Back To Square One
Sooner than expected “Where’s the Value?” moment for Gen-AI
As the ZIRP (Zero Interest Rate Policy) era wound down in 2022, there was a spotlight on how well the technology innovation process (or not) is performing within the enterprise. There was an emerging focus on continuous improvement and improving business outcomes.
Perhaps exacerbated by constraints on growth created by the end of ZIRP, enterprise companies sought to do more with less. They were looking to cultivate a product mindset (Project to Product) and engineering discipline (IT to Engineering) to elevate the business impact of enterprise technology work. Along came Generative AI (Gen-AI), and the focus was upended.
Gen-AI sucked the oxygen of enterprise technology budgets faster than anyone expected in 2023. Even though there wasn’t much clarity of use cases, there was a general agreement that LLMs represent a generational change on what’s possible with software and what can be automated with software.
The phenomenon unleashed a type of business use case fishing unlike anything seen before. The software vendors started asking the system integrators (SIs), i.e., the consulting firms. The SIs launched new delivery practices, seeing what vendors were launching. The buyers, i.e., enterprise companies, exhorted the vendors and SIs to “bring innovation to the table.” Amidst all this, companies commissioned many PoCs to “take advantage” of the Gen-AI revolution, 90% of which don’t make it to production.
In the last 3 months, I have attended 3 different conferences in the Bay Area focused on deploying AI within the enterprise. My takeaway was that there is still plenty of confusion on how to leverage Gen-AI within the enterprise. Given all of that, it is not surprising that the zeitgeist surrounding Gen-AI within the enterprise is changing, slowly but surely, from exploring potential of Gen-AI to creating value in 2024.
The headline for Bain’s recent Gen-AI readiness report is that “2024 is about delivering results and generating real business value”. A similar McKinsey report states that 2024 is the year organizations start “deriving business value” from Gen-AI. A report from my former employer, Deloitte, states that organizations are “prioritizing value creation and demanding tangible results” from Gen-AI deployments. An arch-competitor, Accenture, is urging to “lead with value”.
These reports highlight the existential challenge of scaling Gen-AI PoCs and experiments to production-grade solutions. In addition, there appears to be a growing realization that simply automating current workflows doesn’t have the upside to justify investments. Business workflows must be reinvented with new products and services, and emphasis must shift from efficiency to innovation and growth.
After the brief excursion with Gen-AI, we are back to square one. The dilemma remains the same as it was in the summer of 2022. The questions that took a decade to emerge with SaaS and Cloud took only a couple of years with Gen-AI. How can companies create new business value from Gen-AI work? How can they discover innovation and growth from Gen-AI investments?
With hindsight, ChatGPT’s explosive growth to 100M users in 2 months may have created a decision-making trap for companies. How can you not go all-in on a technology that looks general-purpose and could do it all? LLMs certainly seemed like a technology with much potential to do more with less, which was the need of the hour.
I believe it still has the promise. Except that, we now know it will take longer than expected and that there is no low-hanging fruit. That is a great outcome for everyone who works within enterprise software. The sooner we can go from speculating what technology can or can’t do to actually building solutions that solve business problems, the better.