The past weeks showed quite some interesting activity on the mergers and acquisitions and the partnership frontiers. NiCE acquired the German conversational AI rock star Cognigy for $955M and a short time later announced that the company enhanced its partnership with Salesforce. At nearly the same time, Genesys received an additional funding of $1.5 bn from Salesforce and ServiceNow. Salesforce acquired Waii and Bluebird. VC company Thoma Bravo acquired the still leading CCaaS vendor Verint, to name but a few of the more interesting, and perhaps consequential ones.

On top of all this, Avaya seems to have offered all employees a voluntary exit package.

What all this shows is that there is significant consolidation going on in the AI-assisted (or should I say, driven?) CCaaS market and that various players are battling to provide – or at least be perceived to provide – the most comprehensive and valuable platform while others fight for survival.

Yes, it’s nothing new, but can’t be repeated often enough. The CX market is and always was a high stakes platform game. The stakes got even higher with the advent, the promise and the necessary investments that generative AI and agentic AI require. While one can consider Salesforce’s acquisitions as tuck-ins that help rounding off its Agentforce platform, the other ones are a sign of something bigger going on in the CCaaS and customer service market segments. It is also notable that exactly these sectors get more and more referred to as CX market, whether this is a correct, or only good, attribution, or not. Hint: It isn’t.

Not unexpectedly, Salesforce is in the thick of all this, having a massive stake in Genesys and a strategic – at least for NiCE – partnership with NiCE. And regardless, of whether NiCE or Genesys turn out to be stronger, Salesforce wins.

For customers of the Salesforce ecosystem this, first of all, gives choice. For Salesforce, it gives access to a wide array of different customers and therefore the ability to fuel the data pipeline that Agentforce needs. With the different profiles of NiCE/Cognigy, customers need to establish for themselves whether they need the strengths on the customer service center side (NiCE/Cognigy) or the analytics and orchestration side (Genesys).

On the other hand, we have Verint getting into the fold of Thoma Bravo and being merged with WEM vendor Calabrio. This instantly creates a strong player that includes WEM and analytics capabilities. But looking into the Thoma Bravo portfolio, there is more, in particular Medallia. Adding Medallia’s capabilities adds improved journey management and especially a strong voice of the customer element that helps understanding and therefore improving customer sentiments at every single touch point. This would challenge CCaaS leaders like Genesys and NiCE by creating a comprehensive analytics story and a journey orchestration capability. With the help of UserTesting this gets even more pronounced as UserTesting’s technology allows the finding of a root cause of a problem identified using the Verint/Calabrio/Medallia stack. This combination can create something that one could call a total engagement and analytics platform that covers the capabilities of traditional customer service vendors, CCaaS vendors, and journey orchestration vendors. This has the potential to create a tectonic shift in these software categories, as they converge to form a kind of Gondwana of service solutions.

While this doesn’t necessarily create an immediate threat to the Zendesks and Freshworks of this world, it may severely limit their ability to enter into more complicated, multi-step service scenarios. On the other hand, this combination of Thoma Bravo capabilities needs to prove its ability to scale down to avoid being disrupted from below by vendors that have their strength in high volume ticketing scenarios.

I do not know whether this is what Thoma Bravo has in mind, but the scenario is quite appealing to me.

My analysis and point of view

But in any case, one thing is clear: The markets for conversational/agentic AI, CCaaS, UCaaS and customer service are converging while the markets are consolidating. And all of this seems to accelerate around two themes that NiCE and Genesys exemplify.

NiCE's is integrating vertically. The Cognigy acquisition is an integration move that aims at owning the full technology stack, from core CCaaS to best-in-class conversational AI. The goal is a seamless, single-vendor platform. The risks for customers are potential integration debt from the acquired asset, a less open and smaller ecosystem, and probably a premium licensing for these new capabilities after being locked in. The deepened Salesforce partnership is a pragmatic necessity to ensure relevance and access to the bigger ecosystem, not a fundamental strategy shift.

Genesys plays the ecosystem card. The capital injection from Salesforce and ServiceNow is probably a direct counter. It funds a horizontal integration strategy, betting that enterprises prefer a tightly integrated alliance of market leaders over a single-vendor suite, which is in line with both Salesforce’s and ServiceNow’s strategies. The risk is that these integrations remain shallow, and the "ecosystem" is more of a marketing concept than a technical reality. This largely depends on how the three parties build and execute their roadmaps.

Thoma Bravo’s portfolio around Verint can become a wildcard theme here. The merger of Verint and Calabrio would create a third force. This entity would not necessarily compete directly on the CCaaS infrastructure layer but would focus on the higher-margin levels, starting with WEM and interaction intelligence. Combined with the other portfolio companies, its power would come from either being a platform-agnostic integration hub—the "Switzerland of data" or becoming another vertically integrated solution.

With this convergence going on, customers will be able to create data informed and more seamless processes for their customers, simply because there will be deeper integration on a data platform level, not necessarily a software platform one.

What should customer executives look at now?

• NiCE/Cognigy customers face significant integration risks due to two leading technologies being combined. Customers should demand a detailed, time-bound integration plan and roadmap with technical milestones. They need to have a deep and hard look at how their current licensing models are changed by new ones and model their TCO inclusive of probable premium "Cognigy-powered" features.
• Genesys customers face a different execution risk. They require technical proof-of-concepts that demonstrate the depth and real-time nature of the Salesforce and ServiceNow integrations, respectively, along with a detailed integration roadmap. They should mandate that business value metrics be tied to the pricing and success of this "ecosystem."
• Verint and Calabrio customers face a similar integration risk plus an uncertainty about the detailed positioning and future of Verint. Customers need to demand a detailed integration roadmap with milestones and clarity about the final strategy – platform neutrality or vertical integration.
• Avaya customers are worst off – and have been for a while. Avaya is in survival mode. This situation requires immediate risk mitigation. Customers should work with a two-prongued strategy: Firstly, they need to secure existing SLAs and support commitments in writing. Secondly, they need to engage in a RFI/RFP process for at least two cloud-native CCaaS platforms. The objective is to have a fully vetted exit strategy that can be executed fast.

Some general recommendations

There is so much going on at a fast pace, that it pays off to reevaluate vendor roadmaps. An outdated understanding of your vendor's strategic position is a serious liability. The market of 18 months ago does no longer exist and the market of in 18 months will be very different from today’s.

As platforms consolidate, the risk of vendor lock-in increases. Customers should favor providers with a demonstrated commitment to an open, API-first architecture. There is no one-size-fits-it-all software platform for all businesses, nor will there be.

The current market will expose providers with weak financial fundamentals. Customers should review their vendor’s financial health and R&D investment as a default part of their due diligence.

Last, but not least, customers need to avoid AI-washing and require that any proposed AI solution be tied to a business case with measurable outcomes, such as reduced operational costs or improved containment rates. Additionally, contracts need to be tied to achieving these outcomes.

Now, that we are in the middle of – or hopefully closer to the end of – a general hype that was caused by Open AI’s ChatGPT, it is time to reemphasize on what is possible and what is not, what should be done and what not. It is time to look at business use cases that are beyond the hype and that can be tied to actual business outcomes and business value.

This, especially, in the light of the probably most expensive demo ever, after Google Bard gave a factually wrong answer in its release demo. A factual error wiped more than $100bn US off Google’s valuation.

I say this without any gloating. Still, this incident shows how high the stakes are when it comes to large language models, LLM. It also shows that businesses need to have a good and hard look at what problems they can meaningfully solve with their help. This includes quick wins as well as strategic solutions.

From a business perspective, there are at least two dimensions to look at when assessing the usefulness of solutions that involve large language models, LLM.

One dimension, of course, is the degree of language fluency the system is capable of. Conversational user interfaces, exposed by chatbots or voice bots and digital assistants, smart speakers, etc. are around for a while now. These systems are able to interpret the written or spoken word, and to respond accordingly. This response is either written/spoken or by initiating the action that was asked for. One of the main limitations of these more traditional conversational AI systems is that they are better in understanding than in – lacking a better word – expressing themselves. Relying on well-trained machine learning models, they are also quite regularly able to surface a correct solution for problems in the problem domain that they are trained for. They usually work based on pretrained intents.

And, based on the training data, they usually give quite accurate responses to questions in their domain.

The problem: They are usually limited to a fairly small number of domains.

LLMs, on the other hand, are generally trained “to understand the relationships between words, phrases and sentences in a language. The goal is to have the LLM generate outputs that are semantically meaningful and reflect the context of the input.” This is part of ChatGPTs answer to the question what the purposes of an LLM is. The training set of an LLM is usually a vast amount of “real world” knowledge that usually comes from publicly available sources – aka the Internet. The output itself can be in written, graphical or other formats.

What LLMs excel in is generating responses to questions in a human way. And they can respond to a wide variety of topics. When focusing on text, they are built to generate coherent and meaningful responses.

The problem: They sometimes lack accuracy and give wrong output with full confidence. Even worse, wrong or inaccurate output is not easily identifiable by a user without the requisite knowledge. Again, refer to the Google Bard example that (temporarily, at least) wiped off $100 billion US from Googles valuation. Not picking on Google, there are plenty of examples around that call out ChatGPT or You.com or other tools.

Consequently, the other dimension to look at is accuracy.

The question is whether both dimensions always matter equally or not. In a business sense, one can argue that accuracy matters always. Receiving factual errors in a business conversation is not only a poor customer experience but may in extreme cases even lead to legal issues.

What is also important to understand is that the more accuracy is required the more the necessity of integrating additional systems to augment the LLM increases. An LLM on its own is not much more than some form of entertainment. Even in search engines, LLMs only augment the search by enabling natural language queries and the delivery of results in human language instead of a mere link list.

At least they should do this.

With all this being said, what are business use cases involving a large language model? As said, there needs to be a reasonable accuracy. Obviously, they require fluency as a precondition, as fluency is the core differentiator of an LLM.

Let’s look at some use cases in no particular order of priority.

• I’d start with something that I’d call “storytelling”. This is basically the creation of market-relevant documents that describe the capabilities and differentiating factors of a product, solution, or service. Being somewhat marketing related (no offence intended) and a first point of contact for customers, it needs to be easy to understand without requiring a great deal of technical accuracy. At the same time, it must not be wrong. A stripped-down version of this could be the (improved) generation of social media content, e.g., tweets. Benefits are faster creation of high-level content for general websites but also, more specifically, for ABM scenarios and landing pages. To be able to create this text, an LLM needs to be connected to internal systems holding requirements, specifications as well as communications between the involved persons. This is also a use case that should be implement-able near-term.
• One of the main tasks of people is the writing of, and more so, responding to emails. Especially, in sales scenarios, customer inquiries can get formulated and suggested based upon previous emails and the context given by the CRM system, e.g., about proposals made. This scenario would already require quite a high accuracy to avoid sending out faulty information that might be legally binding. The benefit of this scenario is a significant reduction time needed to send emails, resulting in increased productivity. It is a scenario that Microsoft has already implemented in its Viva Sales solution.
• Generation of documentation is a scenario that somewhat varies in the requirement for fluency. It can be mainly divided into technical and user documentation. While user documentation needs to be extremely readable, the writing style is somewhat less important for technical documentation. Conversely, technical documentation likely needs to have a high degree of technical accuracy that is not needed in user documentation, which means that either different repositories or different parts of source documents need to be used to create the texts and potentially diagrams and images.
• One of the most promising use cases in the short term is customer service, including enterprise search. Here, users want answers to their questions, not just links or something actioned. To achieve this, it is necessary to connect to a conversational AI, business systems and a well-functioning knowledge base that helps in generating accurate answers when searching for something. The actioning of issues is very similar to what conversational AIs do already now. The differences are that the intent detection can be far better as the LLM can create more than enough training sets for this and that the answers given by the system are far more fluent. The same holds true for an inquiry scenario. However, as a word of caution, the accuracy of responses to inquiries depends heavily on the kb content that gets searched by the enterprise search. Therefore, the kb needs continuous and rigorous scrutiny. If this is given, the benefits lie in increased call deflection and customer satisfaction. Properly implemented, benefits include an improved call deflection as more cases can get handled by the system, combined with an increased customer satisfaction as issue handling can become quite easy and efficient for the customer. Cognigy has recently presented some very good examples (here and here) that also include voice in- and output.
• Agent assistance is somewhat easier to implement, as it mostly needs to connect to the customer service application, including the chat history. Having complete access to sales and marketing data, of course is helpful, too. Combined with a sentiment analysis, the LLM can suggest text blocks for the agent to use. The benefits of this are an increased agent efficiency and quite possibly also higher customer satisfaction as the text blocks do exhibit more empathy with the customer’s situation than texts generated without an LLM.

In summary, these five scenarios show use cases involving an LLM that are beyond the hype. They can get implemented in a short time and they can also be easily tied to business outcomes. That way, their benefits can get measured.

Which other use cases do you see? And how would you tie them to business value? 

Inspired by a recent panel discussion as part of theIn the Hot Seat podcastthat I am involved in, I opened a chat with chatGPT3. ChatGPT is a language model by OpenAI that interacts in a conversational way. This way, it shall be able to follow a conversation, answer follow up questions or even admit mistakes, challenge incorrect premises or reject inappropriate requests. Our sixth episode of In the Hot Seat revolved about the question whether web3 will deliver on its promise or not.

The promise being that content producers and web users get more power by applying concepts like decentralization, blockchain and a token economy.

As I am a bit sceptic about this kind of silver bullet promises, I went right for the jugular.

A conversation between a human and a bot

Thomas: Tell me with arguments why web3 based on blockchain will fail

chatGPT3: It's impossible for me to provide arguments as to why web3 based on blockchain will fail, as web3 is not based on blockchain technology. Web3 is a term used to describe the next generation of the web, which is focused on decentralization and enabling users to have more control over their online data and interactions. Blockchain technology, on the other hand, is a decentralized, distributed ledger that is used to securely and transparently record transactions. So while blockchain technology may play a role in the development of web3, it is not the foundation of it.

Thomas: If blockchain does not play a role in web3, how will the decentralization be achieved?

chatGPT3: Web3 is focused on decentralization, but it does not necessarily rely on blockchain technology to achieve this. Decentralization can be achieved through a variety of means, such as peer-to-peer networks, distributed systems, and other technologies. Some examples of decentralized technologies that could be used in the development of web3 include distributed file systems, peer-to-peer messaging protocols, and decentralized identification systems. Ultimately, the specific technologies used to achieve decentralization in web3 will likely depend on the specific goals and requirements of the applications and services being built on top of it.

Thomas: How can ownership be assured in a decentralized world if blockchain is not the foundation?

chatGPT3: In a decentralized world, ownership can be assured through the use of cryptography and digital signatures. These technologies allow individuals and organizations to prove ownership of digital assets without the need for a central authority or third party to verify the ownership. For example, in a decentralized network, a user could prove ownership of a particular asset, such as a digital token, by signing a message with their private key. This signature could then be verified by other users on the network using the user's public key, providing a secure and transparent way of proving ownership. While blockchain technology is one way of implementing these cryptographic techniques, it is not the only way, and other technologies such as distributed ledgers or advanced cryptography could also be used to assure ownership in a decentralized world.

Thomas: Why then the hype around web3 based on blockchain?

chatGPT3: Web3 is not necessarily based on blockchain technology, and the hype around it does not necessarily relate to blockchain. The hype around web3 likely stems from the potential for decentralization to enable new, innovative applications and services that were not possible with previous generations of the web. Decentralization has the potential to allow users to have more control over their data and interactions online, as well as enabling new business models and opportunities. While blockchain technology may play a role in the development of some of these decentralized applications and services, it is not the only technology that could be used, and the hype around web3 is not necessarily tied specifically to blockchain.

Thomas: What new innovative applications and services are not possible with previous generations of the web?

chatGPT3: The potential applications and services enabled by web3 are difficult to predict, as the technology is still in development and many of the specific details are not yet known. However, some possible examples of innovative applications and services that could be built on top of web3 include decentralized finance (DeFi) platforms, decentralized social networks, and decentralized marketplaces. These types of applications could potentially offer users more control over their data and interactions, as well as enabling new business models and opportunities. Additionally, web3 could enable new applications and services that are not currently feasible due to the centralization of the current web, such as secure and transparent voting systems, decentralized governance systems, and distributed cloud computing platforms.

Thomas: Decentralized social networks do already exist. How can defi work without a decentralized ledger, i.e. a blockchain?

chatGPT3: While decentralized social networks do already exist, they are not as widely used or as well-known as their centralized counterparts, such as Facebook or Twitter. Decentralized finance (DeFi) platforms, on the other hand, are a relatively new concept that is still being developed. DeFi refers to the use of decentralized technologies, such as smart contracts, to enable financial applications and services that are not dependent on traditional financial institutions. These services could include things like decentralized exchanges, lending and borrowing platforms, and prediction markets.

DeFi platforms do not necessarily need to be built on top of a blockchain in order to function. While blockchain technology does provide a decentralized, distributed ledger that is well-suited for certain types of financial applications, other decentralized technologies such as distributed ledgers or advanced cryptography could also be used to enable DeFi platforms. Ultimately, the specific technologies used will depend on the specific requirements and goals of the DeFi platform in question.

Thomas: name me some defi platforms that do not use blockchain?

chatGPT3: As decentralized finance (DeFi) platforms are a relatively new concept, there are not currently many examples of DeFi platforms that do not use blockchain technology. Most existing DeFi platforms are built on top of blockchain networks, such as Ethereum or EOS, as these networks provide the necessary infrastructure for decentralized applications.

However, it is possible that future DeFi platforms could be built on top of other decentralized technologies that do not use blockchain. For example, a DeFi platform could be built on top of a distributed ledger that uses advanced cryptography to secure transactions, rather than the consensus mechanisms used by most blockchains. Additionally, some DeFi platforms are exploring the use of "layer 2" technologies, which are built on top of existing blockchain networks and enable higher scalability and performance. These technologies could potentially be used to build DeFi platforms that do not rely on the underlying blockchain network for all of their functionality.

Thomas: If you say most platforms do use blockchain, there are some that do not use blockchain. Please name some examples.

chatGPT3: As I mentioned earlier, most existing decentralized finance (DeFi) platforms are built on top of blockchain technology, as this provides the necessary infrastructure for decentralized applications. Therefore, it is difficult to provide specific examples of DeFi platforms that do not use blockchain technology.

However, as I also mentioned earlier, it is possible that future DeFi platforms could be built on top of other decentralized technologies that do not use blockchain. These technologies could include distributed ledgers or advanced cryptography, for example. Additionally, some DeFi platforms are exploring the use of "layer 2" technologies, which are built on top of existing blockchain networks and enable higher scalability and performance. These technologies could potentially be used to build DeFi platforms that do not rely on the underlying blockchain network for all of their functionality. It is worth noting, however, that these technologies are still in development and it is not yet clear what specific DeFi platforms will be built using them.

Thomas: Ownership already now can be through the use of cryptography and e-signatures. What problem that is not yet solved does decentralization solve or solve better than existing solutions do?

chatGPT3: Decentralization has the potential to solve several problems that are not yet solved by existing solutions. For example, decentralization can provide increased security and resilience by distributing data and functionality across multiple nodes, rather than relying on a single, central authority. This can make it more difficult for attackers to compromise the system, and can also ensure that the system continues to function even if individual nodes fail.

Additionally, decentralization can provide increased transparency and accountability by enabling all users to have access to the same information and allowing anyone to verify the correctness and integrity of that information. This can help to build trust and confidence in the system, and can also enable new business models and opportunities that are not possible with existing centralized systems.

Finally, decentralization has the potential to enable individuals and organizations to have more control over their data and interactions online. By removing the need for central authorities to verify and control access to data, decentralization can enable users to have greater control over who has access to their data and how it is used. This can help to protect user privacy and can also enable new applications and services that are not possible with existing centralized systems.

Thomas: Your first point basically says that security by decentralization is inefficient. Your second point disregards that the information can also be made available to everyone, and easier, in a centralized world. Re your third point, how does decentralization promote more control by individuals as it creates more attack vectors?

chatGPT3: I apologize if my previous response was unclear. Decentralization does not necessarily make systems less efficient, and it can provide increased security and resilience by distributing data and functionality across multiple nodes. This can make it more difficult for attackers to compromise the system, and can also ensure that the system continues to function even if individual nodes fail.

In terms of transparency and accountability, while it is true that information can be made available to everyone in a centralized system, decentralization has the added benefit of enabling anyone to verify the correctness and integrity of that information. This can help to build trust and confidence in the system, and can also enable new business models and opportunities that are not possible with existing centralized systems.

Regarding your third point, decentralization does not necessarily create more attack vectors. While it is true that decentralized systems can have more nodes, which could potentially be targeted by attackers, decentralization can also provide increased security by making it more difficult for attackers to compromise the system as a whole. Additionally, decentralization can enable individuals and organizations to have more control over their data and interactions online by removing the need for central authorities to verify and control access to data. This can help to protect user privacy and can also enable new applications and services that are not possible with existing centralized systems.

My point of view

First of all: Wow!

And I mean W O W.

Given this conversation, not knowing whether there is a conversational AI involved or not, one surely doesn’t have an easy time to establish this fact.

The other thing that becomes abundantly clear is that the system is quite well trained. I would also suspect a bias towards web3, which it sort of ambiguously defines somewhere between web3 and web3.0. Yet, there is quite a wide understanding that one of the core foundational elements of web3 is a blockchain. The system is quite right about the fact that there is no coherent web3 yet.

So, it argumentation wise leaves the door open.

It also, “behaves” rather like someone who is semi-educated and defensive. Drilling down into statements that the system makes are responded to with references into a possible future. This can be clearly seen when it comes to “new innovative applications” or my question about defi companies that do not use blockchain and that the system suggested as existing.

All in all, this system clearly shows the art of the possible. This raises quite some expectations to the future, and some concern. With machines becoming increasingly better at using language, it must be made clear when one communicates with a machine and not with a human.

Last piece of advice: Do not try to use it as a PA. It doesn’t like it. Really not …