Boundary Blindness: How Successful Firms Route Around the Gap That Defines the Market

In the past two weeks, four independent publications have named the same structural problem in slightly different vocabularies. ACORD published its 2026 Insurance Digital Maturity Study in April, covering 210 insurers and roughly sixty per cent of global gross written premium. Its Digital Maturity Assessment Framework spans eight capability areas, each measured from inside the firm.

Munich Re and ERGO published the Tech Trend Radar 2026 on 25 April, placing AI Governance at “Adopt”-tier maturity while conceding that the standards required to make these capabilities safe to deploy are “not yet available.”

Eurobase and Siena Treasury IQ published a Treasury Resilience Checklist on 27 April: fifty-six questions across seven sections, again framed from the firm-internal vantage.

Cory Smith of Signal & Horizon published The AI Fever Dream on 26 April, aggregating MIT’s ninety-five per cent zero-ROI finding, BCG’s seventy-four per cent no-tangible-result findi

ng, and Gartner’s eighty-five per cent failure rate, before prescribing “leadership discipline” as the response.

Four sectors. Four independent authors. Two weeks. One familiar boundary.

Each analysis reaches the firm wall and stops.

The Firm Wall

Claims move from brokers to syndicates, lead to followers, cedants to reinsurers, and eventually to regulators. AI governance depends not only on what one firm’s model did, but on what can be trusted across organisational boundaries.

The market has become interdependent. The tools for proving that interdependence has been handled remain stubbornly internal.

A senior figure in insurance data standards once gave me the analogy that captures the problem:

“ACORD is helping lay the track. But how you get in and out of that boxcar — that’s a much bigger problem.”

Exactly. The rails exist. The data moves. What does not move is the record of what was reviewed, under what authority, with what evidence, by whom, with what outcome, and whether that decision still holds. Insurance has built rails to carry claims data. It has not built the sealed container: the portable record proving who looked inside and whether anyone tampered with it.

The Decision Does Not Travel

The symptoms vary by market, but the structure is familiar. In the Lloyd’s claims chain, review is concentrated with the lead agreement party and, on complex claims, with a second lead acting for the followers. Those further down the chain do not re-review the claim. Structurally, they are not supposed to. That is the point.

The trouble arrives later, wearing the sober clothes of governance. A board attestation, regulatory inquiry, reinsurer, or downstream party needs to know what the lead reviewed, under what authority, and on what evidence. At that moment, the market discovers that the decision may have travelled, but the structured record has not.

Reinsurers re-verify what cedants accepted because the cedant’s review does not arrive in usable form. Auditors reconstruct decision histories from emails, file notes, attachments, and institutional folklore. The work has been done, but its record has not survived the boundary it must later cross.

A claims-platforms consultant put it plainly:

“You’re not getting that streamlining of the decision metadata staying with it. The data moves, the decision doesn’t.”

That is the gap. Senior people have recognised it for years. So why has the market not solved it? Because successful firms have become good at routing around it.

The Workaround Trap

Senior practitioners do not usually say, “We have a market-wide coordination failure.” They say, “Here is how we handle it.”

Loss adjusters and insurers have spent decades developing structured reports, standard headings, file access, audit-ready materials, and insurer-specific manuals. But that success is also the trap. The market has not solved the gap. Two parties have bridged it locally.

Other firms remove the boundary altogether. A UK personal-lines insurer I interviewed has structured itself as a broker-plus-insurer under one corporate group, partly so certain coordination issues never cross a firm boundary. Why solve the boundary problem when one can politely domesticate it?

A senior loss adjuster described old claims where the reasoning was never externalised:

“A lot of the reasoning is in their head.”

The hardest cases are those with no documentation, missing emails, or missing phone notes. That is not merely a documentation problem. It is a portability problem. If the acceptance context does not travel, the next person must reconstruct it

Boundary Blindness

The point is not that firms are behaving badly. They are behaving rationally. Each has built habits, templates, counterparties, controls, contracts, or corporate structures that help it survive the absence of shared infrastructure.

The market-wide gap persists, but the most successful firms experience it least. They have built around it so effectively that it fades from view. Asked about the cross-firm verification gap, senior practitioners describe the workarounds accurately and often conclude that it is not a daily friction point. From their seat, it often is not.

But markets are also made of uneven, stressed, delegated, and multi-party relationships: followers without the lead’s evidence; administrators without visibility; expert witnesses returning years later; regulators trying to evidence outcomes; and parties whose verification depends on decisions they cannot reach.

That asymmetry is boundary blindness. The market is not failing to see the gap because it is hidden. It is failing to see it because every successful participant has made it invisible to themselves.

Where the Workaround Starts to Fail

In one corner of the market, the workaround has begun to give way. Conveniently, it has chosen to fail in front of the auditor.

A Head of Claims at a major UK personal-lines insurer described what happens when an external auditor challenges a declined claim. The auditor accesses the cloud system, listens to recorded calls, contacts suppliers, checks accounts, then sits with claims staff and proceeds decision by decision, asking why each one was made.

The reasoning is not portable in the file. It is reconstructed in the room.

A file may contain documents, notes, recordings, reports, payments, dates, and enough administrative sediment to satisfy anyone fond of folders. But it does not necessarily contain the structured acceptance context: what was weighed, under what authority, against which rule, by whom, with what outcome, and why that outcome remained defensible.

So the market substitutes human reconstruction for missing infrastructure. Senior claims executives spend more time on audits, risk reports, outcomes, and compliance than on managing live claims or training staff.

One further reason makes adoption urgent: AI is moving into production. If the verification record of a human decision does not travel today, the verification record of an AI-mediated decision will not magically travel tomorrow. The gap is structural. AI gives it better shoes and a faster horse.

The Question That Follows

If boundary blindness explains why the gap persists, what would have to be true for a regulated market to move from individual workarounds to collective infrastructure?

SWIFT, ACORD, Stripe, Visa, and other categories emerged when bilateral friction became expensive and visible enough that shared rails beat private improvisation. But those transitions require timing, governance, incentives, trust, and a neutral framework that participants do not feel is someone else’s enclosure.

Whether those conditions now apply to UK insurance, banking, treasury, AI governance, or cross-border trade is open.

When multiple sectors name the same stopping point — the firm wall — the market is telling us something. The next infrastructure layer will not merely move data. It will move the acceptance and decision context.