Companies of all stripes have begun embracing generative AI as a way to increase productivity and possibly even replace some workers. But the base technology has been around for decades. Over the last five to 10 years, big enterprise software companies have been building AI into their platforms without the same level of screaming hype we are seeing at the moment.
The problem has always been placing AI. It hasn’t helped that there has been a tendency for the early adopters to personify and productize it, even if it wasn’t really a product. Salesforce called it Einstein; Adobe, Sensei; and IBM, Watson. But AI isn’t one thing you can nail down and call a product, per se. These companies aren’t necessarily selling AI like they would CRM or Photoshop or health records software. Instead they infused AI into different parts of the product line, making the products they sell potentially better.
Since generative AI hit the scene at the end of last year with the release of GPT-4, the hype machine has been out of control. Suddenly, AI was in front of everyone in a way it never had been before. But even with the higher profile, it’s not clear this is an actual market, and it’s even less clear how many companies are truly embracing it in spite of all the ballyhoo.
Clearly, though, companies are curious about what AI can do. Enterprise software vendors have been adding functionality and branding it to get the most juice out of their marketing. Startups have been popping up, too: Cohere, Anthropic, Writer and Jasper are among those trying to build businesses around generative AI technology, whether by building foundational models or tools to take advantage of generative AI in an enterprise setting.
As we find ourselves in the midst of this AI-induced frenzy, it could be useful to take a step back and see just what we are dealing with here. Is there actually an AI market in the pure sense, or is it enabling technology that will soon be built into everything, making how we view it less clear?
Let’s start with some data
Sure, everyone is talking about AI, but what are Fortune 500 and 2000 companies actually doing when it comes to AI? It would appear that while clearly there is a lot of interest (to put it mildly), big companies are treading lightly, as large companies tend to do.
Morgan Stanley, in its quarterly survey of big company CIOs, found that 56% of respondents indicated that generative AI would have an impact on this year’s IT budgets, but just 4% said that is translating into “significant” projects, suggesting that they are still very much in the exploratory stage.
Insight Partners recently published its annual State of Enterprise Tech report for 2023 and found similar results to Morgan Stanley’s. But George Mathew, managing director at Insight, cautions that the company took the survey at the end of last year, just as generative AI was bursting into our consciousness.
At the time of the survey, around 15% of companies with over $10 billion in revenue had plans to invest in generative AI, but Mathew says that throughout this year, he’s seen a lot more interest from the Global 2000 companies that Insight works with on surveys.
“So much has shifted between the end of 2022, and the first two quarters of 2023, particularly when it comes to the question of generative AI,” Mathew told TechCrunch+. “And what we’re now seeing, even after some of the survey data came back . . . is that that investment into generative AI is actually accelerating, and it’s accelerating at a pace that we’ve actually never seen before in enterprise software.”
“I think the nuance that I would just say is that it is still in the mode of trying things, trialing, working through the initial training runs of what you’d want to build as far as generative AI applications go,” he said.
Constellation Research found similar results in its H1 CxO Business Confidence Survey, with 56% of respondents saying they plan to launch a proof of concept around generative AI this year.
Platform piece or product
To some extent, the question of where AI fits doesn’t matter, simply because it will pervade software, regardless of where you place it. But it’s still a question worth exploring, especially for startups that might be thinking about it as a category, rather than an enabling technology.
John-David Locke, an analyst at Gartner, falls firmly on the platform side in spite of a message we might be hearing otherwise. “Gosh, there’s certainly a lot of people out there with really bad research, saying AI adoption is at 4% and it’s going to run 6% of overall IT spending. All of that is nonsense. First and foremost, there isn’t an AI market. AI is a technology that’s going to get sprinkled across. It’s part of a platform,” Locke said.
Regardless, he said that it’s going to be ubiquitous. “The difference with AI is every piece of technology we have right now is a conduit market for AI. Every device, every printer, every server, every piece of software, every web application can have AI sprinkled on top — like sugar on cereal, it’s gonna taste better with AI.”
But it may not actually be a simple matter to nail down. “AI to me is an enabling platform AND a product if you have a business graph. For it to be a product, you need to have cheap compute, massively refreshing datasets, large network interactions and cheap power,” said Ray Wang, founder and principal analyst at Constellation Research.
“Only a few companies will ever be able to make AI the product, but because of that, there is an unfair power dynamic with Big Tech digital giants and startups, governments and individuals.”
Mallun Yen, founder and partner at Operator Collective, also sees it as something that will be built into everything, but will disappear as a category as it becomes ubiquitous. “Today there are companies offering AI products, but as what happened with mobile, most ultimately won’t be selling ‘AI products’ just as ‘mobile products’ isn’t really a category now,” Yen told TechCrunch+.
“Instead products will integrate and leverage AI both in terms of functionality but also how they are built and developed. Some of the most interesting startups we’re seeing today are anticipating how the world will develop, make, and provision products in an AI-native world.”
It’s clear that AI has captured our imaginations this year in a way that no technology has in quite some time, and it will eventually have an impact on every company and every software application.
As it takes over, AI will likely fade into the background. Like television shows in the 1960s bragging that they were being broadcast “in color,” before it became a given, applications will simply have AI running under the hood without having to explicitly say so.