Sellers want to believe that we are in the midst of an AI revolution, one that is changing the nature of the way we work. But the truth, according to several recent studies, is that it’s much more nuanced than that.
Companies are keenly interested in generative AI as vendors push the potential benefits, but turning that desire from a proof of concept into a working product is a much bigger challenge: they face the technical complexity of implementation, whether it’s technical debt from a legacy technology stack or simply a lack of people with the skills Occasion.
In fact, a recent Gartner study found that the top two barriers to implementing AI solutions are finding ways to quantify and demonstrate value at 49% and a talent shortage at 42%. These two elements may constitute major obstacles for companies.
Consider it Study conducted by LucidWorksan enterprise search technology company, reported that only 1 in 4 of those surveyed reported successfully implementing a generative AI project.
Amer Baig, senior partner at McKinsey & Company, speaks at the conference MIT Sloan CIO Symposium In May, he said his company also found in Recent study Only 10% of companies implement generative AI projects at scale. It was also stated that only 15% saw any positive impact on profits. This suggests that the hype may be far ahead of the reality for most companies.
What is disruption?
Page finds complexity to be the primary factor that slows down companies even with a simple project requiring 20-30 technology components, where a proper LLM is just the starting point. They also need things like proper data and security controls, and employees may have to learn new capabilities like agile engineering and how to implement IP controls, among other things.
Outdated technology stacks can also hold companies back, he says. “In our survey, one of the top obstacles cited to achieving generative AI at scale was actually having too many technology platforms,” Page said. “It wasn’t the use case, it wasn’t the data availability, it wasn’t the path to value; it was actually technology platforms.”
Mike Mason, chief AI officer at consulting firm Thoughtworks, says his company spends a lot of time preparing companies for AI — and their current technology setup is a big part of that. “So the question is, how much technical debt do you have, and how much is the deficit? And the answer will always be: it depends on the organization, but I think organizations are increasingly feeling the pain of that,” Mason told TechCrunch.
It starts with good data
Much of this preparedness deficit is due to data, with 39% of respondents to a Gartner survey citing concerns about a lack of data as the main barrier to successful AI implementation. “Data is a huge and daunting challenge for many, many organizations,” Page said. He recommends focusing on a limited set of data with an emphasis on reuse.
“The simple lesson we’ve learned is to really focus on the data that helps you across multiple use cases, which typically ends up being three or four areas in most companies that you can actually start and apply to your top priority business challenges with business values and delivering something that delivers,” he said. actually down to production and volume.”
A big part of being able to successfully implement AI has to do with data readiness, but that’s only part of it, Mason says. “Organizations are quickly realizing that in most cases they need to do some AI readiness work, some platform building, data cleaning, all that kind of stuff,” he said. “But you don’t have to take an all-or-nothing approach, and you don’t have to spend two years before you can get any value.”
When it comes to data, companies also need to respect where the data comes from – and whether they have permission to use it. Akira Bell, head of IT at Mathematica, a consulting firm that works with companies and governments to collect and analyze data on various research initiatives, says her company has to tread carefully when it comes to working that data into generative AI.
“As we look at generative AI, there are certainly going to be possibilities for us, looking across the ecosystem of data that we use, but we have to do it carefully,” Bell told TechCrunch. This is partly because they have a lot of private data with strict data use agreements, and partly because they sometimes deal with vulnerable populations and have to be aware of that.
“I come to a company that takes my job as a trusted data steward very seriously, and as a CIO, I have to be entrenched in that, both from a cybersecurity perspective, but also through how we engage with our customers and their partners. Data, so I know,” she said. How important is governance?
Now, she says, it’s hard not to get excited about the possibilities that generative AI brings to the table; Technology can provide much better ways for an organization and its customers to understand the data they collect. But their job is also to move carefully without standing in the way of real progress, which is a difficult balancing act.
Find the value
Just as when the cloud emerged a decade and a half ago, IT managers are naturally being cautious. They recognize the potential that generative AI brings, but they also need to pay attention to fundamentals like governance and security. They also need to see a real ROI, which is sometimes difficult to measure with this technology.
In a January TechCrunch article about AI pricing models, Sharon Mandel, IT director at Juniper, said it was difficult to measure the return on investment generated in AI.
“In 2024, we will test the hype around synthetic genes, because if these tools can produce the kinds of benefits that you say they do, the return on investment is high and may help us eliminate other things,” she said. So she and other IT managers are running pilot programs, moving cautiously and trying to find ways to measure whether there is actually an increase in productivity to justify the increased cost.
Page says it’s important to have a centralized approach to AI across the company and avoid what he calls “too many skunk initiatives,” where small groups work independently on a number of projects.
“You need support from the company to actually make sure that the product and platform teams are organized, focused, and working at pace. And of course, it needs the vision of senior management.
None of this guarantees that an AI initiative will be successful or that companies will find all the answers right away. Both Mason and Page said it’s important for teams to avoid trying to do too much, and both stress reusing what works. “Reusability translates directly to speed of delivery, which keeps your business happy and makes an impact,” Page said.
Whatever companies implement generative AI projects, they should not be paralyzed by challenges related to governance, security, and technology. But they should not be blinded by the hype: there will be many obstacles for almost every organization.
The best approach might be to get something that works and shows value and build from there. And remember, despite the hype, many other companies are struggling too.