OpenAI in the Enterprise: Everything, Everywhere, All at Once?
Welcome to AI Pioneers at Work, a newsletter for technical leaders building AI products in the enterprise. To join our community of >1,000 product and engineering executives using LLMs and modern AI, please sign up here.
The dramatic moments at OpenAI over the past 10 days were a reminder of the significance of a ~770 person startup to some of the largest companies in the world. Much of the recent attention focused on how much Microsoft had at stake through their partnership, but they are not the only major enterprise with a lot riding on OpenAI’s success. Contrary to popular belief, AI adoption in the enterprise is not lagging consumer adoption – it is a top strategic initiative in many board rooms and has been an urgent priority for many executive teams. The reality is that the largest companies in the world are well on their way to becoming AI companies, and an impressive 92% of the Fortune 500 are using and / or are becoming reliant on OpenAI in some capacity for their AI strategy.
At our recent AI Pioneers Summit, we gathered >250 product and engineering leaders from 160 companies who have been early adopters of AI and LLMs in the enterprise. We additionally conducted research on adoption of AI within the largest companies, and surveyed many Fortune 500 execs on their AI-enabled products and deployments. We were pleasantly surprised to find that 75 out of the 100 largest companies in the world, as represented by the Fortune 100, had already announced modern AI applications using LLMs for internal or external use, with the vast majority using OpenAI’s platform in some meaningful way (beyond ChatGPT, consistent with many published surveys and reports). Is OpenAI the fastest adopted enterprise technology in history? If so, then 2023 will likely be remembered as the year that every company became an AI company, and the year that OpenAI landed in the enterprise to AI-enable everything, everywhere, all at once. How did this happen?
All At Once: “We Will Not Be Left Behind”
Though very few enterprises report vendor adoption, it is likely that OpenAI’s GPT-4 model is on track to be the most rapidly used technology in B2B history. Why did the internet, cloud computing, and mobile uptake have much longer adoption curves? It is often the case that a new platform technology requires the overhauling of existing infrastructure, the acquisition of new technical talent, and a mindset shift that is difficult for incumbents to navigate. This traditional paradox does not seem to apply in the early innings of AI, in no small part due to the power and breadth of OpenAI’s GPT-4 model and the plug and play API abstractions of their platform which fit well into existing teams and products. We have also observed a growing recognition among executives in large companies about the critical importance of AI and the benefits of embracing it early on. As Cheryl Ainoa, EVP of Walmart Global Tech, summarized at our AI Pioneers Summit: “Walmart may have been behind in other tech trends in our history, but we intend to lead in the field of AI.” Walmart’s current initiatives span from incorporating AI into their "Text to Shop" program, to optimizing internal operations with their "Generative AI Playground" and "My Assistant" tools for employees. Technological readiness and eagerness to adopt have aligned at an unprecedented speed, leading to a level of enterprise adoption that was previously unimaginable.
Logos above: The 75 Fortune 100 Companies that have announced Generative AI products
Everywhere: Adopt the Language of the Fortune 100
The Fortune 100 comprises both tech companies and traditional industrial conglomerates. These companies have complex IT policies and legal requirements for deploying new technologies, often resulting in slower deployment. How did OpenAI find its way across the largest companies at warp speed, in a segment of the market traditionally seen as slow moving? In addition to the rapid time to value and the power of GPT-4, organizations across numerous verticals credit OpenAI and Microsoft’s early focus on the enterprise as one reason they have been able to overcome privacy and deployment restrictions. If you asked your favorite enterprise CTO, they’d tell you their engineering team is only allowed to use Azure OpenAI, not the OpenAI endpoint itself, as it has different promises around availability, privacy, etc. Startups such as Credo AI have also emerged to help organizations meet forthcoming AI regulations and satisfy compliance requirements. Even traditionally late adopting industries such as healthcare and manufacturing are now freely experimenting with modern AI, either with OpenAI or with open-source AI platforms, yet another reason to believe there will be ubiquitous adoption across the Fortune 100 by this time next year.
Everything: There Are No Boundaries
Many people with a history in enterprise tech adoption assumed that due to risk and potential for harm, the initial deployments of AI technology would be limited to narrow internal use cases. Our research into the Fortune 100 found the opposite - out of the 75 companies using generative AI, 62 have launched or announced public-facing AI apps and only 13% have publicly announced internal-only applications. Though we suspect many more internal use cases have been deployed in private, it is clear that enterprises are moving quickly to embrace external use cases despite the perception of risk. Why move so quickly? We found that the main driver of AI adoption is not just competitive pressure but economic need - sectors like healthcare, finance, retail, and telecom have urgent requirements to use customer data to improve customer service and support at a time when labor costs are rising, customer interactions are increasingly digital, and budget for new headcount is constrained. Becoming AI-enabled for the Fortune 100 is not just about staying ahead of the competition – it is also critical to drive real efficiency from the promise of AI in a tough economic environment.
2023: The Rise of AI Pioneers at Work
As with every disruptive movement, the wave of AI adoption in the enterprise did not come solely from OpenAI’s sales and marketing effort, but instead relied on the groundswell movement in AI engineering. Over the past year, we have seen the rise of innovative product and engineering leaders who have had the courage, technical acumen, and leadership required to build and deploy LLM-enabled applications. We call these individuals the AI Pioneers at Work - those who have taken the risk to lead the field and break ground by using LLMs in startups and large companies alike. There are a handful of these individuals in every organization, and we will be profiling their products, use cases, technology decisions, and personal stories in the weeks to come. Within our community we hope to answer many of the major questions facing AI engineers and product leaders today:
· Other than OpenAI, what alternative platforms are you using?
· How are you diversifying your technical and vendor risk?
· Is open-source a viable option for your company?
· Should you be training or fine-tuning your own models?
· How do you balance accuracy and affordability for every use case?
· What are the highs, lows, and lessons learned in shipping AI products?
· How can you network and learn from like-minded AI leaders?
We are excited to bring this community together and if you are interested in joining, please sign up here!