The News

Dreamforce 2023 has started. It is the first post-Covid physical Dreamforce. The event has more than 40,000 participants from all over the world, which is almost small, considering past events.

As usual, Dreamforce was opened by a keynote that was accompanied by a flurry of announcements. In an interesting twist, the keynote was accompanied by an analyst watch party held by Salesforce’s analyst relations. There was a “competing” watch party by the CRM Playaz.

Not surprisingly, the topic of AI was front, right and center of the keynote after some emphasis on a culture of giving and the celebration of Salesforce as the #3 software vendor worldwide with an expected revenue of $34.8B US while continuing to lead the CRM market by a considerable margin. This is, indeed, quite an achievement.

However, the focus of the keynote was set in the watch party by Salesforce’s Chief Enterprise Strategist Bruce Richardson with the following posting in the chat: “gen AI to boost global economy by $4.4 trillion. Since the release of ChatGPT in November 2022, it’s been all over the headlines, and businesses are racing to capture its value. Within the technology’s first few months, McKinsey research found that generative AI (gen AI) features stand to add up to $4.4 trillion to the global economy—annually. Source: McKinsey (August 23, 2023)”

Consequently, much of what Salesforce currently does, capitalizes on this opportunity and helps businesses work with this opportunity. The key vehicles for this are the new Einstein 1 Platform and Einstein Copilot that comes together with a low code development environment Einstein Copilot Studio. 

Einstein Copilot is an “out of the box conversational AI assistant built into the user experience of every Salesforce application”. The Einstein Copilot Studio is “an easy way for companies to build an entirely new generation of AI-powered apps with custom prompts, skills, and AI models”. 

Both are residing on top of Salesforce’s Data Cloud and are going into pilot state now.

The Einstein 1 Platform essentially fuses Salesforce’s Data Cloud and Einstein, based on a single meta data framework with the goal of enabling AI powered workflows that raise productivity.

The bigger picture

Everybody, and their dog, are talking generative AI these days (count me in). AI, as in automation, efficiency, and productivity, is an important, if not the most important topic these days. Many a company has made announcements mainly basing on and leveraging the huge buzz that OpenAI has managed to create. Few have delivered outside from some fairly simple to build scenarios, and most of them have concentrated on the same use cases. Nearly none of them speaks actively about, or embraces, the notions of privacy, security, and trust. Yet, most of the systems need extensive access to corporate, or even sensitive personal, data to lead to satisfactory results. I am yet to see a vendor that offers a dedicated AI security architecture, apart from Salesforce. Some may have one, but they fail to talk about it.

The second important topic these days is the increasing importance of conversational, i.e., conversational user interfaces. As I already wrote back in 2017: Voice an text will replace point and click. 

My analysis and point of view

From a topic point of view, Salesforce clearly gets it. This is clearly shown by the statement “entirely new generation of AI-powered apps with custom prompts, skills, and AI models”. The interaction with applications will shift more and more from a computer centric point of click metaphor to a human centric communication metaphor. Sure, visual user interfaces will continue to have their place – think dashboards or charts – but they will overall blend into a human centric interaction model. To be effective, this needs lots of machine learning that exposes knowledge and functionality via AI models. This will be the case for user interactions as well as interactions with external parties like customers, suppliers, partners, you name it.

The foundation for this is data. Clean data. Acknowledging that data will never be perfectly clean, basing it on an enterprise-wide harmonized meta data infrastructure is the way, also to address the topic of data islands. In my eyes, the underlying “database” of business applications will more and more become a data lake, means a lake house architecture for business applications. The challenge for Salesforce will be that the company is not the strong business software vendor that creates a metadata infrastructure. Instead, it needs to compete with the likes of Microsoft and SAP. Especially SAP is interesting here as – although seemingly smaller by annual revenue. SAP commands the as per yet most important system that a company has: the ERP. Around which system will companies build their metadata infrastructure, their ERP or their CRM? Related to this question: What will be the core platform? I guess that this is a question that is yet to be answered.

Benioff’s keynote was refreshing from an AI point of view. It was probably the first one this year that did not only rave about the fascinating prospects that generative AI “changes everything”. Yes, he said that, too, even added to it – it changes “everything and anything”.

However!

Salesforce also addresses the number one concern not only of customers but also of executives. Trust. During the keynote, Benioff and Salesforce did not only address the obvious business opportunity but also covered the very valid and real concern that was also an outcome of the recent State of the Connected Customer report.

External AI systems are a major concern for companies that use them. They do not know how the input data, including prompts is used, stored, and proliferated. They also do not know how accurate the responses of especially LLMs are.

In his keynote he said “LLMs are very convincing liars”. I think that is a very apt description.

Again, Salesforce clearly gets it.

Still, there is ways to go. There are many questions that are to be answered when it comes to the effective, efficient, and ethical usage of AI, as I have also written in my recent column article If AI can’t be trusted, efficacy and efficiency won’t matter on CustomerThink. Some major open topics are explainability and the question how, and using which data, the own models are trained. Salesforce delivers pretrained industry specific models yet states in its tenets of trusted, ethical and humane AI that “your data is not our product”. Pretrained models and this tenet are seemingly at odds with each other. This should be possible to be resolved easily. The topic of explainability is a hard one, but needs to be covered, too, as the sixth tenet rightfully says, “transparency builds trust”.

The news

During the Salesforce AI Day on June 12 as well as the Salesforce AI Industry Analyst Forum on June 20, Salesforce provided a lot of interesting information on how the company addresses the challenge – or should I say problem – of trust into artificial intelligence. Salesforce sees this gap caused by hallucinations, lack of context and data security as well as toxicity and bias. According to Salesforce, this gets compounded by the need for integrating external models into business software.

To address this problem, Salesforce has announced its AI Cloud that combines an “Einstein GPT Trust Layer”, Customer 360 and its CRM to offer AI-powered business processes that are built right into the system, based on an AI that can be trusted. The main vehicle is the Einstein GPT Trust Layer that takes care of

• secure data retrieval from business applications,

• dynamic grounding to reduce the risk of hallucinations and to increase response accuracy by automatically enriching prompts with relevant business-owned data,

• data masking, the anonymization of sensitive data to avoid its unintentional exposure of sensitive data to external tools,

• toxicity detection to make sure that generated content adheres to corporate policy, is free of unwanted words or images, and unbiased,

• creating and maintaining an audit trail,

• the external (or internal) AI not retaining, storing, any corporate information that gets sent to it via the request.

This trust layer sits in between the used AI models and the apps and the respective development environments. All requests to the models, along with their data, get routed through this layer, ensuring authorization protected retrieval of data, the grounding of prompts using it as well as data masking for anonymization. Responses by the models get routed through it as well. This enables an audit trail as well as toxicity detection. Models can be ones within Salesforce, ones developed and deployed by the customers in their infrastructure and third party models.

Figure 1 The Salesforce AI Cloud Architecture; source Salesforce

To round this off, Einstein Studio allows the building and deployment of models, their training using data within Salesforce and, at a later stage, the building of own models using a no-code environment.

The bigger picture

Although AI is not new, it is safe to say that generative AI is a game changer. OpenAI managed to get AI out of the realm of data scientists and into the hands of mere mortals. And most of us use business applications on a daily basis.

One of the most daunting problems of the use of AI is that there are a number of considerable risks involved with its usage. The one that is currently talked about most in the context of generative AI is the one of accuracy of responses to prompts, which is often referred to as hallucinations. This is not only problematic in consumer usage but even more so in business usage.

What comes on top in a business context is very much related to data privacy bias and profanity. Both have also been discussed in the consumer arena. Do you remember Microsoft’s infamous Tay bot? Or more recently of Samsung, Amazon, Apple, and other companies ordering their staff to not use ChatGPT et al.?

The management of all of these risks is of paramount importance to businesses, for regulatory reasons as well as for the need of protecting own intellectual property. No business can afford customer data and/or sensitive corporate data leak into external tools. This is doubly true in strongly regulated industries. But how to ensure this, when the models are not fully understood and when it is not even clear where and how data is stored? How to adhere in a GDPR-request to delete a customer’s data in this case? The management of these risks requires organizational, educational, and cultural measures in companies. These need to be supported or enforced with the help of technology. 

The obvious technical resolution for this is an AI security layer that I outline in my (upcoming, as of this writing) column article on CustomerThink as follows.

Figure 2 An AI security layer; source Thomas Wieberneit

This is of course simplified, not exactly trivial, but possible.

My analysis and point of view

One could, or rather should, say that trust and security are two of the most important assets in business. Customers need to trust that businesses do not collect an inordinate amount of data and that they furthermore use the data given by customers only for consented to purposes. In addition, they need to trust businesses that they keep their data safe. A multitude of regulations mandates this. Being trustworthy is even more important in times of AI as a service, when businesses cannot even tell anymore where customer data is stored, as it is learned by the AI and stored in a very decentralized manner – as part of an unknown number of parameters.

To enable this trustworthiness, what lies closer for a tier one platform vendor than ingraining an AI security layer directly into the own platform? The gateway to external services is already provided by the platform and can be reused by the AI security layer.

This is what Salesforce has done in an exemplary manner with the aptly named Einstein GPT Trust Layer. Kudos for this.

Figure 3 - How the Einstein GPT Trust Layer works; source Salesforce

In my opinion, the most interesting part is the zero-retention portion. Salesforce cannot guarantee on its own that external providers do not store any data. Whenever a prompt is sent to an external vendor, this data is leaving Salesforce’s systems boundaries. This means that external vendors assume temporary control of this data to provide their services. Masked or not, this data that can potentially be demasked, is handled by them.

To accommodate for this, Salesforce has established “zero-retention policies” with these vendors. According to information given during an analyst briefing, these policies ensure that the vendors won’t store any in-flight data, including inputs and outputs, nor won’t they use it for any purposes besides generating a response to the prompt.

This is quite an important statement that also indicates GDPR compliance, if “policy” can be translated to contract. On the other hand, this makes me curious how the refinement of prompts works in this case. Obviously, for highly security-oriented customers, this also suggests the preference of Salesforce or customer-owned models over external ones.

Overall, this is a great offering that addresses important concerns  of the C-suite.

The only qualm that I have is the price tag, which is quite steep, starting at currently $360,000 US. For sure, customers can derive good value out of it, this is not the problem. Where I see a challenge is that this is out of reach for most SMBs. I’d love to see an adaptation of this offering combined with offerings like Salesforce Easyor similar.

I wait to see when other vendors come forward with a comparable offering. Especially the other tier one but also the tier two vendors need to make a move now. Salesforce truly let the Genie out of the bottle and put them in a tight spot.

Kudos again!

In the course of the last year or three, I have written a few times about the supply chain and how it affects the customer experience, sometimes badly so. In my last article ‘the impact of the supply chain on the customer experience’, I explained some of the issues and laid out a couple of high level solutions.

Let me elaborate a little on what I wrote then. Of course, there are challenges in two main areas:

• The demand side
• The supply side

The customer expectations are quite simple: A customer expects that things just work, that the vendor, and information given by the vendor, is reliable, accurate and comes timely; that deliveries after order are coming reasonably fast and that there is transparency about the order and delivery status.

This is all quite abstract, so let us fill this with life.

An example to the contrary goes as follows: You buy something, e.g., some whiteware, from a brand with a very high reputation; you are asked for upfront payment (several thousand dollar, remember, high reputation), get an approximate delivery date two months out – along with a warning that there may be a delay of a month or so.

So far, so good. I could live with that although even the two months should be explained …

Now, imagine that one month into the wait time you get an e-mail notifying you about a changed delivery date, pushing it out by a month. Well, not nice, but also not out of the range of expectations. After all you got told so.

However, the next month you get a similar mail. And the next month, and then in some more months.

This is where the customer experience falls apart. What do you do? You call the vendor. Customer service tells you that they do not know when the delivery can happen, and don’t know a reason other than ‘chip shortage’. It is all up there in the clouds (pun not intended).

This is akin to saying: I do not have any insight into my supply chain.

What? How can this happen?

This is quite unhealthy for a business.

Not having insight into the own supply chain can happen generally only under two circumstances

1. Data is not shared internally
2. The suppliers do not provide any data

The first reason can be mitigated pretty easily. It is a matter of internal collaboration – i.e., processes – and then of setting up a proper data exchange between the relevant systems. That way, all organizations have the data at hand that they need to at least fulfill some functions that lie at the base of the pyramid of customer expectations. Still, the experience is poor, as the customer does not only expect timely and accurate information, but actually a delivery at the agreed upon date and time.

The second challenge is not overcome that easily. As I have written earlier, overcoming it also involves a fundamental strategic change. A change away from optimizing the supply chain using minimizing cost as the sole KPI, moving towards a balance between cost, reliability, and speed, therefore addressing the core customer demands at the base of the pyramid in more situations. As we have re-learned in the past two years, a sole focus on cost makes supply chains inherently unstable.

Why re-learned? Because supply chain disruptions are not a new phenomenon. We have seen supply chain disruptions time and again, albeit usually not on a global scale like now. Reasons for these disruptions can range from traffic jams, strikes, vendor-supplier disputes, outages or accidents, up to natural disasters, pandemics or wars.

A second part of this necessary strategic change also involves more and earlier sharing of information with suppliers, so that they can improve their own planning. The other way round, this is true, too: Suppliers need to be willing to share relevant information about their capacities and/or load situation, too.

The third party of the chain are the transportation agents that offer air, ground or sea transportation.

Of course, this also involves a lot of trust and the assurance that this information is not abused in a power play of any kind, or worse, but used for the sole purpose of serving the end customer better by cooperating.

But just to be clear: Not all corners of the triangle cost, speed, reliability can be improved at the same time. Working this triangle is not a silver bullet but will help reducing the likelihood of not being able to deliver.

So, what data needs to be shared, and how, so that the supply chain can run at its optimum at any given time?

There is data on two levels

• Macro
• Micro

On macro level, demand signals need to be made available to the suppliers, so they can plan production (and their own sales) early. Conversely, available capacity data needs to be available to buyers, so they have a chance to distribute their orders amongst a number of suppliers, as needed. At the same time, transportation capacities need to be secured, too. This is, where the transportation agents come into picture. It doesn’t help, if buyer and supplier get into an agreement and the goods cannot get transported.

Again, this requires a lot of mutual trust that the supply can be produced in time and quality, that payments are made in time, and especially, that the provided data is not or cannot be abused. Alternatively, it requires an infrastructure that enables the participants to trust in the infrastructure instead.

On the micro level, it needs the real-time exchange of detail information that enables a just-in-time supply that allows for detailed scheduling and synchronizing work processes and to possibly prepare for contingencies. This may include IoT data out of the network but also external data like weather, traffic, and lots more.

This can easily become an extremely high volume of data that needs to be processed fast to be useful.

All of this might be a good use case for a system that is based on a blockchain that is provided by an independent party or, even better, by a party that all participants have a stake in (the more the better). The blockchain could be used to secure the promises (demands and available capacities).

On the other hand, there are still significant voices (e.g. Bruce Schneier in his readable post on the dangers of cryptocurrencies and the uselessness of blockchain) that ask where the ultimate benefit of a blockchain is, whether it is a solution in search of a problem or whether it solves existing challenges in just another (not necessarily better) way or even question the promise of blockchain security altogether (e.g. this DARPA analysis).

In any case, every scalable transactional system that allows for efficient and automated negotiation and processing should be capable of delivering these services. After all, the juice lies in the business logic and not in contracts being smart contracts or not.

So, what does it take to address the supply chain problem, given businesses have developed a strategy that focuses on a balance between cost, reliability, and speed?

1. A secure (technical) environment that makes it easy for the participants to trust each other or that makes it even possible to replace trust into each other by trust into the environment. This environment must also ensure that the individual participants are not put at a systematic advantage or disadvantage. This requires a strong governance and is in all likelihood implemented by a network of systems that interact based on a standardized protocol.
2. An ecosystem with enough participants on demand, supply and logistics sides that allows for both, the efficiency of scale and the possibility to build flexible networks between suppliers and buyers instead of relying on tight chains. As said above, the participants must not be put at a systematic advantage or disadvantage.
3. Enough and accurate data that is securely collected, managed, bound by consent and purpose; it is to be only used to fulfill the needed services. This data must be equally accessible and usable by all participants (for this purpose).
4. Scalable and transactional and analytical systems that enable the necessary real-time data ingestion and transaction handling. These systems predict demands and/or shortages on macro level and might also help in mitigating micro-level disturbances.

But as said: Trust between all members of the supply chain, from the end customer through the complete value network is key to the customers having a great experience.

Supply chains must evolve into flexible supply networks, based on mutual trust.

The enterprise software world is one where few, if any, companies can stand alone. Cloud computing has greatly increased the availability and usability of business applications. However, no vendor can claim to supply a complete homegrown solution. The SaaS world is one of partnerships. Partnerships are best when all the partners are trustworthy.

Business software customers still need to pick a brand and stick with it for a while. That brand—the platform—represents a serious investment of time and trust. Sure, you own your data, and changing platforms is supposed to be easy. But ask anybody who’s done it whether they would want to do so again next year.

Platform customers have to weigh more than just the price and features of each platform provider. They must also consider what the future holds for that platform and its partners. None of the big platforms are likely to disappear or merge anytime soon, but their fates and fortunes are not constant. They also have histories and reputations. Increasingly, they have prominent social and/or political stances as well. Those issues matter enough to some business owners and investors to affect their decisions. There are also simpler matters, like whether the customer likes the feel of the software environment, or has concerns about possible changes.

Can I Trust You?

In the end, these decisions boil down to variations on a single question: “Can I trust you with my business?” Contracts and commercial laws deal with businesses that act in bad faith; this is more about responsibility, reliability, and safety. Examples:

• Will you continue to support the applications I use?
• Is my data secure?
• Do you have good disaster plans in place?
• How will your price structure evolve?
• Will you expand your offerings?
• How does the app marketplace work for developer partners?

Platforms give their users clear answers to these questions. The provider creates the rules for the platform, and any partner or developer that wants to associate with it has to stick to those rules. Thus, a customer can trust any vendor on that platform to adhere to its standards, or face consequences. Broader kinds of partnerships provide some protections, but the customer is still essentially shopping at a bunch of different stores rather than the platform’s one superstore.

App vendors who rely on a platform for distribution are probably among the most trustworthy businesses in all of enterprise software. They answer to their customers, their platform, and their own standards of excellence, all at once.

One good example is Fastcall, a CTI app developer on the Salesforce AppExchange. “We serve Salesforce customers and only Salesforce customers, so everything we do is focused on earning and keeping their trust,” says Richard Rosen, co-founder of Fastcall. As a dedicated AppExchange developer, Fastcall can concentrate fully on adhering to Salesforce’s standards and practices regarding security, partner policy, and app compatibility. “There is no other market for us, the AppExchange is our home,” Rosen says. “It’s just good business to be as easy to work with as possible.”

• First of all: They must not be part of the problem.
• They need a relentless focus on solving real business problems with technology.
• They need to deliver easy and simple to implement solutions to these business problems.

Don’t be a part of the problem

1. Normally, the cost comes before the value can get harvested. This was true in the old days of on premise software and is also true in a subscription economy. The fee is to be paid. And consulting, e.g. the implementation of business software, is usually paid by the hour.

2. Especially after a crisis the clients may not be able to invest enough to be able to get significant value out of the investment. This even happens at the best of times. How many projects needed to get stripped down so much that software users feel the urge to circumvent the newly implemented processes? It happens all the time.

Focus on solving real business problems

Deliver easy and simple to implement solutions to real business problems

Call to Action

A Focus on Customer Experience Guides on the Way

• Treat your employees well
• Offer high-quality products/services
• Listen to your customers
• Pay your fair share of taxes
• Engage into ethical business best practices

Treat Your Employees Well

Offer High-Quality Products and Services

Listen To Your Customers

Pay Your Fair Share of Taxes

Engage Into Ethical Business Best Practices

My Take