Main content

"For technology to stick, it needs to be useful" - why G2's Matthew Miller predicts generative AI downturn in 2024

Barb Mosher Zinck Profile picture for user barb.mosher October 31, 2023
Never mind the hype, feel the practical reality! A contrary view to the currently accepted norm...


Generative AI is the technology for 2023. But will it be for 2024? We know it’s not going anywhere, but Matthew Miller, G2’s Principal Analyst, AI, Automation, and Analytics, thinks its use is trending downward - and he’s not alone. 

Miller and I talked about two types of AI products:

  1. Pure play AI vendors - These are the vendors producing specific AI tools for data scientists and developers. Google’s Vertex AI sits at the top of this set of AI categories.
  2. Products that cut across categories and incorporate some AI (we focused mostly on generative AI for this discussion). 

Let’s break that second group down a bit. As a software marketplace provider, G2 tracks 150,000+ products across 2100 categories. Miller says that G2 is seeing a proliferation of generative AI across those categories. Two weeks ago, 1600 products had some generative AI, and now there are 2600 products. That’s two percent of ‌the total products G2 tracks. 

In the G2 State of Software, Q4, 2023 study, AI is experiencing the fastest growth across all 26 high-level software markets at 39% year over year. Its growth is 2x greater than the next category, which is design. Even marketing only comes in at 17% growth year-on-year (YoY) (September 1, 2022 to September 1, 2023).

Synthetic media is the fastest-growing category, with a 222% YoY increase. This includes image generation tools, audio and visual generation tools, and content generation. The next two growth categories are AI Writing Assistants and Text-to-speech tools. 

Chatbots are another fast-growing category, coming in fourth. Miller pointed out that some are more intelligent than others, but with the rise of LLMs and GPT 3 and 4, these vendors have been able to ingest powerful conversational AI capabilities into their bots to improve support and increase customer engagement. 

One area that Miller is watching is the platform play, where a vendor is providing a set of content generation tools to support all content needs. Canva is a good example with its Magic Studio, including AI tools for image and video generation and text generation. There are other tools out there, but most are not that great, Miller says.


Miller acknowledges the movement of AI technology, especially after the launch of Chat GPT last November and the subsequent launches of GPT 3 and GPT 4. GPT has enabled many companies to incorporate generative AI into their products easily.

But he also thinks a slowdown is inevitable:

I think not only will we see a slowdown across categories, across products. We'll probably see some of these features fizzle away, not necessarily be part of the products, be part of the categories.

G2 has 30 analysts looking across its product categories, and they found 200 categories that they were confident some products had generative AI capabilities. This number will go up and down, he said, but overall it will trend downward. 

The question is, why? Miller has a starkly simple response: 

For technology to stick, it needs to be useful.

Buyers look for technology and capabilities that meet their requirements. That means the capabilities help them do what they need. According to ‌G2 data, over time, these generative AI features haven’t helped buyers meet their requirements. 

On the other hand, sellers want features that provide ROI. So, if these generative AI features are not helping buyers, ‌sellers are paying for features that aren’t producing or improving ROI. 

That’s why Miller predicts there will be a:

fizzling and lessening of the number of categories that have generative AI features.

When you look at Gartner’s Hype Cycle for Emerging Technology, you find generative AI at the peak of inflated expectations. The research firm also placed generative AI at the peak of inflated expectations for Revenue and Sales Technology. It’s a similar story. Generative AI is not always producing the returns expected. 

WriterBuddy also did some research on the over three thousand AI tools available today. They isolated the top 50 that generated over 80% of the AI industry’s traffic. Of those, ChatGPT accounted for 60% of traffic, followed by,, and

What they also found was that traffic for AI tools peaked in May at 4.1 billion website visits and has been slowly dropping off since then. WriterBuddy doesn’t necessarily think this drop-off is due to the hype wearing off, but they do think it’s one of the reasons. Other reasons included shifts in consumer preferences, regulatory changes, and the adoption of mobile applications (Chat GPT launched mobile apps in June and July of this year, potentially leading to more people using the mobile app than the website application).

Finding the right value in generative AI

So why would gen AI use be trending downward? Companies are looking for AI technology to solve problems, but many more are still not sure of the role it should play in their processes. It’s available in many different applications, from marketing to sales, to customer support, development, and more, but where does using it truly help? 

I think demand for the capabilities will grow as they figure it out and implement the proper strategies, but there will continue to be a learning period where things will get tested and if they aren’t helping achieve a business goal, they won’t continue to be used. 

On the vendor side, there is a lot of movement in implementing generative AI, but they need to use it where it will be the most helpful. So, there will continue to be a lot of testing. Sitecore’s Chief Product Officer, Dave O'Flanagan, said it best:

The other thing I've learned in my years using doing personalization, conversion rate optimization at Boxever - you never know exactly what's going to work, right? You could think that the product is awesome, but maybe the less sophisticated version has more utility. So I think one of the things that I'm keenly aware of as a product leader is that it's not yet clear what the right capabilities are. I think there's lots of opportunity. I think there's a lot of hype around this. We don't quite know yet what the right features are. So we're entering a phase in Sitecore where we're going to be super iterative and responsive.

Does every software vendor need an AI story? Customers are asking for one, even if they aren’t sure what they will do with it. So, if you’re a software vendor without a response to the question, you might be in trouble. And for those sticking to basic capabilities, it might be time to think bigger, based on the challenges your customers face. 

A good example of going beyond the basics is Jasper, an AI writing assistant. Jasper introduced new capabilities that will provide companies with an end-to-end marketing copilot, “infused with company intelligence, campaign acceleration, performance analytics, and insights.” You can upload information such as brand positions, product positioning, campaign strategy, creative briefs, customer research, and more to help the AI produce content that supports your brand. It also now offers an AI-based project management tool, and analytics insights that will help inform content strategy. 

My take

We are entering ‘predictions season’, - like it or not - so expect a lot of chatter about where generative AI is heading in the coming year(s). The truth is, it’s going nowhere. But that doesn’t mean that every software vendor with an AI feature is going to find success because of that feature. 

Usually, we say that technology is a supporting tool and that ‌processes need to be first defined to know what technology is needed. With generative AI it feels reversed. Companies are being told they need it before they even understand what for and where they will get the most value from it.  

Maybe the best prediction is that we’ll meet somewhere in the middle. The technology will continue to improve and companies will get smarter at understanding the processes that improve the most when using generative AI. It won’t become a race to get AI implemented - any AI - it will be a race to improve the processes that increase employee efficiency and help customers solve their problems quickly. That requires some testing, some playing, and the willingness to acknowledge that while it’s cool and new, it’s not always necessary - until it is. 

A grey colored placeholder image