Gen AI: Corporates’ Shift from Big Brands to Startup Solutions

Generative AI and chatbots are not something the world has never seen before 2022. It’s not even about Siri or Alexa, but the good old ELIZA, one of the first examples of Natural Language Processing, who would be a 57-year-old lady now. However, only half a century after, when Chat GPT and other notable large language models proved the technology as commercially viable across a vast spectrum of industries, businesses understood that they needed Generative AI solutions, as soon as possible.

Few of them, however, realize what they need Generative AI for, and even fewer understand the complexity of the task and the resources required. Here’s where we come in – accelerators and consulting companies.

Made-to-measure or ready-to-wear?

A good suit, tailored according to the individual measurements from preferable fabric, colour and with a particular occasion in mind, is a worthy investment. People, wearing such suits, do not worry about their look. They know they look perfectly and feel accordingly. A customised AI technological solution, which is made to reach particular business goals, has enhanced security and perfectly integrates into corporate systems, is a real James Bond suit.

This is a good comparison, which gives a general idea. But let’s dive a bit deeper into the reasons most enterprise companies prefer not to implement ready-made AI solutions, even from market leaders:

First of all, effective Generative AI integration is impossible without highly individual work for each company, which requires a separate team, informed about the company’s strategic development plans, goals, and resource availability. A Generative AI solution, which appeared workable for one company, will probably appear useless for another one.

Secondly, a smaller startup will fully immerse into the company’s specifics and offer a made-to-measure solution from a team of AI experts, who are capable of working with open-source models, securely training them on corporate data, and placing them on the client’s servers. This allows to create an on-premise solution and comply with the requirements of secure data deployment and storage, which is a priority for enterprise companies.

What do companies need Generative AI for?

As Gen AI is relatively a newcomer to the corporate market, the major way to gain experience and make progress is through trial and error, which means launching pilots. Until we have enough benchmarks across various sectors, this is by far the most productive way to find a solution that perfectly fits the company’s unique needs.

Nevertheless, there are certain trends in corporate requests for Generative AI solutions:

  1. Smart text and voice bots based on LLMs to provide high-quality assistance to customer service and support queries of different complexity levels.

  2. Employee AI assistant (i.e. sales manager’s helper, which analyzes a real-time conversation with the potential customer and simultaneously generates ideas and customer offers for a specialist)

  3. Copilots for developers

  4. HR solutions for recruitment and onboarding automation

  5. Marketing tools: images and avatar generation, writing articles, and product reviews.

‘No Gen AI is needed’ – this is the conclusion that some customers don’t expect to come to, but readily agree with after analyzing the company’s current state and business goals. AI for the sake of AI is a waste of resources, which the technology is called to eliminate.

Generative AI Market Opportunities

According to PitchBook’s estimation, the Generative AI market will reach $42,6 billion by the end of 2023 and is expected to grow at a CAGR of 32% to reach $98,1 billion by 2026. These predictions do not take into account the potential of generative AI to expand the total addressable market of AI software.

This is compared with 22.6% CAGR for the AI industry as a whole, which means that GenAI will continue to overperform relative to the larger industry.

If estimates aren’t convincing enough, here’s an illustrative fact from our experience as an accelerator. After the turbulent 2022, which is associated with the economic recession and a rapid decline of venture investments, Intema acceleration programs switched focus from fundraising to launching pilots with corporations.

In 2023, Intema held two acceleration programs with absolutely different dominant technologies: Metaverse and Generative AI. Throughout the program, we connect startups with corporate customers to discuss potential technological solutions, arrange demos and, if successful, make agreements on the potential pilots. The Metaverse acceleration program resulted in 4 pilots with corporate clients, which is great taking into account the technology’s specifics and complexity.

The Generative AI program, even several weeks before its termination, had 7 pilots in discussion with a range of corporations. So is this just the effect of a hype that used to surround Blockchain and Metaverse before? Or is it because Gen AI is a real game-changer?

It All Comes Down To the Question: Is GenAI Worth the Hype?

First off, it is not uncommon for a new promising tech or an idea to get overhyped in the short term, perhaps to the disadvantage of its longer-term prospects. If we continue draw parallels between GenAI and Blockchain, at its preliminary maturity stage, blockchain has been described by many as a technological revolution, which will reshape the world, much like GenAI is touted today. However, years later, in 2018, Gartner announced that blockchain has entered the Trough of Disillusionment, which also corresponds with more than a 30% drop in consumer interest from peak levels and a 45% decrease in VC investment from 2018 to 2019.

As opposed to blockchain, at its early maturity stage, GenAI already has many use cases across a vast spectrum of industries that are commercially viable. Their number is expected to grow as more industries adopt GenAI solutions. In its recent publication, Gartner placed generative AI technology at the peak of the so-called “hype curve,” which indicates that there might be a correction in expectations and some sort of disillusionment in the near future.

Conclusion

Does it mean that after such a massive demand for Generative AI solutions, the technology is doomed to get off the radar? This scenario is unlikely, for GenAI has already proved its fundamental tenability and flexibility in various spheres of human activity, from science to art to supply chain.

However, a slowdown in technology development is inevitable, with the major cause here being the urgent need to control and regulate the use of GenAI. So far, this instrument has been utilized relatively freely, without any legal constraints. Legal regulation will set a new trajectory in the technology’s evolution path, and it’s hard to predict where it will go, for GenAI with its current abilities is wholly unprecedented in human history.

The other factor, expected to limit Generative AI in the future, ironically is the growing size of large language models. Sooner or later the capabilities of AI chips won’t catch up with the development of the technology, and the aspiration to build Artificial General Intelligence and the growing volumes of data require highly complex engineering and much more computing power.

These limitations, however, open a vast field for research, experiments, and non-standard approaches to LLMs lossless compression, computing power growth, data storage, etc.