The news

On June 23, 2025, Salesforce announced Agentforce 3, the third iteration of its Agentforce platform. Agentforce 3 is a major upgrade to Salesforce’s digital labor platform. It gives customers the visibility and control that is needed to scale AI agents that already have proven useful at many companies. The release covers several additional capabilities.

Salesforce has introduced the Agentforce Command Center, an observability console designed for managing AI agents. This tool allows businesses to track and scale AI agent activities. It is built into Agentforce Studio and includes features for monitoring performance metrics such as latency and error rates through live analytics.

The update also brings native support for the Model Context Protocol (MCP), which enables Agentforce to connect with any MCP-compliant server without requiring custom coding. The platform's Atlas architecture has been enhanced to improve latency, accuracy, and resiliency. Support for additional Large Language Models has been added, including Anthropic's Claude Sonnet via Amazon Bedrock, with future support planned for Google Gemini.

Agentforce 3 includes over 200 pre-built industry actions, with half of these being new additions. The agents are now capable of performing web searches and providing citations for the information they retrieve. The platform's availability has been expanded to Canada, the U.K., India, Japan, and Brazil, adding support for six new languages.

More than 30 new partners have been added to the AgentExchange marketplace. These include companies such as AWS, Box, Cisco, Google Cloud, IBM, PayPal, and Stripe. Salesforce has also introduced new pricing options, including per-user plans for its Sales, Service, and Industry Clouds that provide unlimited usage of actions for employee-facing agents.

The bigger picture

Agentic solutions are the biggest thing since sliced bread in the industry right now. They are even more overhyped than Metaverse and Blockchain have ever been. Still, the number of case studies and use cases that get published is rising fast. This is a sign of buyer interest, increasing adoption and ultimately shows that agentic AI is not a solution in search of a problem but can address actual business challenges. However, these solutions need to be implemented in a thoughtful and planned approach to harvest their benefits. This expressively also requires a hard look at pricing, which can easily dig deep into the gains achieved by productivity increases, increased efficiency, or other gains. The other hugely important topic is the change management that is required. The implementation of agentic AI has the smell of impending lay-offs. This is bad for employee morale and therefore a topic that needs to be actively addressed by vendors as well as by buyers.

My PoV and analysis

With this release, Agentforce has arrived in the Enterprise. The main missing link is the still emerging A2A support. After focusing on the agents themselves, followed by their development and deployment in the previous releases, this release is about agent management and scalability, including scaled development through the Salesforce ecosystem.

Without a good, if not complete, built-in observability, Agentforce was still more or less a tool for rapid prototyping of and experimentation with AI agents. This is not bad in itself, but not enough for an enterprise scale product. Still, it helped tremendously by getting mindshare and establishing Salesforce as a category leader, which deserves kudos. In puncto interoperability, SAP has beaten Salesforce to the punch with its Joule Everywhere announcement during the yearly Sapphire event. Interoperability and A2A capabilities will be the next battleground for platform domination in which the business applications behemoths, including Microsoft, find themselves.

What I find quite notable, even laudable, is how strong Salesforce (or SAP for that matter), emphasize upon agents cooperating with humans, both “sides” working on what they are doing best, instead of agents replacing humans. Unluckily, especially Salesforce sends mixed signals with CEO Marc Benioff making statements like “up to 50% of the work at his enterprise software company is now being done by AI” during an interview with Bloomberg. In the light of Salesforce (and other tech companies) laying off employees in the thousands or tens of thousands because of efficiency gains through the use of AI and AI agents, this is more than just a bit contradictory. This clearly suggests that the messaging should be cleared. Words must match actions, especially with a topic that is as sensitive as this and considering the huge visibility that Salesforce has.

Using Agentforce as customer zero certainly helps with proving that the product works and with gaining more credibility. This is especially in the light of the fast deployment of the system. As time to value is important, this proof point and the fast delivery of prebuilt industry specific agents plus the Agent marketplace are crucial for continued success in a market that moves extremely fast.

Last, but not least, I am looking forward to seeing two particular sets of functionalities. The first one is a true agent to agent workflow that includes external systems and agents, not only Salesforce provided ones. The second one is even more interesting in the light of the need to move fast. It is one thing to offer a Command Center that helps administrators and users to understand what is going on, what is more important is the ability to immediately suggest and generate solutions for identified problems. This is where the real efficiency gains can be harvested.

As usual, a good part of the announcement is for functionality that is still in pilot or beta stage, in particular the MCP andA2A support as well as the Command Center.  

Having said all this, Salesforce has done it again. The company evolved a solution to become enterprise ready from its first humble days during Dreamforce. Congratulations for this. And let’s see how the competition changes the playing field.

The News

On December 17, 2024, Salesforce announced Agentforce 2.0 after introducing Agentforce 1.0 during the company’s Dreamforce event. With it, it repositions Agentforce as a “Digital Labor Platform” that is capable of supplying businesses with an infinite workforce. This way, they shall be able to address internal challenges like labor shortages, fixed capacity, stalled productivity, or burnout. In addition, they increase their ability to work with increasing customer demands like no patience, their wish for personalization and empathy and with a knowledgeable expert, etc.

“Agentforce 2.0 is the newest version of Agentforce and the first digital labor platform for enterprises — a complete AI system for augmenting teams with trusted, autonomous AI agents in the flow of work. This new release introduces a new library of pre-built skills and workflow integrations for rapid customization, the ability to deploy AI agents in Slack, and advancements in agentic reasoning and retrieval augmented generation (RAG) - enabling teams to scale their workforce with a custom Agentforce capable of handling complex, multi-step tasks with even more precision and accuracy.”

Agentforce 2.0 comes with a library of prebuilt skills, which are jobs that digital agents can perform. These skills do not only cover Salesforce software but also cover partner software and, with the help of Mulesoft, any other system. The agent builder, that is part of Agentforce 2.0 enables the building of new agents/skills using natural language. New agents can be built using the skill library or custom logic.

Agentforce 2.0 is live and will get additional capabilities throughout Q1/2025.

The Bigger Picture

AI has hit an inflection point. We came from rule-based chat bots, added assistants and are now focusing on agency. This is possible because technology improved, the ability of grounding inferences into company data improved and furthermore, the ability of technology to show empathy by also employing what I would call world knowledge (after all, this is what an LLM is trained on) can be combined into responses to queries. Add to that the increasing desire of (business) users to not be forced into a machine-enforced user interface but to interact in a human way, and there is something that is very exciting. I am a technologist, after all.

At the same time, there is undoubtedly a need for more automation in business operations. Customer demands are increasing – shareholder demands are, too. So is the pressure on employees. Pretty much every senior manager (and their dog) is talking about a shortage of skilled labor, which is a notion that I do not fully buy into as there is also a still substantial unemployment. Given that, saying that there is a skill shortage is akin to saying that who is unemployed is unskilled or lazy. There are also many lay-off rounds which cause a drain of skills from companies. So, these people cannot be re-skilled as well?

To me, this dichotomy sounds to me like we are in the need of having some deep ethical discussions about the value of (human) labor vs. digital labor.

But back to Agentforce.

My PoV and Analysis

Agentforce 2.0 is BIG. Big not only, because it promises to deliver a whole new level of automation to businesses, but also because it is the way for Salesforce to extend its reach out of the “CRM” niche and to support the whole enterprise. Salesforce wants to achieve this by essentially making Agentforce the user interface to all business applications, not only Salesforce applications. The vehicle to achieve this is Slack, supported by MuleSoft.

At the same time, Salesforce boldly creates a new market category. No other significant company has taken a stance like this before.

This is a significant repositioning, or should I say pivot, including a direct attack on all other enterprise software vendors and I am now waiting to see their responses.

Salesforce itself uses Agentforce to power its customer service and to improve its customer service efficiency. According to its own data, this already now results in an automated handling of more than 80 per cent of all inquiries and into a reduction of escalations to humans by 50 per cent. Zendesk already during its Relate event last May predicted customer service automation rates of 80 per cent and above. Looks llike there is an increasing number of instances where this prediction has become reality.

What neither the article about the success of Agentforce in Salesforce customer service nor Mark Benioff said during the announcement is what happened to the human agents to which cases are no more escalated. Are they tasked with “more valuable” work or did they get replaced by Agentforce. I hope for the former, but my fear is the latter.

Agentforce, undoubtedly, hits a nerve. In the short time since Dreamforce, Salesforce signed up more than 1,000 customers of which several hundred are already live. I’d count this as a huge success although I’d love to hear more about the use cases that these customers deployed, what their success metrics are and how they fare along these metrics.

Let me put this into a bigger context here again. Whether skilled labor shortage is a perceived or a real problem, Agentforce allows businesses to focus on shareholder value by replacing humans with Agents. And I think that they will. The term “Digital Labor Platform” already strongly hints into this direction although Benioff repeatedly emphasized on “humans supported by agents”. Still, the conversation moves into the direction of agency fast. And, the vast majority of the announcement is about efficiency, I did not hear much about businesses being able to do entirely new things, offer entirely new services that they couldn’t offer before.

In summary, it is hard to overestimate the potential impact that a technology like Agentforce has, especially in the light of Agentforce being a work in progress. Why is this? Because Salesforce managed to combine a set of powerful technologies into a coherent whole.

Still, I’d really like to err on the side of caution and would like to see the business metrics and successes used by Agentforce customers.

Lastly, as said above, I see a high need for a strong ethics discussion. Agentforce invites businesses ta automate business processes instead of rethinking them. I strongly urge potential buyers to not fall into this trap. Low hanging fruit are just that: low hanging. Identify where the real value lies in processes and data, and apply AI from there. Remember the adage that a fool with a tool remains a fool.

Congrats to Salesforce for bringing a bold vision to life. It is now up to all of us to use this sensibly.

The News

Yesterday, December 3, 2024, Salesforce reported its Q3 / 2025 numbers.

The company reported a total GAAP revenue of $9.44B US, which is a year-over-year change of eight percent and also slightly above the guidance of up to $9.36B that it gave after Q2. In addition, Salesforce raised the lower end of its full year FY25 guidance by $100M to $37.8B. What is also visible is the continued focus on operative margin and operating cash flow, which are up by 280 basis points and 29 percent, respectively.

These numbers exceeded ‘market expectations’, with the corresponding impact on the Salesforce share price.

"We delivered another quarter of exceptional financial performance across revenue, margin, cash flow, and cRPO,” said Marc Benioff, Chair and CEO, Salesforce. “Agentforce, our complete AI system for enterprises built into the Salesforce Platform, is at the heart of a groundbreaking transformation. The rise of autonomous AI agents is revolutionizing global labor, reshaping how industries operate and scale. With Agentforce, we’re not just witnessing the future—we’re leading it, unleashing a new era of digital labor for every business and every industry."

“We continue to drive disciplined profitable growth with third quarter GAAP operating margin of 20.0%, up 280 basis points year-over-year, and non-GAAP operating margin of 33.1%, up 190 basis points year-over-year,” said Amy Weaver, President and CFO of Salesforce. “To date, our total capital returns have surpassed $20 billion and we remain focused on driving shareholder value.”

The Bigger Picture

Salesforce has managed to become the biggest vendor of enterprise software in only 25 years. This is nothing short of amazing, especially, when considering that this market has been changing and is continuing to change fast and that Salesforce started out as a player in a small niche of this market. Add to that who has been there already. Since its humble early days, the company pivoted a couple of times, from a single application company to a multi-application company, from best of breed to best of suite, to platform company, to AI company.

Which is straightforward, as AI, combined with conversational interfaces, are the current battleground of the enterprise software market. Companies that are able to establish themselves with these pivotal technologies, have a great chance at becoming quite sticky. It is one thing to migrate a database from one vendor’s system to another’s.

It is a totally different ballgame to migrate proprietary models that are ingrained in a software stack.

My Point of View and Analysis

In my eyes the quarter’s Salesforce numbers, in contrast to what Benioff boldly stated, are not exactly overwhelming. They are in line or below the growth figures of other major vendors.

 Yet, Salesforce is telling a very compelling story.

Salesforce is in another iteration of the clash of titans, this time it is not about dominance in the wider CRM market, now it is about AI and about the whole enterprise. It looks like Salesforce has found a way to make it into ‘core enterprise software land’ by virtue of its Data Cloud and Agentforce. It is basically attacking the vendors that support the full enterprise value chain via its data platform, its ecosystem, and AI.

The compelling story is the story of the rise of a new market of “digital labor” and, of course, Salesforce being the “largest supplier of digital labor” in this market. Personally, I find the moniker ‘digital labor’ dangerous as it, in contrast to contemporary wording, does not convey the message of ‘human augmented by machine’. It is a pivot away from it, albeit a straightforward one.

In fact, it is the story of Agentforce that allegedly fundamentally changes how businesses can operate. It is the story of AI agents doing a lot of the work that will not be done by humans anymore for various reasons; these reasons may include work being dull or dangerous, a lack of ‘skilled labor’ or mere linear thinking in terms of efficiency. According to Benioff, “productivity is no more tied to workforce growth but to this intelligent technology that can be scaled without limits”. The intelligent technology being Agentforce, that has turned into a platform, along with the whole company Salesforce. The problem with bold statements like this is that there are no real proof points yet. Agentforce is live only for a few weeks at the time of this writing. Yet, I agree that this market is “so much bigger than the data management market”. This is largely because the data management market is a prerequisite, a part of the foundation of the agent market – as evidenced by Salesforce’s failed attempt to acquire Informatica and the subsequent acquisition of Own.

While Benioff cites some success stories, these are still the same as the ones used during Dreamforce.

Yes, Agentforce works off the huge Salesforce databases, and yes, there are guardrails in place, but still … call me a pessimist, but as long as there are still too many incidents of language models, I would prefer if vendors like Salesforce openly addressed this topic along with possible ways to overcome it. I’d also like to see a few more implemented use cases along with their value.

Other vendors are working with less exuberant messaging. Maybe, this is part of Salesforce’s success.

Now, it is time to deliver.

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.