Enterprises are prioritizing artificial intelligence projects, but shortchanging the collaboration and orchestration needed across all business functions including HR, IT, finance, operations, and other areas. What's needed: Chief AI Officers.
That take is the short version of a Constellation Research report "The Urgent Case for a Chief AI Officer," by Andy Thurai and Ray Wang. The argument is that boards are giving AI projects above all other projects--including cybersecurity. And just like the rise of the chief information security officer you'll need a C-level focused on AI.
The bet is that enterprises will embed AI into all aspects of the business by 2026. Without a leader, AI projects are likely to run amok.
According to Thurai and Wang, who interviewed two dozen current chief AI officers (CAIOs), AI projects are already rolling with proofs of concepts. What's lacking is a long-term strategy to become an AI-first enterprise. Toss in a bevy of AI features--and add-ons to match--from vendors and this landscape is going to get complicated in a hurry.
Thurai and Wang wrote in the report:
"The case for a CAIO is simple. AI is still immature; hasn’t been proven in most enterprise settings; and needs many surrounding acts such as security, privacy, governance, provenance, ethicality, and truthfulness before it can offer real value to enterprises. With great power comes great liability: Such is the case with AI. There is a lot of liability in using external tools as well as using decisions proposed by AI with full confidence in real-world applications. A single point of responsibility and accountability will ensure the right orchestration and collaboration across the organization. Furthermore, organizations need a champion for AI, just as they did for digital transformation."
While the CAIO will have to set a long-term strategy, know business and technology and balance risks with rewards, there will be a more immediate chore: Delivering returns on investments.
Generative AI, commoditization and ROI
With boards pushing for AI projects, there’ll come a point where these same executives ask for results. Enter the CAIO. In recent days, Salesforce, Microsoft, Oracle, ServiceNow, SAP and Workday have all announced generative AI features across their cloud application portfolios.
We've talked names. We've talked generative AI sprawl as domain specific LLMs proliferate. And we've also talked about sticker shock as add-ons add up.
Next up: Generative AI commoditization. Yes, your enterprise software vendors will raise prices in various ways as they monetize generative AI. That approach will work for a bit.
But at some point, say 2025, enterprise software buyers will look to cull the generative AI upsell herd. Technology vendors call this optimizing the cloud spend. This copilot culling will happen because generative AI features will be commoditized. At that point, CIOs and CFOs will acknowledge early productivity wins and then start questioning returns.
Did that ServiceNow Now Assist feature really drive returns? Or was that the generative AI in that other suite? How will you know if Joule or Einstein was responsible for that productivity boost? Maybe we should give credit to Anthropic's Claude via AWS or OpenAI's ChatGPT or some open source model.
Here's the theme to watch: CAIOs are going to have to instrument their employee generative AI usage of each bot, query and pieces of content created with LLMs to quantify productivity gains. And that chore could get complicated. But first companies will need to add the CAIO role.
