{"id":17167,"date":"2026-11-04T13:55:16","date_gmt":"2026-11-04T13:55:16","guid":{"rendered":"https:\/\/sr2rec.com\/?p=17167"},"modified":"2026-11-04T13:55:18","modified_gmt":"2026-11-04T13:55:18","slug":"ai-didnt-kill-the-enterprise-architect-it-made-them-essential","status":"publish","type":"post","link":"https:\/\/sr2rec.com\/en\/consulting\/ai-didnt-kill-the-enterprise-architect-it-made-them-essential\/","title":{"rendered":"AI didn’t kill the enterprise architect. It made them essential."},"content":{"rendered":"\n

You have the technical answer. The architecture is sound, the delivery plan is credible and the team can stand up the pipeline. <\/p>\n\n\n\n

Then the bank or insurer scores you down on a question the architecture can\u2019t answer: do these people understand how our business runs? If you\u2019ve sat through that debrief, you know it.<\/p>\n\n\n\n

AI has made the first half of that gap cheaper to close. It\u2019s currently doing nothing for the second. This is our read of the market for the people who lead financial services (FS) and insurance consultancies.<\/p>\n\n\n\n

What AI made cheaper<\/h3>\n\n\n\n
<\/div>\n\n\n\n

Building. Standing up a data pipeline, writing the code and producing a first model all take less time and fewer people than they did, and clients know it. That changes what they will pay a premium for – and<\/em> it also changes what a consultancy has to show at bid stage, because the buildable part no longer separates one bidder from the next.<\/p>\n\n\n\n

Any bench of capable engineers can now show a client a working prototype. The consultancy that wins is the one whose people can explain why that prototype will (or won\u2019t) survive contact with the client’s business.<\/p>\n\n\n\n

What AI didn\u2019t make cheaper<\/h3>\n\n\n\n
<\/div>\n\n\n\n

Deciding what should change, and what good looks like in a particular industry. In FS and insurance that judgment is concrete. It means knowing which controls a change will touch, which parts of the operating model are fixed by regulation and which are habit, and what a credible sequence looks like for a firm running legacy platforms.<\/p>\n\n\n\n

Regulators haven\u2019t taken that decision away from people. The Financial Conduct Authority (FCA) says it does not plan to introduce extra regulations for AI and will rely on existing frameworks<\/a>, and that its rules emphasise accountability for senior managers<\/a>. Someone senior at the client still owns the outcome. The consultancy advising them needs someone who can talk to that accountability in the client’s own terms.<\/p>\n\n\n\n

Who is the scarce profile?<\/h3>\n\n\n\n
<\/div>\n\n\n\n

The senior generalist:<\/strong> the enterprise architect, the transformation lead, the delivery director. They understand an operating model well enough to redesign it, not just build against it. They rarely describe themselves as AI specialists, and a CV search for AI skills won\u2019t surface them.<\/p>\n\n\n\n

Many consultancies try to cover the gap with a bench of specialists. A bench of specialists can build, but it seldom contains the person who says an approach will not get through the risk committee, or that a controls team will want to see the design before the build starts.<\/p>\n\n\n\n

What does this mean for bids?<\/h3>\n\n\n\n
<\/div>\n\n\n\n

In our view, the answer is to name the person who has done it, and put them in the room early. An architect with a track record in the client’s kind of firm carries more weight than the same bid with a strong generic bench and a generic summary of experience. A specialist added in Week 3 of a programme can\u2019t repair a weak answer given in Week 1.<\/p>\n\n\n\n

It also means saying no more often. Some mandates need a profile you don\u2019t have and cannot credibly name before the deadline. Bidding on those costs time you could spend on mandates you can win.<\/p>\n\n\n\n

How do you test your own bench?<\/h3>\n\n\n\n
<\/div>\n\n\n\n

Start with three questions:<\/p>\n\n\n\n