Application publishers integrating conversational AI into their business solutions are navigating a complex, very complex galaxy.
Between technological opportunities and competitive challenges, the adoption of AI is transforming user interactions and development strategies.
Let’s take a look at how the forces at play are influencing this dynamic market and the strategic choices ahead.
What kind of AI are we talking about?
Here, we address conversational AI functions, i.e. those whose purpose is to perform automatic language processing (NLP) within business applications. To simplify things a bit, we can classify the main use cases as follows:
- Transactional AI enables a user to interact with his or her application directly in natural language, through a digital workplace or collaborative solution (MS Teams, Google, Slack…) or a simple chatbot window on a web access or mobile application. Ask for vacation time, book a meeting room, approve an invoice…
- Generative AI is capable of proofreading texts to improve them, proposing summaries and drafting standard texts (blog posts, sales proposals, etc.), which will certainly need to be refined and corrected, but which will save a great deal of time.
- Retrieval Augmented Generation (RAG) offers natural language access to knowledge bases or databases. Technical documentation, internal procedures, reference texts and data can all be queried directly.
The forces at work
We propose here a macroscopic representation of the “application-integrated conversational AI” environment as an application editor might encounter it.
We use Porter’s forces analysis to give a synthetic view of the situation. And we try to look a few years ahead, to get a somewhat dynamic vision.
And to avoid half-hearted appraisals, we decided to contrast the scoring vigorously, by making it binary for each of the 6 selected axes (2 or 8 on a scale of 10).
1. Your customers' bargaining power
2024 = 2 ➜ 2027 = 8
The contribution of conversational AI is new and still underestimated. Your customers need a little time to get used to it, and the solutions to fine-tune themselves. But in a few years’ time, it will be a decisive value criterion in the choice of a solution.
2. Your suppliers' bargaining power
2024 = 8 ➜ 2027 =2
Today, there is a classic first-mover advantage, and technology suppliers (NVIDIA, OpenAI, Microsoft) control a large part of the value chain. However, the proliferation of high-quality open source solutions should help to redress the balance in favor of the vendors.
3. The threat of new entrants
2024 = 2 ➜ 2027 = 8
Technological and usage breakthroughs are conducive to the arrival of new players. As in all markets (automotive, smartphone, etc.), some will play on price and business model. Others will offer high-end solutions. We can also expect segmentation, with specialized offers for specific uses.
4. Alternative solutions
2024 = 2 ➜ 2027 = 2
As far as we know, and over a 3-year horizon, there is no alternative to current conversational AI technologies (such as language model neural networks or LLM) to transform your users’ experience. Naturally, implementations will be more or less complete and more or less successful… But the landscape is unlikely to change radically. Future developments (quantum computing, general-purpose AI…) will only appear much later – if at all.
5. Competitive strength
2024 = 2 ➜ 2027 = 8
As functionalities become commonplace, acquired positions are eroded. As a result, the first-mover advantage is gradually disappearing. Unless the first mover has taken advantage of this to create a monopoly situation (an event we’ll set aside here). We can therefore expect competition to intensify.
6. Legal constraints
2024 = 2 ➜ 2027 = 8
The arrival of the various European regulations has not finished spilling ink and creating heated polemics. But the General Data Protection Regulation (GDPR) has applied since 2018. The Digital Markets Regulation (DMA) and the Digital Services Regulation (DSA) since 2023. And the AI Act has been passed and will apply from 2025 for the most part. In short, we’re going to have to learn to live with this legislative arsenal.
Challenge Hofstadter?
This analysis, necessarily superficial and general, is our sole responsibility. Perhaps developments will be more rapid than we describe.
Hofstadter’s (empirical and recursive) law tells us that every project takes longer than expected, even when Hofstadter’s law is taken into account.
In the field of conversational AI, the recent past has shown us the opposite. Things are moving faster than expected!
Open AI opened its service to the general public on November 30, 2022, just over two years ago. It’s not impossible that there will be more surprises in the near future.
There are, however, many reasons why the capacity of underlying technologies such as LLMs should be rapidly capped. In particular, Yann LeCun (Chief AI Scientist at Meta) and Jacob Browning (postdoc in NYU’s Computer Science Department) express this in the article: AI and the limits of langage.
In any case, these tools will be with us for a long time to come. We’ll tame them as we did the others. The solutions will be lighter, cheaper, better integrated, sometimes less fanciful, and hopefully less discriminating. Stripped of its cloak of invisibility, conversational AI will fulfill what is its destiny, for application publishers and users alike: that of a good servant.
And that’s something to celebrate, because it’s what we need right now.
Software publishers: AI, a lever for competitiveness for your application
The radar above shows just how much the current situation in conversational AI will have to change over the next few years: most criteria are being reversed, and this cannot happen without consequences.
Three attitudes are possible:
- Stunned: pretend to be unaware of the transformations underway. “After all, there’ll always be time to do something once we’re up against the wall. And above all, there’s no problem that can’t be solved by not solving it!”
- Haste: fear of being left behind and outdated in the eyes of customers and prospects. Let’s “do it fast” (or pretend to) at the risk of not doing it right, and developing disposable tech.
- Determination: refine use cases, test the market, prepare technical and sales teams, choose technologies and partners, plan resolutely and methodically.
Those who choose the third path will build “to last”.
It’s up to you now. We suggest you adapt Porter’s analysis to your business, your market and your situation. Then define the appropriate conversational AI strategy for your solution…
_____________
You are a software publisher? We hope this article has given you a better understanding of the changing market forces.
Agora Software is the publisher of conversational AI solutions partner of application publishers. We rapidly deploy conversational application interfaces that enhance the user experience of software and platforms. By integrating Agora Software into your solution, your users benefit from rich, multilingual and omnichannel interactions.
Would you like to integrate conversational AI into your applications?
Let’s talk: contact@agora.software
If you enjoyed this article, you might also like Hallucinations, LLM and conversational AI.
Finally, join us on our Linkedin page to follow our news!
Want to understand how our conversational AI platform
optimizes your users’ productivity and engagement by effectively complementing your business applications?