AI is redrawing the top of the org chart, reshaping C-suite and board roles, while enterprise architecture gets promoted to the moat for scaling AI. The newest EA platforms can map a company faster than ever. None can change the reporting line above them.
José Freitas is Lead Enterprise Architect at IATA, the trade association for the world’s airlines, which sets many of the standards global aviation runs on. His fifteen-plus years span chief architect roles in banking at IBM across the DACH, CEE, and Benelux markets, plus M&A architecture work at Deloitte. Earlier this year, he argued data lineage decides AI’s fate in regulated industries. This time the target was the org chart.
“The authority to produce enterprise impact does not belong to a single division, but at a higher level, such as the CEO or the board,” said Freitas. “When enterprise architecture sits under the CIO, it is shaped by what IT wants, even when the board or the rest of the business may not be there.” In theory, the discipline spans the enterprise. TOGAF connects strategic objectives to capabilities, processes, information, and technology, giving architects a way to translate board-level ambition into what the operating estate can deliver.
The steward’s brief: “People think you define the strategy. No. You understand the objectives defined by the board, and your obligation is to convey them using principles and other tools, so the business comes to its own strategy,” Freitas said. “You are just the steward.”
One peer can’t referee: “The CIO is just a peer of all the other peers. He doesn’t have the authority to drive these things,” he said. “You need empowerment at a higher level.” Capability ownership is where the problem shows, part of the same authority fight now running from boards to dev teams. “Sometimes people feel they are both responsible. It’s like co-parenting, although it’s not really your kid. You feel it like yours, but in reality it belongs to the organization.” Without escalation routes and a named owner, a capability model is useless, and the org chart rarely shows where work actually runs.
For Freitas, AI proved most useful when it stopped proposing futures and started exposing the present, including the inherited processes an organization had carried for decades without questioning.
The 20-year X-ray: Process tools that pull data straight from SAP and Salesforce surface what workshops miss. “You may be dragging a process you constructed twenty years ago into the reality of today and not even question it,” Freitas said. “When you look at it, you have the opportunity to improve and calculate what it implicates in investment and resources.” The gains sit in reshaping decisions rather than automating the old workflow.
Where the tools stop: Populating a platform assumes a capability model exists, and that model still gets built through negotiation. “You shouldn’t go to a business stakeholder and tell them, ‘These are our business capabilities.’ That doesn’t work, especially in an area where you were never a professional,” he said. “You need all of that negotiation before to come to a common language.” At some point, someone also has to be named accountable. “You cannot fix this with any AI tool.”
An LLM without a problem: “People will always elaborate: 'I want an AI, and I want an LLM.' But when you have quality data with a rules engine on top, you know all the conditions, you know the appropriate reply,” Freitas said. “There is no value in these scenarios.” Even where an LLM could do the job, the economics increasingly demand a reason to use one, with soaring costs reshaping model choices while most companies still hunt for bottom-line impact.
The hardest constraint, however, wasn’t the platform or even the reporting line. It was finding architects capable of operating across both. Titles mean different things everywhere, seniority rarely matches knowledge, and the role demands an unusual combination of domain expertise, communication, and conflict management. “You need the knowledge of the trade and the capability of engagement. Explaining complex stuff in an easy way, and managing conflict. The clarity is for the stakeholders to make their best choices. We are just an enabler.”
That capability has to be developed, yet many organizations expect it to appear on its own. Freitas credited IBM’s career framework with shaping him and faulted companies that skip structured mentoring, onboarding, and job design. “You don’t have a career framework. You don’t have mentorship. You have nothing that propels you within the organization. Organizations don’t do that, and they pay a price.”
His proposed floor is six hours of structured development per month. “Maybe six hours per month, with a specific curriculum and these kinds of meetings, would be extremely valuable,” Freitas said. “When retention is an issue, understanding that the company has a vision for you as a professional gives you an appreciation and a willingness to do more.”