The AI Content Trap: Multiplying Mediocrity at Scale

Marketing has always suffered from a volume addiction; however, the advent of generative AI has turned a bad habit into a terminal condition. In the recent discussion with Volker Hildebrand in our CRMKonvo, we explored the uncomfortable reality that while AI has made marketing faster and cheaper, it has largely failed to make it better. The cynical view, which I happen to hold is that marketers frequently confuse the amount of content produced with the actual impact on the customer. We are now in an era where everyone has the same tools to flood the market with what in the words of Volker just “multiplies mediocrity” – or in mine creates instant mediocrity.

The core problem is that generative AI multiplies mediocrity by definition. It ingests existing data and spits out an average of what is already there; consequently, when every startup uses these tools to build their websites and social posts, they all end up saying the same. If you look at the CRM space today, the messaging is often nearly indistinguishable. Everyone promises "revolutionary" efficiency and "seamless" integration. As Volker noted, this is a trap for startups; if they cannot differentiate their story, they simply will not survive the noise.

TL;DR

If you want to watch the full CRMKonvo, please go ahead here (optimized for smartphones) or here (optimized for tablets/computers).

Else, be my guest and continue to read.

Or do both …

The Productivity Mirage

Vendors love to sell AI based on productivity gains. They promise you can save 20 percent of your time on content creation. But as we discussed, productivity is a hollow metric if you do not have a plan for that saved time. If you save 20 percent of your time just to produce 20 percent more "slop" or low-quality content, you have solved nothing, nor have you saved anything. You have actually made the problem worse by increasing the background noise for your customers. The real question for any marketing leader hence is how that saved time can be reinvested into understanding and resolving the customer's actual pain points.

Product Marketing: The Center of the Universe

Volker makes a compelling case for Product Marketing as the organizational "center of the universe". In a tech landscape where products are increasingly complex, the role that sits between engineering, sales, and the customer is the only one capable of maintaining narrative integrity. He argues that this role should report directly to the CEO. This is because the items product marketing owns: pricing, packaging, roadmap strategy, and win-rate optimization: are the literal lifeblood of the company.

If you bury this function under a traditional marketing silo, it becomes a "content factory" for sales decks and brochures. When it reports to the top, it becomes a strategic filter. In the age of AI, this filter is more necessary than ever. AI can draft a battle card, but it cannot understand the nuanced political reality of a specific enterprise buying center.

The "center of the universe" concept is about architectural integrity. Product Marketing must ensure that the technology actually solves a business problem rather than just serving as a shiny new feature to mention in a press release. If the messaging does not address what keeps the customer up at night, then all the generative AI in the world will not improve your win rate.

From Personalization to Individualization

We have been chasing "personalization" since the 1990s; Volker’s PhD thesis touched on it back in the 90s. Yet, most of what passes for personalization today is still just "segmented mass marketing". The real shift happens when we move toward true individualization. This could be the death of the "campaign" as we know it. After all, a campaign is, by its nature, a scattergun approach that is irrelevant to most people in the target group.

True individualization, powered by a combination of predictive and generative AI, means the customer journey is unique to the person. If Thomas visits a site, he sees the architectural whitepaper because the predictive engine knows he’s an analyst. If a procurement officer visits, they see the ROI calculator. This isn't just swapping a name in an email: it's a dynamic reconstruction of the entire engagement layer.

The Rise of the Agentic Customer

Perhaps the most significant strategic shift on the horizon is the rise of the "agentic" customer, the shift from B2B (Business to Business) to B2A (Business to Agent). We are rapidly approaching a time when the initial 80 percent of a purchase journey isn't conducted by a human researcher, but by an AI agent. When 80 percent of the interaction is bot-to-bot, traditional marketing fluff becomes useless. How do you market to a bot? You can't appeal to its emotions with a fancy hero image or a catchy slogan. You have to provide structured, high-quality, verifiable data that the agent can ingest. The bot cares about structured data, proof points, and reliability. This will force a radical return to "the fundamentals" Volker repeatedly mentioned: reliability, relevance, and proof points. Marketing leaders must prepare for a reality where their "customer" is no longer a person, but an agent looking for the most rigorous solution to a defined problem.

A Pragmatist’s Guide to Avoiding the AI Slop-Pocalypse

If you are an enterprise leader looking to "AI-enable" your customer experience, stop listening to the vendor slide decks for a moment. Take a breath. You might be about to make a very expensive mistake if you don’t follow three simple rules.

Here is the pragmatist's guide to not making that mistake.

Prioritize Predictive Over Generative

Currently, everyone is obsessed with the "Gen" in GenAI, but for CX, the "Predictive" side is often more valuable. Don’t just buy tools that just help you write faster. Instead, use AI to identify patterns: which customers are about to churn, which leads are actually ready to buy, and what content actually helps close deals. As Volker noted, tools that track actual consumption or customer journeys (did they stop at the pricing page?) provide infinitely better data than "clicks". Use AI to find the needle; don't just use it to make a bigger haystack.

Kill the "Auto-Pilot" Content

If your marketing team is using AI to generate content and pushes it directly to customers without a rigorous human-in-the-loop review, you are actively eroding your brand equity. AI-generated content is, by definition, an average of everything that already exists on the internet – it’s instant mediocrity. It cannot innovate; it can only regurgitate. Use AI for drafts, for brainstorming, and for reformatting (e.g., turning a long webinar into short clips, or a blog), but never for the final "voice". Authenticity is about to become your scarce resource.

The Training Gap is a Strategic Risk

You cannot simply buy a subscription for a tool and expect "transformation". The most common failure point is the "dump and run" approach. If you aren't investing in training your people on how to prompt, how to critique AI output, and how to integrate these tools into a unified RevOps workflow, you are just buying shelfware, or worse, something counter-productive. AI is a skill, not just a software category. If your team doesn't understand the "human-in-the-loop" necessity and how to work with AI, they will eventually be replaced by the very mediocrity they are producing.

Welcome to the second quarter of 2026. The dust of the generative AI explosion seems to have finally settled, so actual business realities can be seen. For the last few years, the enterprise software market has been drowning in vendor promises of AI magic. Now, companies are waking up to the hard truth. AI is no longer a futuristic promise; it is a budgetary line item with concrete expectations. As our guest Clint Oram accurately pointed out in our CRMKonvo sit-down, businesses are actively hunting for 20 to 40 percent productivity gains from their knowledge workers. But are these gains real, or just another SaaS vendor hallucination? The market is scrambling to figure out what actually works and what is just expensive hype.

TL;DR

If you want to watch the full CRMKonvo, please go ahead here (optimized for smartphones) or here (optimized for tablets/computers).

Else, be my guest and continue to read.

Or do both …

While the underlying LLMs have become core components of daily workflows, the execution at the enterprise level remains often fraught with mediocre strategies. At the same time, we are seeing a profound shift in how work is accomplished with the help of AI. This year will be defined by a massive, societal scramble to understand if, and if so, how, this technology supports the bottom line of the companies using it. Let us see if there is actually any substance there, or if we are just increasing vendor revenues. The focus must shift from adoption at any cost to architectural integrity, and it already does in some areas.

Vendors love to sell you a tidy vision of a fully autonomous future, but the reality on the ground is different and far messier. Integration into legacy systems is painful. The data architectures required to make these systems hum are often neglected in favor of flashy superficial updates. We must rigorously question every new "feature" that hits the market and understand. Its value for us.

The Customer Experience Disaster

Perhaps the most glaring failure in the current AI landscape is the impact on customer experience. Companies are desperately trying to cut costs by replacing human agents with AI bots, and the results are regularly embarrassing, even infuriating. Clint shared a chilling anecdote about interacting with an AI screening agent over the phone. He described the interaction as an Interactive Voice Response system on steroids. It was a frustrating wall erected between the customer and a human being. The AI could comprehend the words, but it lacked the fluid capability to navigate a non-scripted conversation. Deflection, not service.

When organizations deploy AI merely to deflect customers rather than serve them, they are missing the entire point of a modern CRM strategy. This is not innovation. This is just a cheaper, faster way to annoy your buyers. Well, that’s kind of an innovation, too; but probably not a desirable one. The market must understand that AI replacing human agents is still failing miserably. The technology simply cannot yet handle the nuance of human frustration. Conversely, AI augmentation of human agents is where the real value lies for time being. When AI works in the background to provide context to a live agent, the customer wins. We must stop treating AI as a cost-cutting guillotine and start treating it as an enablement engine.

Context Blindness and the SDR Spam Machine

The root cause of these failures is what Clint terms "context blindness". LLMs are incredibly articulate, but they are incredibly stupid without specific, grounded data; even with that, it is still worth to follow the trust-but-verify principle. If you drop an AI into a workflow without connecting it to your CRM, CDP, or a robust knowledge graph, it will confidently generate useless responses. The industry is finally realizing that localized context is the missing link.

This is acutely visible in the sales area. Inbound AI Sales Development Reps (SDRs) are performing reasonably well at qualifying leads. Tools like Regie.ai and Win.ai are successfully routing inbound interest. However, outbound AI SDRs are an unmitigated disaster. They disappoint prospects with formulaic, obvious AI spam. Buyers immediately recognize the lack of human nuance and discard these messages. You cannot automate relationships with generic prompts. The tools exist, but companies are burning through their prospect lists with low-quality, automated outreach. It is the new version of spam, and it is destroying brand equity. To fix this, vendors must prioritize architectural integration over generative party tricks.

Seniority Beats Juniority: The Unlikely AI Masters

Here is the most fascinating observation from Q1 2026. The so-called digital natives are losing the AI race. We assumed the younger generation would master AI effortlessly. Instead, older, experienced professionals are extracting vastly more value from generative tools. Why? Because effective AI usage requires deep domain expertise. You have to know what questions to ask, and even more importantly, you must have the experience to judge whether the AI's output is correct or just plausible nonsense.

Clint points out that delegating tasks to AI is identical to managing a junior employee. You must give precise instructions, provide context, and meticulously review the work. Senior managers who know how to delegate are thriving. They are building complex slide decks and strategic documents in minutes instead of weeks. Seniority beats juniority in the AI era because wisdom and context cannot be downloaded. The prompt is only as good as the professional typing it. But remember, not having juniors today means not having seniors tomorrow.

The Exhaustion of Accelerated Productivity

Finally, we cannot ignore the human cost. The pace of technological change is causing massive AI burnout. Professionals are generating high-quality work at breakneck speeds, leading to a relentless, continuous stream of high-impact decision-making. The mental fatigue of constantly managing AI agents and making rapid-fire strategic choices is exhausting the workforce. You are no longer doing the rote work; you are just making decisions all day long. And this will increasingly distance you from your domain expertise, which means that decisions may be taken at an increasing level of uncertainty.

This also means that the conversation around a four-day workweek might no longer be an HR perk; it might become a physiological necessity. When you remove the friction of content creation, you are left with the intense cognitive load of continuous evaluation. We are burning out our best people by forcing them to operate at the speed of a machine. The balance must be restored before the productivity gains collapse under the weight of human fatigue.

Reality Check for Enterprise AI Buyers

Let us cut through the vendor noise and establish ground rules for buying AI in 2026. If you are an enterprise buyer looking to inject AI into your customer experience architecture, stop buying hype and focus on reality. Here are three crucial learnings and recommendations.

Integration Realities Require Starting Small.

Do not attempt to boil the ocean with a massive AI rollout. As Clint advised, you must start small, think big, and move quickly. Identify a highly specific friction point in your business, such as territory planning, and deploy a targeted AI solution like BoogieBoard to solve it. Technology implementations stall when organizations lack a clearly defined problem. Stop buying AI just to have AI. You need a targeted business case. Force your vendors to prove their worth on a micro-scale before expanding.

Data Quality Triumphs Over Generative Hype.

Your shiny new LLM is entirely useless without context. Stop obsessing over foundational models. Instead, focus on your internal data structures first. Curing context blindness means feeding your AI localized, curated data through retrieval-augmented generation (RAG) and structured knowledge graphs. If your CRM data is garbage, your AI will simply generate garbage at unprecedented speeds. Context is the only thing that separates a useful tool from a hallucinating liability. Do not let vendors convince you that their AI will magically organize messy data.

Third: The Human-in-the-Loop Necessity.

Do not replace your human workforce with cheap AI alternatives. The technology cannot replicate the empathy and strategic judgment of a human being. Focus entirely on AI augmentation. Use AI to feed context to sales representatives at the exact right moment, rather than using it to spam prospects with outbound garbage. Keep the human in the loop to handle complex escalations. Your customers deserve a premium experience, and humans are still the only ones who can help them have it.

As enterprises scramble to deploy AI, the Great AI Debate’s eighth installment reveals a widening gap between what vendors are selling and what actually works at scale. Dr. Michael Wu and Jon Reed spent this episode cutting through the hype around language models, domain expertise, and the financial reality of building sustainable AI systems; and they didn’t pull punches about where the field is failing.

TL;DR

If you want to watch the full CRMKonvo, please go ahead here (optimized for smartphones) or here (optimized for tablets/computers).

Else, be my guest and continue to read.

Or do both …

The Domain Expertise Imperative: Correlation is Not Causation

One of the most dangerous, and frankly lazy, narratives pushed by AI maximalists is the idea that artificial intelligence negates the need for deep domain expertise. This is a fundamental misunderstanding of how these models work.

As Dr. Michael Wu frequently points out, almost all machine learning and AI systems today are built using supervised or reinforcement learning. They are, at their core, sophisticated correlation engines. They do not understand causality. They can surface 50 variables that move together, but they cannot tell you whether A causes B, B causes A, or if a hidden confounding variable C is responsible for both.

If an LLM correctly states that smoking causes cancer, it is not because it understands the biological mechanisms of cellular mutation; it is because it has been fed enough human-generated text asserting that relationship. It creates the illusion of causal reasoning without the substance.

This is precisely why domain experts, whether in healthcare, supply chain logistics, or financial services, are more vital than ever. The AI can process the data at unprecedented scale, but it takes a human domain expert to identify spurious correlations, recognize hidden causative factors, and make the final judgment calls. The machines are there to augment the experts, not replace them. So far, they simply can’t.

Deconstructing the Anthropomorphic Illusion

We must also fiercely reject the anthropomorphizing of AI. It is a fundamental human flaw to project human traits onto inanimate objects. It’s a psychological quirk that leads people to form emotional attachments to digital chatbots or, absurdly, even "marry" them because a machine, unlike a human, requires no compromise and has no expectations.

In the enterprise, this anthropomorphism manifests in the careless misuse of terminology. We call these systems "intelligent," giving ourselves an excuse to let our guard down. We use terms like "grounding" to imply that an LLM has a fundamental anchor in truth.

Let’s be clear: LLMs are not inherently "grounded." You can feed them contextual data, such as a causal graph or a vector database via Retrieval-Augmented Generation (RAG), to influence their outputs at inference time. This is valuable and important, but it is not true grounding. The underlying architecture remains a Transformer: a probabilistic, non-deterministic engine that samples from a distribution to guess the next word. If you ask it the exact same question twice, chances are it gives you two different answers. Heck, chances are that the answer is plain wrong. That’s why they all have a disclaimer. It is not reasoning; it is calculating probabilities. Until we develop architectures that sit on top of a truly grounded foundation of world logic, we must remain vigilant against the illusion of machine certainty.

The Myth of "Bigger is Better"

The market is currently obsessed with the idea that the path to Artificial General Intelligence (AGI) is simply a matter of scale; more compute, more parameters, bigger models. I disagree.

We are reaching the point of diminishing returns for massive, generalized models. A model that can write a Shakespearean sonnet, pass the bar exam, and debug Python code is intellectually fascinating, but functionally excessive for most enterprise needs. I do not need my supply chain optimization algorithm to “understand” 18th-century poetry, nor to code in Python.

The future of enterprise AI lies in rightsizing. Once we distill these massive models into smaller, highly focused, domain-specific architectures, perhaps utilizing a Mixture of Experts (MoE) approach, the cost to run them drops drastically. Smaller, smarter, and tightly scoped models deliver vastly superior ROI because they solve specific business problems without the bloat and compute costs of a massive frontier model.

The Value Equation: Digitization vs. Transformation

Ultimately, we must confront the value equation. Are we actually transforming our businesses, or are we just using AI to do the same inefficient things – just faster and more often?

If you take a broken, convoluted business process and simply layer an AI agent on top of it to speed up the keystrokes, you have not achieved digital transformation; you have merely digitized a bottleneck. True enterprise value comes from using these technologies to fundamentally reimagine workflows, ask questions of your systems that were previously impossible to query, and empower employees to operate at a higher strategic level.

Governance in the Age of AI

Finally, we cannot ignore the geopolitical and regulatory backdrop, such as the EU AI Act. While some decry regulation as a handbrake on innovation, pragmatic risk frameworks are essential; my emphasis is on pragmatic here. Bad actors will not slow down, so neither can we. However, we must limit our deployments to lower-stakes environments until we fully understand a model's limitations. You cannot hold an AI accountable when it makes a critical error in a clinical diagnosis or a financial audit. Accountability remains—and must always remain—with the humans who deploy it.

The score in 2026 is clear: AI is a powerful utility, but it is not magic. Enterprises that succeed will be those that pair domain expertise with right-sized, cost-effective models, applying them to genuinely transformative use cases while maintaining strict, human-led governance.

Three Things Enterprise Software Buyers Should Take Away

These are the hard truths for CX software buyers, paired with the uncomfortable questions they need to ask vendors to cut through their marketing noise

Call out the "intelligent" BS

Vendors love claiming their CX chatbots are "grounded" and can "reason". Spoiler: they can't. They are just probabilistic auto-completes guessing the next word based on correlation.

Ask the vendor:

• Since your model is fundamentally probabilistic and non-deterministic, what hard, deterministic guardrails actually stop it from hallucinating a fake refund policy for our customers?
• "Are you using the term 'grounded' just to describe a basic RAG (Retrieval-Augmented Generation) setup, or is there an actual foundational world-model logic driving this?"

Keep your human experts

AI is completely clueless about cause and effect. If you want to resolve complex customer issues instead of just failing at scale, you need your human domain experts to babysit the AI and catch its inevitable blind spots.

Ask the vendor:

• LLMs don't inherently understand causality, they just mimic it based on training text. How does your system allow our domain experts to review, correct, and input hidden causal factors that the AI missed?
• Is this tool designed to genuinely augment our tier-3 support agents, or is it just a generic copilot that's going to force them to spend more time auditing its mistakes than actually helping customers?

Stop paying for bloated models

You don't need a massive LLM capable of writing Shakespeare just to route a simple customer complaint. Demand smaller, domain-specific models. They do the job better, and you won't get fleeced on unnecessary compute costs.

Ask the vendor:

• Why should we pay the massive compute premium for a generalized frontier model when a smaller, right-sized model trained specifically on CX data and use-cases would be cheaper and more effective?
• If we strip away the massive scale of the underlying LLM, what unique, domain-specific value is your company actually bringing to our specific industry's workflows?

For the longest time, companies have been trapped by enterprise software vendors.

First by shrink-wrapped software packages.

Then by SaaS offerings.

Both situations led to what one even in a SaaS world can call shelfware – although these days the shelf is a virtual one instead of a physical one. Buyers still get enticed to purchase more capabilities than they need, which leads to them paying more than necessary while often using software packages that offer overlapping capabilities.

One of the promises that SaaS started with, was to end this. Sadly, it looks like this promise was not kept. And this is no wonder; after all vendors want to be sticky. And they need to have increasing revenues. This means that they need to offer an ever-increasing number of capabilities, aka features, to warrant their pricing and eventually regular price increases. Combined with the frequently used strategy of offering related capabilities, i.e., seats for an adjacent software that is not yet needed by a customer, this led to two things: bloat and shelfware. Both go at the expense of the enterprise buyer.

Since the dawn of packaged software, the argument to buy, i.e., to voluntarily step into this trap, is the same: Buying is cheaper than building.

Which probably was correct. Buying from a specialist was the logical choice. Engineering talent was, and still is, scarce. Building software includes a lengthy process of requirements engineering, years of development and ultimately never-ending maintenance.

Just that most of this is true for most implementations of purchased enterprise software, too.

And the buying process is arguably broken. Need identification is often done without the right stakeholders, the software selection becomes a procurement-heavy process that is more based on checking boxes than in fulfilling user needs and the implementation turns out to be a death march. Who has not read – or at least heard of – the statistics that show implementation failure rates of more than 60 percent?

But then, who got ever fired for buying IBM, or Salesforce or SAP, for that matter. Or Oracle? Take your pick.

The result?

As a consequence, we see processes that are not improved or that not necessarily differentiate the company, as they are either implemented to follow the “same ole” or along “best practices”, which often translates to “average”, i.e., mediocrity. Users are forced to adapt to the tool, and not the other way round. Their pain is not solved. This, combined with shelfware, contributed to low adoption and shadow IT which ultimately harms all efforts of a digital transformation.

And it is costly.

Entry low-code, no-code and GenAI

Low- and no-code environments are basically there since, well … forever. At least as measured in Internet time. I had my first experiences with one back in 1995 (yeah, I am that old).

Depending on who got its fingers on these environments, results have been good or not so. Anyone remember the infamous Lotus Notes app graveyards? It needs guardrails.

However!

The combination of low-code, no-code and generative AI has the potential to reduce the marginal cost of software development to almost zero. Instead of engaging into a multi person year software implementation project, it is now theoretically possible to “vibe-code” a bespoke application in a short time and at low cost. The scarcity of IT personnel is mitigated, and the procurement process is no hurdle anymore.

At least theoretically. Again, it needs guardrails and the right tools for the right people.

Still, and this is important, it is now possible to create what one could call a throwaway MVP, or a working prototype, that covers a requirement’s happy path at almost zero cost. To be clear, this is a capability that we didn’t or only barely have with all the traditional low-code and no-code environments. And this is a big deal.

With this prototype it is possible to quickly identify whether a real problem is solved or at least mitigated; and this before big money is spent for the customizing and deployment of a new SaaS solution. This flips the procurement process to something for which one could use the term prompt-to-product.

A new paradigm?

As said, the traditional software procurement lifecycle: requirements identification → software selection → implementation is flawed. It relies on abstract and static written requirements. Text is ambiguous whereas software is explicit. The gap between a fuzzy, written requirement (e.g., "The system must support flexible workflows") that we see all too often and the delivered reality is where millions of dollars in enterprise value can get tanked.

With the help of Generative AI, it is possible to establish a methodology that moves the build phase to the very beginning. It serves not as the delivery mechanism, but as an agile discovery tool in a three phased process.

Phase 1: Dynamic Discovery

Instead of collecting stakeholder needs in a static document, this process begins with live prototyping. Business stakeholders work with an AI engineer or directly with an LLM-enabled no-code environment to describe their problem in natural language, rapidly developing a working prototype that supports the happy path to the desired outcome. This is agile development on steroids. The prototype does not need to be secure or scalable; it only needs to fulfill the job and be interactive. As a result, it becomes very clear what the users actually want. Plus, some implicit requirements get surfaced early in the process instead of after the purchasing decision and project budget assignment.

Questions like “Does the user actually want a dashboard, or just a daily email summary?” or “Does the data structure actually fit the way the team works?”, and more, are answered before they require costly change requests.

Phase 2: Stress Test

After the prototype solves the business users’ pains, IT leadership is in a better position to decide whether to buy, build, or opt for composing a low-code solution. Based on the assumption that the existing software packages do not cover the requirements, this decision can be taken by answering three main questions based on the generated code.

• Does this tool need to read/write to business-critical system like the ERP, or does it live in isolation?
• Does the logic involve high-liability calculations (tax, payroll, health data)?
• Is the logic static, or will it require constant updates based on external factors (e.g., changing shipping tariffs)?

Phase 3: Strategic Fork

Based on the answers, the organization moves down one of three paths. Crucially, the outcome of phase 1 is valuable on all three paths.

Build

If the prototype is self-contained, low-liability, and specific to the company’s internal operations, the decision is to build.

The prototype code gets refined to cater for edge scenarios and for compliance and security, if the development environment of the prototype didn’t already take care of these. After that, it can get deployed.

Because the cost of generation stays at near zero, the resulting software is disposable. If the process changes over time, the application is not patched but simply discarded and regenerated.

As a result, the company has a solution with perfect process fit, low implementation cost and zero licensing fees.

Buy

If the prototype reveals that the requirements are more complex than anticipated, for example, if there are more regulations to consider, the decision is to buy. In contrast to the traditional process, this is now an informed decision.

The organization stops building but uses the functional prototype as a key part of the RFP that demonstrates the desired process. The conversation shifts from "Can you meet our requirements?" to "Here is exactly how our process works; demonstrate that your software can replicate this specific behavior."

The result is risk mitigation for both the company and the winning vendor. The prototype proves that building potentially creates unmanageable technical debt. It also prevents buying vaporware by forcing vendors to prove capability against a live model. For the vendors, it takes away considerable uncertainty in assessing the project size.

Compose

If the prototype requires the flexibility of custom logic but the governance of a standard platform (Microsoft, SAP, Oracle, Zoho, Salesforce, etc.), the decision is to compose it using a low-code environment.

The AI-generated logic gets transferred to an existing low-code/no-code platform. This platform handles identity management, UI standardization, and hosting, while the generated code still handles the unique business rules.

This enables speed of deployment with the safety net of IT governance.

What does this mean?

Executives should flip the purchasing process using three key actions.

• Provide an infrastructure that allows for rapid, AI-supported prototyping, aka vibe-code environments. Ideally, this environment already embraces security and compliance rules.
• Train users, business analysts or IT personnel to use this environment to bridge the gap between business and AI.
• Instead of asking for written requirements only, make it the creation of prototypes in this environment that serve as core elements of the demand mandatory.

This flipped process opens the build vs. buy question to no longer being binary. It creates a build-to-define process that ensures that the decision of how to deliver required functionality in a better informed, de-risked way that has a higher chance for success at lower cost. It isn't killing SaaS but stopping to buy hope. Low-code/no-code in combination with GenAI can help to know more exactly what gets delivered, regardless of whether you build or buy.

These days, we are drowning in conflicting information about the value of generative and/or agentic AI. I, myself am researching for good studies that dive into the ROI that is generated by this technology, with limited success. Most information is anecdotal, or comes from success stories, which cannot get used too literally.

Two major 2025 reports from MIT and Wharton, respectively, paint starkly different pictures of AI adoption and adoption success. While the meanwhile often quoted MIT NANDA “report” on the state of AI in business often gets quoted with 95 percent of all businesses not getting any ROI from their gen AI initiatives, a recent study by the Wharton Business School shows a very different result with 74 per cent of enterprises showing a positive ROI. Why is one so pessimistic and the other so optimistic? As I have written before, a closer look at the data reveals the 95% "failure" narrative is a myth, or even a scare, and the real story is probably a different and far more differentiated one, which Wharton names Accountable Acceleration.

Is GenAI really a 1-in-20 lottery ticket or is it rather a core business function? So, let’s have a look.

Methodology matters – debunking the 95% failure rate

In contrast to the NANDA “report” that relies on a fairly small sample of about 150 survey responses and 52 structured interviews, the. Wharton report bases on a large-scale, quantitative and longitudinal study. It surveyed around 800 senior decision-makers at businesses of different sizes and is tracking trends for the third consecutive year. Therefore, its data is built for statistically valid conclusions.

In addition, MIT NANDA’s rather sensational number of a 95 per cent failure rate is the result of a different focus. This “report” solely looks at custom, task specific agentic AI implementations that reach production state and measurable P&L impact. It does not look at the value of widely adopted generative AI tools like ChatGPT, Copilot or Perplexity. Instead, it dismisses their impact by labeling them as as tools that primarily enhance individual productivity, not P&L performance. And this, despite their obvious success that the report also identifies. 40 per cent of general purpose LLM implementations go into production with about 90 per cent of all employees using privately purchased licenses for ChatGPT, Perplexity, et. al. This is an indication that the corporate implementations are not exactly looking at the right pain points.

In contrast to this, the Wharton report identifies this as the primary source of value creation. Its data shows the top use cases are data analysis, content creation and its summarization and presentation (p. 30). This is the ROI. The MIT report is like arguing 95% of companies get no return from email because it only boosts individual productivity. This doesn’t make much sense as an increase in productivity is a return. Wharton, in turn, identified how businesses identify and measure ROI and found that almost 75 per cent of all businesses report a positive return on investments, with smaller businesses (p. 45) and some industry sectors (p. 47) seeing a higher return. But the picture is clear.

The ROI story – mainstream and measured

Given all this, the Wharton study seems to be a far more credible source of information about the current usage and ROI of generative AI. There is only a small fraction of businesses that sees a negative ROI, with some businesses seeing it as neutral or too early to assess.

The numbers vary across industries, but the general picture is clear. Generative AI is not a high-risk gamble. On the contrary, if implementations are managed appropriately, it does deliver results and consequently, Wharton concludes that we have now regular usage in core business operations, with embedded ROI metrics. More than 70 per cent of the surveyed organizations are formally measuring ROI, with a focus on productivity gains and incremental profit.

There is both, an increasing usage of generative AI and a continued significant investment, including into own research and development. Wharton dubs this phase as “accountable acceleration”, moving on from exploration- and experimentation-oriented phases in the previous years. I guess, there needed to be a catchy marketing phrase …

And, as said above, it’s showing results. Of the surveyed businesses, 74 per cent are seeing positive returns. This includes 35 per cent reporting "significantly positive ROI" and 39 per cent "moderately positive ROI". This data flat out contradicts the 95 per cent failure narrative of MIT NANDA. Apparently, we are not looking at a divide with a 95 per cent chasm but rather a far smaller spectrum of adoption speed and maturity.

Buy – or rather build?

Now that we are clear about well-managed generative AI initiatives showing a positive impact, the make or buy question looms again. NANDA quite unequivocally says “buy”, as they find that the failure rate of custom implementations is double the failure rate of strategic partnerships with a vendor and systems integrator.

Wharton roughly shows an equal distribution of efforts in new technology, enhancing existing technology and internal research and development efforts.

Which is not necessarily a contradiction, as it basically says, “buy and adapt”. Some scenarios can be supported by delivered software, some needs training or fine tuning, other models need to be specifically built.

Businesses are adopting a far more sophisticated hybrid strategy than simply make or buy. Firms are not just "buying" off-the-shelf tools; they are investing significant capital to build custom, proprietary solutions that drive competitive advantage.

The MIT report's "buy, don't build" advice, in contrast, is very simplistic. While I stick to the recommendation that I made in my article Beyond the Hype: Unlocking GenAI ROI in the Enterprise, the truth is more nuanced. It is basically the same as for the implementation of enterprise software in general. Stick to best practices where there is no significant or lasting competitive differentiator and invest into custom implementations where a differentiator with a competitive advantage lies.

The real barrier

Which leads us to what prohibits businesses from taking even more advantage from the use of generative AI. As Wharton shows, the main problem is not the technology, at least not only. There are credible studies that show a decreasing likelihood of success with increasing complexity of the modeled scenario, including CRM-Arena Pro and TheAgentCompany, but then these agree with the Wharton findings that success is very well possible.

The technology is mostly ready.

The real barrier is people and organization – essentially corporate culture. Management needs to set an example, people need to be educated and, very importantly, governance needs to be in place that both helps the users and protects the business. This requires change management. According to Wharton, the toughest challenge facing businesses are a lack of skills, employee fear for their jobs and management’s ability to constructively manage the necessary change by convincing the employees that the use of AI is not about technology grabbing their jobs.

From sensationalism to strategy

What we can confidently say is that the generative AI landscape is not as dystopian as the NANDA “report” makes it appear. This narrative is created on a too small data set and a questionable definition of “value”. It is fundamentally biased by an underlying agenda – which comes from the authors pointing a way into their solution of overcoming the technology problem. The Wharton "Accountable Acceleration" study provides the more data-driven, and optimistic view. Generative AI has gone mainstream. The return of investments can get measured, is being measured, and it is proving positive, with few exceptions.

The "GenAI Divide" isn't a chasm between success and failure, caused by poor technology.

It is rather a spectrum of maturity and of connecting investments into generative AI into strategic business KPIs. And, of course, selecting the right problems to solve. The real challenge is not to find a silver bullet that automagically solves business challenges. It's the boring work of business transformation. It is about investing into and aligning people, investing in training, setting smart guardrails, and executing a hybrid build and buy strategy.

This is the real, actionable roadmap to success.

If you want to explore, how this roadmap could look like for you, get in touch with me for an informal conversation.

My past two column articles on CustomerThink dealt with how to determine the return of agentic investments and whether agentic AI delivers at all.

The question of ability to deliver is particularly interesting for me, as I am researching measurable results other than cost savings in contained business areas for some months now, and regularly find a very strong focus on customer service and marketing, with customer service functions being best able to report measurable results. This is evidenced by the number of success stories I find, supported by the publication of a recent TEI of Zendesk customer service study. 

However, most of this is anecdotal evidence, or vendor sponsored/commissioned. And which vendor likes to speak about failures? Similar for buyers who understandably do not like to be in the spotlight with investments that turned out to be less than successful. There hasn’t been too much in depth research on whether generative and/or agentic AI deliver to promise or not. 

Luckily, there has been at least some research evaluating the capabilities of LLM based AI agents in business environments published this year. CRMArena-Pro by Salesforce Research naturally has a focus on CRM tasks across B2B and B2C scenarios. The authors identified nineteen tasks commonly executed in CRM systems and categorize these tasks in the four business skill categories database querying and numerical computation, information retrieval and reasoning, workflow execution, and policy compliance and includes a confidentiality awareness evaluation. TheAgentCompany on one hand covers a wider area along the business value chain but on the other hand has a narrower focus on software engineering companies. One other main difference between these two studies is that TheAgentCompany has a focus on more complex tasks that require multiple steps for their execution.

Both studies find that LLM-based agents deliver in simpler contexts while they are still failing in more complex ones.

Not that this comes as a surprise.

While this consistent result gives an indication on what scenarios to avoid, there is little help in what actually to do – besides of starting with simple scenarios and to focus on workflows. 

This is where a recent study from the MIT NANDA project gives additional insight. The study, titled “The GenAI Divide – State of AI in Business 2025” starts off with a quite unsettling finding: Only five percent of organizations are extracting value from their integrated AI pilots. The rest – a staggering 95% - doesn’t receive any measurable P&L impact. 

Why? 

According to the study, the problem is the missing ability of integrated systems to retain feedback, adapt to context, or improve over time.

That’s a big ouch, given that enterprises often invest multi-million dollar budgets into these initiatives.

This is also in stark contrast to what emerged as shadow AI. About 90% of all employees are using privately purchased licenses for ChatGPT, Perplexity, Claude, or other tools

Figure 1: The steep drop from pilots to production for task-specific GenAI tools; source The GenAI Divide

And they are happy with how these consumer-grade tools help them in their work environment. So, apparently, employees are convinced that genAI tools can and do help them in their work environments.

So, the important thing is not to cry out failure but to identify what businesses and vendors can do to make investments into generative and agentic AI a success. The GenAI Divide – State of AI in Business 2025 gives plenty of insight for both. 

So, let’s have a look under the hood of this report

What can buyer executives do?

First things first, buy, don’t make. This doubles your chance for success.

• Look at the right problem and use the right KPIs. A lot of budget, around 50% of it, is sunk in sales and marketing. While this seems sexy and results can be showed off with easy to gather (and misattribute) KPIs, the real gains seem to lie in back office automation. To mitigate the still existing weaknesses in the automation of complex, start simple. But think big. This is corrobated by the findings of both earlier studies, TheAgentCompany and CRMArena-Pro.
• Look at the right partners. Do not only rely on artificial benchmarks but on business outcomes. The world of AI is different from traditional SaaS. AI is service as a software, so consider your vendors outsourcing partners instead of software partners. Have them prove that their AI tools can be configured to your needs – and hold them accountable.
• Usability and flexibility are key. One of the core reasons for the success of tools like ChatGPT, Perplexity, Claude or Gemini is the simple user interface plus their ability to be guided through the problem-solving process in conversations. These conversations resemble iterations and are regularly the way humans work. Technologies like RAG and RAC help, but still need improvement. Again, start with the simpler problems and go from there.
• Trust your people. Your employees do know what they need. This is abundantly clear through their use of privately sourced AI tools. Do not centralize AI initiatives but have them driven where they matter while providing a meaningful yet robust set of guidelines.

Plus, here’s a bonus tip: All the above is particularly important for enterprises. As opposed to SMBs, enterprises tend to be more vulnerable to internal politics. For the sake of sustaining success, this should be avoided.

What should vendors do?

Vendors should have a close understanding about what their clients are – and have a hard look at what they actually need and want. GenAI and agentic AI are the shiny new tools, still, they need to solve actual business problems. This means that it is important to not build generic tools but solutions that are capable of embedding themselves into business workflows, improve them. This requires a strong focus. Grow from there. If the big, established vendors don’t do that – there is a startup that is capable of disrupting them and eating their lunch. Start with solutions to problems that are not business critical, and extend from there.

One of the major complaints that users have about the tools at hand is that they do not learn and are too rigid. So, genAI tools need to incorporate learning mechanisms that help them to learn via supervised, or unsupervised, means. They also need to provide a context window that is big enough for the complexity of business challenges that are addressed. Lastly, even if a process looks the same, it often isn’t. As a consequence, build in a deep ability for configuration, maybe even customization.

A bonus tip for startups. Many a buyer does not even consider you. They don’t know you, they don’t trust you. Heck, they don’t even know whether you’ll be around tomorrow. So, your best bet to get into larger accounts is by partnering up with vendors or consultants that your buyers know and trust.

Vendor or buyer. Do you need some help? Give me a call. I am there.

The news

On October 9, 2024, Zendesk held its AI Summit in New York’s Chelsea Industrial. The AI Summit is an event mainly for customers to inform themselves about what is new at Zendesk but also to network with each other. The event featured an interesting lineup of customer and partner speakers, headlined by New York Times bestselling author and podcast host Kara Swisher.

My estimate is that there have been more than 250 customer representatives in attendance who not only could listen to the speakers but also get in-depth demos of Zendesk’s updated offerings, following real-life use cases.

True to its name, the event centered around the use of AI, in particular bots, to increase not only efficiency, but also customer- and employee satisfaction. CEO Tom Eggememeier opened the event with an emphasis that Zendesk’s AI is built to support humans by stating that it “is designed for humans”, and Zendesk’s service solution is built to strengthen the human – AI partnership.

Kara Swisher talked about the promise and peril of AI, giving the audience some food for thought on the day after Geoffrey E. Hinton, the godfather of machine learning turned AI warner got co-awarded the 2024 Nobel Prize in Physics for his “foundational discoveries and inventions that enable machine learning with artificial neural networks”. While Swisher sees the value that the use of AI can bring, she, too, warned about the hurdles that still need to be overcome, namely the concentration of power that the technology creates and its immense hunger for energy. The tie into the Zendesk story is that customer service is a prime candidate for the use of AI, which turns information to insight while customer service in general has quite some room for improvement – and this is smack in Zendesk’s territory.

Some of the news revealed is that the functionalities that got announced at Zendesk Relate earlier this year are now in general availability. Zendesk announced them as a series of innovations, including AI-powered agents for omnichannel support, enhanced agent copilot, powerful voice, and an agent builder. The company reiterates its vision of having 80% of all service interactions automated. To achieve this, the company also delivers improved agent autonomy and an enhanced agent co-pilot. It adds that it wants to automate up to 50% of voice interactions by partnering with poly.ai.

The bigger picture

There is a continuing need for enabling a good/great customer experience and to engage correspondingly. In customer service scenarios, this makes it crucial to get responses right, fast. At the same time, customer service is a profession that still suffers from high employee churn, with the corresponding cost of re-hiring, re-training and low employee morale. As Swisher said, there is scope for improvement and a good way to achieve this is to focus on creating a positive “human-AI relationship” that helps human agents do what they can do best by helping them with what an AI can do best, which is delivering knowledge and insights as opposed to have humans research it.

Looking at this from a customer angle, customers want faster and better service, i.e., higher quality of resolutions at a faster time to resolution.  

Achieving this requires the right tools. These tools need to help agents focus on the right cases, give them help for their own improvement and, most of all, keep the mundane off their backs.

From businesses, this requires an outside-in mindset that is not solely focusing on efficiency and productivity as its outcomes, but on becoming “better” in a business sense by looking at employee- and customer outcomes.

My analysis and point of view

Zendesk tells a very compelling customer service story by combining autonomous agents, copilots and a sophisticated QA solution that helps steering a customer service organization into the right direction – namely outcomes for customers, which includes agent coaching as opposed to merely monitoring them. The latter does create a climate that a customer service organization’s most valuable assets – the agents – perceive as negative. Where the story is a bit difficult for me is where CX is used as a synonym for customer service, but this is a different topic.

The audience of the AI summit clearly shows that this story resonates. The demo-booths were well attended, and customers asked lots of questions about how to optimally leverage the capabilities. In fact, those customers, who I had the chance to briefly talk to, looked beyond the usual narrative of increasing efficiency. Instead, they looked at how to provide better internal or external service in an environment that needs to manage with limited resources and personnel. Having said this, not every customer is there yet. Many are still too focused on cost reduction by headcount reduction. This is still the current reality that also AWS General Manager of AI and ML Tia White spoke out, while maintaining a more optimistic outlook about refocusing employees on higher value tasks.

On the other hand, as also a recent Salesforce study shows (free download, requires registration), customer service professionals experience that customer expectations are continuously on the rise; customers ask for more: more service quality, more personal touch, and more speed of resolution. This combination puts customer service organizations between a rock and a hard place. And this is where helping customer service agents with automation and (generative) AI helps. As Zendesk CTO Adrian McDermott and SVP Product and Solution Marketing Lisa Kant emphasized, the adoption of AI is both, a marathon and a sprint. To support this, it needs both, fast time to value through short implementation cycles with well running “pre-trained” models, the continuous improvement of these models with clean data, as well as the careful selection and implementation of meaningful business scenarios plus the analytics to identify them.

The sprint part is the fast implementation and the continuous model improvement that come out of the box.

The marathon part is staying on focus by staying use case aligned and defining relevant KPIs that measure success.

The chosen vendor software must enable a feedback loop that helps establishing this.

In my eyes, Zendesk delivers this for customer service scenarios, although some aspects of it like the prediction of a response perception by the customer or the (semi-) automated improvement of knowledge bases, require complementing software. So, buyers who are in the market for a new omni channel customer service solution should have a good look at Zendesk.

_{Disclosure: Zendesk has paid for most of the travel and accommodation cost associated with my attending the AI Summit. This came with no obligation whatsoever. All thoughts and opinions expressed are solely mine.}

Oracle Cloud World is in the books, Dreamforce just wrapped up, Hubspot’s Inbound event is still on, and there is one key theme that overarches all three events.

And no, it is not Larry Ellison getting all cozy with AWS (or Azure, for that matter). It is also not that his keynote was distinctly geeky, after some years of Oracle putting business solutions to the front. Or that Mark Benioff apparently tore up his keynote in the last moment. It is also not that Hubspot CEO Yamini Rangan found that the sales process is broken and that customers know more about you as you about your customer.

No, the theme is ... drumroll ... you will have guessed it ... AI agents. Oracle's Steve Miranda talked about them at length in a line of business context, while Larry focused on IT, security, and database-oriented agents.

For Salesforce, agents are even more of a topic, dubbing Dreamforce the biggest AI event and Salesforce the most successful AI CRM - both technically right but probably somewhat short selling the full value of both.

For Salesforce, the next big thing is Agentforce, it's AI Agent platform.

And Hubspot announced Breeze, its AI to power the customer platform, which, you guess it, includes agents. CEO Yamani Rangan talked about marketing, sales, and service agents. Co-founder and CTO Dharmesh Shah then spent considerable time in his keynote, talking about agent.ai, Hubspots “professional network for AI agents”.

What struck me watching all three keynotes - Ellison's, Benioff's and Shah’s - is the change from last year's messaging to this year's messaging.

Last years it was all about Co-Pilots, or digital assistants, this year, it is about autonomous agents. Oracle announced more than 50 agents, Salesforce more than 100, Hubspot a humble 4, plus the ones that are available in agent.ai.

So, what is the key difference, compared to last year?

There are basically two. One is that a co-pilot serves and supports the human. With this, the core message was that the machine helps the human; the human is supported by the machine. The human will always be in the loop.

This year, the emphasis is on autonomous agents, whether we already have them or not. And I argue that the goal of autonomy is achieved only for fairly simple use cases as per now – with progress being fast. Still, autonomous agent reads a lot like automation, and there is a distinct difference between the human deciding and taking action vs. the machine "deciding" and "taking action".

This translates to, as Miranda astutely said in his post-keynote press conference, "the early use cases will be a little less autonomous and human assisted". In other words, human helps machine, until they are "more autonomous" and do not need the assistance anymore...

Benioff put it somehow similar in his keynote when he listed the many, many agents that Salesforce already now (well, GA is in October 2024) offers with Agentforce. So many in fact, that I felt inclined to comment that "only the CEO agent is missing to have a fully virtual company".

At the same time, all three, Hubspot, Oracle, and Salesforce, insist in their messaging that their objective is not the replacement of humans by AI but taking away the mundane work, thus improve the employee experience – or employee’s value, in the words of Shah.

According to a number in the early 2024 report New trends in AI use at work by Slack’s Workforce Lab that gets referenced in the recent Salesforce report Trends in AI for CRM, mundane work amounts to a whopping 41 per cent of a desk worker’s workload. Desk workers state that they spend this amount of their time on tasks “that are low value, repetitive or lack meaningful contribution to their core job functions”. If right, this is a clear opportunity for AI and automation, but also a definite opportunity for some process improvement.

Zendesk foresees a whopping 80 per cent of all service interactions being resolved without human intervention. This essentially makes the human the AI's supervisor - which technically is a human in the loop scenario, until that supervision can get automated, too. And automation at scale changes the human role to exception handling. A high visibility example that frequently can be observed is the mandatory stop of trading at stock exchanges. Though this trading technically is not necessarily performed by AI agents, it is often fully automated.

The second stated objective in the Oracle and Salesforce keynotes is improving the customer experience, something that Salesforce's Patrick Stokes demonstrated with an impressive live(?) demo during the Dreamforce keynote. In the case of Hubspot, it is helping companies grow, which is a different – more neutral – formulation, and focusing on what Rangan called acceleration. Acceleration happens with the help of agents that assist employees – thereby putting less emphasis on autonomy and more on efficiency.

The question is what these vendors customers' objectives are. Do they think outside-in or inside-out? Inside-out does imply a focus on own efficiency and reducing cost, as expressed by Klarna and discussed in this article by diginomica’s Stuart Lauchlan. Outside-in thinking would be the use of agents to improve their ability to serve their customers.

Will they use AI agents as a convenient tool to cut costs by cutting employees, something that also heavily depends on AI pricing? In other words, are AI agents the next iteration of outsourcing? Or will companies use agents to become better without laying off employees? Obviously, one scenario is more toxic than the other one.

In any case, I am very much looking forward to the “monster piece” that Jon Reed promises in his Oracle Cloud World AI agent article.

To close: let me ask two sets of questions for you to comment. There’s one for the vendor side and one for the buyer side:

Vendors: How do you convince your customers that your agent platform is more valuable if used to help the company become better instead of making it "leaner"? How do you educate them?

Buyers: What are your key objectives when implementing AI agents? What are the KPIs you use to determine success and how much do these KPIs need to change to achieve success?

The news

Dreamforce 2024 has (almost) started and the announcements are pouring in. Unsurprisingly, many of them are about AI, generative AI, Slack, and of course, agents. One of the major announcements that Salesforce made these days is about the release of Agentforce. According to Salesforce, Agentforce is ”a groundbreaking suite of autonomous AI agents that augment employees and handle tasks in service, sales, marketing, and commerce, driving unprecedented efficiency and customer satisfaction. Agentforce enables companies to scale their workforces on demand with a few clicks. Agentforce’s limitless digital workforce of AI agents can analyze data, make decisions, and take action on tasks like answering customer service inquiries, qualifying sales leads, and optimizing marketing campaigns. With Agentforce, any organization can easily build, customize, and deploy their own agents for any use case across any industry. The future of AI is agents, and it’s here.”

The platform is intended to bring chatbots to the next level by graduating them from co-pilots that “rely on human requests” to autonomously operating agents that retrieve the right data on demand, build action plans and execute these plans without intervention.

The bigger picture

According to Salesforce’s Trends in AI for CRM report, a staggering amount of 41 per cent of employee time is spent on low impact work. On top of this, 65 per cent of desk workers believe that generative AI will allow them to be more strategic. Salesforce also maintains that “every company has more jobs to be done than the resource available to do them.” Zendesk postulates that the number of interactions in customer service will grow by a factor of 5 in the near future and that of all resulting customer service interactions, 80 per cent will be resolved without human intervention.

At the same time, many a business complains about a shortage in skilled work.

So, autonomous agents are clearly a kind of holy grail for software vendors, as they in theory allow their customers to scale with increasing demand in an efficient manner. Correspondingly, the heat is on, and agent has become the hottest buzzword in the current generative AI hype.

My analysis and point of view

I want to use this announcement to place some fundamental stakes in the ground.

One is about the belief that autonomous agents are a silver bullet. Sorry to disappoint all drinkers of kool-aid: They are not.

The other one is about data. The perfect agent infrastructure will not deliver good results if the underlying data and the data used to train these systems, are flawed. For time being, this gets partly addressed by technologies like RAG and an increasing use of synthetic data, but this does not solve the issue of low-quality corporate data. This issue is also one of the root causes for the other one.

More on these points later.

Mark Benioff is bold as always by saying that Salesforce wants to “empower one billion agents with Agentforce by the end of 2025”. This is a little less than 1/8 of Earth’s current population. So, he either expects a lot of job attrition or he is talking about software agents. The latter might be a distinct possibility as, at least for the foreseeable future, software agents will be highly specialized. A deployment of meaningful size correspondingly requires a significant number of these agents.

However, the more interesting question is how autonomous these agents can be. Their quality, and hence the customer satisfaction that Salesforce rightfully mentions as a core objective, heavily depends on data. And this data is regularly of poor quality. That is a problem. A big problem. This data problem is not new and needs to get addressed before a meaningful deployment of agents – or co-pilots, for that matter – can happen. Yes, agents can take over the more mundane tasks and are steadily moving into the not-so-mundane areas. The better this data problem is addressed, the better human as well as artificial agents can work. Again rightfully, Salesforce maintains that this is one of the core reasons to have Agentforce sit on top of its Data Cloud and are using Salesforces AI Trust architecture.

Still, this is one of the core reasons why agents in the near future will not be as autonomous as one wants them to be. I wished that Salesforce would have addressed this issue a bit more actively. Customers achieving a 40 per cent increase in case resolution which outperforms their old bot is great news, but it in reality tells only half of the story.

Having said all this, I think that it is worthwhile for Salesforce customers to have a close look at Agentforce. However, I encourage them to do so with a clear plan of improving customer outcomes including customer experience, instead of looking at efficiency gains – aka laying off people. Not all of Salesforce’s customers will think like this. While there is no clear evidence that engaged employees lead to a higher customer satisfaction, it is quite obvious that empowering employees instead of antagonizing them and not creating anxiousness amongst staff reduces friction and therefore has a positive impact on productivity, if not necessarily customer experience. Still, improving customer outcomes is what improves business outcomes at the end of the day.

Salesforce could have strengthened this thinking by not choosing a consumption-based pricing but an outcome-based pricing, which is something that Zendesk recently introduced.

All in all, with Agentforce, Salesforce delivered a powerful tool to its customers who with the help of Salesforce and system integrators can use it in very meaningful ways.

It will be interesting to learn more about how Agentforce will be deployed with which objectives and which KPIs business will use to identify a successful implementation.

The News

On September 12, 2024, Zoho released a new, AI-rich version of Zoho Analytics that brings self-service BI to any persona in business.

The release added more than 100 features and now offers powerful AI and ML capabilities. These enable diagnostic insights, predictive analytics, and automatic report and dashboard generation. “Additional advancements to Zoho Analytics include a custom ML model-building studio, seamless integration with OpenAI, and third-party BI platform extensions. The new version of Zoho Analytics has added power, intelligence, and flexibility to serve a broader range of businesses and users than competitors in the market.”

The Zoho BI and Analytics Platform now offers more than 500 connectors to other systems, including streaming analytics.

Overall, the new release offers new capabilities across four main categories:

• Data Management Hub: Zoho Analytics has expanded its data management capabilities, ensuring more accurate and applicable decision-making and deeper insights to accelerate business success
• BI Infusion with Generative AI: Zoho Analytics has introduced Generative AI capabilities across the BI platform to accelerate the adoption of insights for a broad spectrum of user personas.
• Data Science and Machine Learning (DSML): Zoho Analytics now features the DSML Studio, allowing users to simply and quickly build custom machine learning models, which, for example, analyze or predict customer churn. n of insights for a broad spectrum of user personas.
• Platform Extensibility: Zoho Analytics is more deeply extensible, allowing businesses to sync and standardize data stored across multiple tools and platforms for comprehensive analysis and insights. Zoho Analytics is a composable platform on which any analytical solution can be built.

Alun Rafique, CEO and Co-Founder of Market Dojo says about the new Zoho BI and Analytics Platform that "as developers, Zoho Analytics 6.0's new AutoML capabilities caught our attention immediately. Our team is currently working on a model to analyze and predict customer churn, among other projects. Zia Insights' diagnostic capabilities have been a game- changer as well, helping us understand the underlying causes of a spike or drop in sales or performance so we can take quick action. Again, as solutions providers, it's critical we understand what's working and what's not within our own company before helping other businesses and professionals. Zoho Analytics 6.0 continues to give us that insight and more. Ask Zia' is another useful feature. With a simple prompt, I'm able to get the right insights I need instantly. The new chart types are amazing, and I'm particularly impressed with the racing and sunburst charts."

The bigger picture

The Zoho BI and Analytics Platform has its origins in 2009 as one of the first self-service cloud BI platforms. It can be installed behind firewalls locally or can be deployed on cloud platforms such as AWS, Google Cloud, and Microsoft Azure. Additionally, it is available as an embedded platform that can be leveraged by other vendors. Originally, it was aimed at line-of-business users and in the meantime has evolved into a full-blown BI platform with around 17,000 customers, as of August 2024. Within the Zoho One suite, it is used by 70,000 business on a daily basis. This makes Zoho Analytics the second-most used app, after Zoho CRM, of the Zoho One suite.

The Zoho BI and Analytics Platform covers the complete analytics workflow from data preparation through visualization and exploration. It addresses challenges that a broad range of users, from data scientists and developers to line-of-business users, have.

This new Zoho Analytics release features more than 100+ updates, including new visualizations, enhanced dashboard building, audit and admin controls, revamped mobile apps, right-to-left (RTL) support, and more.

The new version of the Zoho BI and Analytics Platform shall address five key challenges, namely:

• Data velocity and diversity
• Data management and -governance
• Complex analytical needs
• Limited adoption of analytics caused by diverse user needs
• Rapid technology changes.

Zoho addresses these challenges using the four different categories listed above.

1. Deepening the data and integration management using a powerful data management hub to establish a strong data and data management foundation
2. Infusing AI, particularly generative AI, across the platform to accelerate BI adoption
3. Increasing the ability to build machine learning models for analytics workloads by making this capability available to a broader user audience
4. Enhancing the platform’s extensibility to serve more use cases.

My point of view and analysis

As my colleagues Brian Sommer and David Smith said in an article by Lynn Greiner on CIO.com, the Zoho BI and analytics platform has some key strengths, including the deep integration into the whole Zoho suite of applications while simultaneously offering “some 500 integrations to other firms’ software products and data”, which makes it a very good platform for generating insight from disparage sources. In combination with Zoho’s own and the ability to leverage external AI, this can lead to strongly improved business analyses and forecasts.

With the help of decision intelligence that answers the very important question “why” and recommendation, immediate value can get created for business users.

Lastly, the strong support of very different user personas helps in making powerful analyses available throughout the enterprise.

You can download and read my full analysis on the Zoho BI and Analytics Platform directly from the Zoho site.

Disclosure: My report “One Platform to Help them All – The Zoho Analytics Platform” was commissioned by Zoho. Zoho did not and did not try to influence the analysis part of the report in any way.