The "Divine Assembly" of an AI-first Enterprise

The "AI Maturity Paradox" is the starkest indicator that our current approach to enterprise AI is failing. Average maturity scores have dropped even when investment has gone up, falling from 44 to 35 in the last year. Clearly, the operationalization of AI is stalling and hitting its valley of despair moment. Enterprises still haven’t figured out how to pivot to the structure needed to make AI pay off at scale.

By and large, I see more of an attempt to retrofit AI programs into either legacy IT, or a transformation structure. Invariably, both structures struggle to deliver, as the reality is that AI is more than a technology upgrade; it is an operational and a mindset shift. By forcing AI delivery into a model that does not recognize the inherent need for experimentation, learning and iteration, enterprises have created a cycle of failure, friction and frustration.

We also see a trend of board or C-level leadership roles coming up at some enterprises such as the “Chief AI Officer”. While it’s great to have that level of leadership and sponsorship, I was also trying to think about what changes in roles or accountability would need to happen with the existing teams, and what sort of crew a CAIO would need to assemble.

Keeping in mind the leadership visibility, level of investment, and the urgency with which these programs are typically run, I have split the roles or accountabilities into two buckets – The “Day 1” or set-up roles, and the agile execution engine roles. More than the technical skills, I believe certain personality traits and EQ strengths will be needed for each of these roles to make the overall team successful.

The "Day 1" Foundation: Building the Framework

Before the first prototype is launched, these five foundational accountabilities must be active and working together. They are not gatekeepers but rather they must serve as the "internal service providers" that build the secure, scalable floor for innovation. The CAIO needs to set in place SLAs and ways-of-work to make sure that none of these become bottlenecks or “Engines of No”.

• The Data Steward (Internal Service Provider): An existing team augmented with a "Data Platform" mandate. Working with the CAIO, I believe the data steward knows how to prioritize their work into bite-sized chunks so that the initiatives can start without waiting for a major data-cleanup program as a prerequisite. They must act as service providers to the squads, ensuring the CMDB and knowledge silos are ready for AI consumption without becoming a bottleneck.
• The Economic Navigator (The Up-front Truth-Teller): Defines investment standards and payback paths. This is not a standard finance role, and it can be especially tricky for AI programs where the payback is not just cost savings and could even involve changes to the underlying business models. To be successful, we need people that can effectively straddle the world of finance, business and technology enough to stitch together the business cases. They set up value-capture mechanisms up front, tune it over the course of the first few implementations, then fade into automated governance within 6–12 months.
• The Trust Officer (The Educator & Architect): Responsible for massive enterprise education. The biggest needle mover will be based on how they help the organization understand the implications of various decisions, sensitize them to potential pitfalls and help keep the programs moving.  They also need to automate policy via frameworks like ServiceNow AI Guardian, supported by the existing IT Security organization for risk mitigation so that they don’t end up becoming the bottleneck for broader adoption.
• The Platform Architect (The Scalability Guard): Mandates a "Clean Core" strategy. They ensure that the plumbing, the APIs, Integration Hub, and data schemas, are in place, and are robust enough to handle the unpredictable load of agentic AI. Rather than being a technical gatekeeper, they act as a "Scalability Guide," providing delivery squads with the architectural blueprints that prevent "spaghetti automation" before it starts. Their success is measured by how effectively they enable rapid prototyping without creating future technical bankruptcy.
• The Infrastructure & Security Liaison (The Perimeter Guard): Creates "Pre-Approved AI Zones" working with the CISO’s office. We’ve all seen the uncertainty and limbo that happens when there is a security freeze, and nobody knows who has to take the decisions necessary to unblock work – that’s where this accountability has to step in. They clear the path for data residency and encryption up-front so squads can iterate without waiting for a new security audit every two weeks. This involves pre-validating data residency, encryption-at-rest protocols, and LLM access layers so that squads can iterate at speed. They don't just secure the perimeter; they build the "secure tunnels" that allow data to move at the speed of business, transitioning from manual audits to automated, real-time security guardrails.

The Execution Engine: Breaking through the corn maze

With the base in place, the execution accountabilities need a mindset of breaking the "entrenched" thinking and systems that hinder scale, as well as the mandate from senior leadership and air cover from the CAIO to do so:

• Value & Workflow (Strategic Ruthlessness & Creative Destruction): Squads must have the mandate to re-imagine and re-design processes as needed. Folks with this accountability advocate for radical simplification and "Out-of-the-Box" (OOTB) standards. The toughest part of this job would be breaking down entrenched departmental silos to rebuild them into streamlined, AI-first workflows that prioritize speed and outcomes over the legacy “this is how it was always done” mindset. If a workflow is "entrenched," the mandate is to freeze the rollout until the underlying process is simplified to make it agentic-ready.
• The Human Orchestrator (Radical Empathy in Action): Be the bridge to solve the "Vision Gap" by bringing front-line employees directly into the rapid-prototyping loop. It’s critical that change enablement be recognized as a pre-design necessity rather than just a post go-live afterthought. Their success is not measured by training completion rates, but by Co-creation and adoption. They will be successful if they can help re-frame the initiative as a growth initiative for everybody involved and not just a bottomline play for the organization – growth for the front-line will come from enhanced skills and getting them the ability to exploit AI for their own growth. They ensure the AI is perceived as an ally that augments the human, rather than a competitor for their paycheck.

The Bottom Line: The CAIO’s Evolution

The Chief AI Officer (CAIO) acts as the Strategic Rewirer. They do not manage a permanent department; they manage the transition of these accountabilities into the DNA of the business. The personality and leadership drive needed to help their team break down organizational barriers and inertia, and identify the right leaders to take on these various accountabilities are what’s going to make them successful.

Eventually, I believe the CAIO role must become obsolete and merging into a broader business or strategy leadership role. If the CAIO still exists as a specialist after five years, the "rewiring" has failed.

Just to make this interesting, I asked my friendly LLM to reimagine these roles as figures from Indian mythology. After a few back-and-forths, it came up with a "Divine Assembly" that surprised me with how much it resonated. It identified the CAIO as Vishnu, the strategic sustainer who maintains the balance between innovation and enterprise integrity. The Workflow Reinventor becomes Shiva, the force of creative destruction needed to dismantle legacy "cowpaths," while Ganesha represents the Human Orchestrator, removing adoption obstacles through empathy and co-creation.

The foundation is supported by Varuna, ensuring the pure flow of data; Kubera, the Economic Navigator guarding the treasury of AI investments; and Dharma, upholding the ethical order of automated trust. Finally, Adishesha provides the stable "Clean Core" as the Platform Architect, while Kartikeya stands as the vigilant commander of infrastructure security. While I got a chuckle out of this imagery, the archetypes are pretty strong and may be useful in getting the message across.

How do you see this playing out? Are there other critical gaps you see in the modern AI operating model?