Generative AI is now virtually ubiquitous in global businesses, with major companies having strongly prioritized commitments to it and AI deployments having spread at a near unprecedented pace for adoption of a new technology that is still accelerating, new research from Bain & Company reveals.
Bain’s latest proprietary cross-industry Generative AI Survey1 indicates that almost 9 in 10 companies (87%, up from 83% in October last year) have already deployed or are now piloting the technology, with adoption continuing to climb rapidly across all use cases.
Bain’s analysis shows very rapid ramping up of businesses’ spending and other commitments to generative AI use. More than 60% of businesses surveyed put it among their top three priorities for this year and next, with 87% ranking it among their top five priorities for the next three to four years.
On average, Bain’s research shows companies are already budgeting some $5 million per year for generative AI activities and technology infrastructure – with that average number rising to $50 million per year for 20% of the largest companies, indicating businesses’ increasingly large-scale commitments to generative AI implementation.
Bain’s analysis of the firm’s, regular Generative AI Survey1 of senior executives in 200 businesses, across an even split of technology and non-technology companies, also shows companies’ fast-scaling commitment to the technology in terms of growing size of teams working on it. Companies have around 100 people spending some time engaging with the new AI technology in some form, with large companies having as many as 240 team members, the data shows.
As companies around the globe race to seize generative AI’s potential and capture its competitive advantages, the analysis shows the greatest focus for executives in terms of business goals from AI capabilities is to harness benefits from boosts to revenues alongside enhanced efficiency and productivity. Both goals are cited by 68% of companies surveyed among their top three primary objectives.
However, the survey data also reveals a need for businesses to strengthen their focus on how they can best use generative AI, with only around 36% of executives indicating their organization has a strong, well-defined vision for AI deployments, with a sequenced roadmap and clear value expectations. In addition, a fifth of organizations (21%) have ideas for generative AI deployment but have not yet made coordinated efforts.
Despite this, the data also indicates that generative AI is meeting or exceeding businesses’ expectations in 75% of instances overall. Alongside, around four-fifths of respondents (~80%) observe that prototyping for generative AI use is faster than was experienced with earlier, traditional AI technology and machine learning.
For the cases where AI deployments have fallen short of expectations, the most common issues cited are around poor output quality or the technology not meeting performance needs. This tech-market fit issue is common for new technologies. The next most common set of levels of issues is around user adoption and off-the-shelf tooling which did not deliver the expected value.
But the survey also shows indications that these issues are being encountered less frequently for some key use cases, with performance for use cases in sales, sales operations, marketing, customer service and customer onboarding cited by respondents as improving in the latest findings. Concerns over risk, data security and privacy, and uncertainty around regulation have also declined. Most firms still see room to improve their generative AI preparedness across the areas of data readiness, data security and talent.
“The scale and pace of generative AI adoption across the business landscape is remarkable. It speaks to this technology having a truly far-reaching and transformative impact for companies across sectors as it continues to develop – and as deployments continue to accelerate,” Gene Rapoport, Bain & Company partner and leader of AI initiatives for Bain’s Private Equity practice, said. “It’s equally impressive that, with most major business already putting money and muscle behind generative AI implementations, the majority are seeing a path towards realizing real business value. But what is also clear is that CEOs and executive committees need to take clear ownership of activating AI in their organizations and ensuring a clear, well-defined vision for its use. The businesses that do are going to emerge quickly as those that lead with AI and secure the best results and the greatest competitive advantage.”
Sanjin Bicanic, Bain & Company partner and member of Bain’s Advanced Analytics Group, added: “Many software companies are adding AI features to their products at a breakneck pace, but our research shows those solutions are not yet fully featured enough to create value for the Enterprise. This gap in perceived value combined with the availability of frontier models as APIs are the two main reasons why we’re seeing so many companies choosing to build to capture value quickly. As solutions get better, we expect to see more buying, but the landscape is shifting rapidly and it’s not yet clear where building might be a correct long-term solution.”
Four evolving themes
As adoption of generative AI increases across all use cases, concerns around organizational readiness have grown, Bain also finds. Four themes emerge from the analysis that show how companies are thinking about the technology:
- Are we delivering value yet? Across industries, conversations about generative AI are more earnest, moving from excitement and hype to more realistic assessments.
- Five promise areas: As companies get their hands dirty with generative AI, they are reporting some use cases show the best signs of success, including sales, software development, marketing, knowledge worker assistants and customer service.
- Tech companies are finding out first: Compared to Bain’s 2023 Q4 survey, companies in the tech industry were more likely to say their data and security protocols were ready for generative AI as well as being further ahead in adoption. This contrasts to companies across other (non-tech) industries, which reported about the same levels of readiness in both surveys, indicating they have not yet hit this bottleneck.
- Buy or build? Both approaches are being tested across use cases. Companies are buying third-party solutions when available but are investing in tailoring them for their needs.
