Reviewing Everything That Arrives: Full-Coverage Assurance in Tektome

Most reviews don’t fail because the team wasn’t diligent. They fail because no one ever defined what was supposed to be reviewed, and so something quietly slipped through.

The sampling problem.

A submission arrives. An engineer reads the sections they know are contested. A coordinator skims the spec for clauses that caused problems last time. The model gets a quick check against the obvious constraints. The package gets signed off. And somewhere inside it, untouched, sits the thing that will surface on site six months later.

This isn’t negligence: it’s physics. A modern higher-risk building submission can run to thousands of pages of drawings, specifications, reports, and model data. No team has the hours to read all of it. So they triage. And triaging means gaps. And gaps mean risk.

Ingesting the whole package.

When a submission lands in Tektome, the platform ingests the complete document set: drawings, specifications, structural and MEP reports, fire strategies, energy statements, BIM models, and any accompanying correspondence. Not a representative sample. Not the documents flagged as changed. Everything.

The dangerous gaps aren’t usually in the documents everyone is watching. They’re in the appendix nobody re-read after the M&E scheme changed. The spec clause imported from a previous project and never updated. The drawing note that contradicts a requirement buried three sections back.

Full ingestion is the precondition for everything else. You cannot cross-reference what you haven’t read.

Cross-referencing across documents and models.

A single document, read in isolation, can look entirely compliant. The problem emerges at the joins.

Tektome reads across the whole package simultaneously: comparing what the architectural drawings show against what the structural engineer has assumed, checking whether the fire strategy reflects the evacuation routes as drawn, verifying that specification clauses are consistent with the model geometry. The platform holds the firm’s own requirements, built up over previous schemes and refined through resolved judgment, and checks every part of the submission against them.

This is the cross-referencing layer. It surfaces conflicts that survive a single-document review. And because it runs across models as well as documents, it catches the coordination gap that lives at the boundary between disciplines: the issue no single consultant owns.

Re-running on every revision.

Most review processes treat a revision as a delta. The team looks at what changed, checks those parts, and re-issues. The rest is assumed to be stable.

That assumption is where revisions introduce new risk. A change to one part of a scheme can invalidate a requirement elsewhere, in a document that wasn’t touched, against a clause no one thought to re-examine.

Tektome re-runs the full review on every revision cycle. Not just the changed documents. The whole package, from the beginning. The platform compares each run to the previous state, so the team can see not just what changed in the submission but what changed in the compliance picture, including requirements that were satisfied and no longer are.

No sampling, no gaps by default.

Sampling is a reasonable response to a resource constraint. Tektome doesn’t eliminate that constraint: it changes what the constraint touches.

Because the platform handles the mechanical reading and cross-referencing, the human reviewer’s time goes where human judgment is genuinely required: resolving conflicts, making calls on ambiguous requirements, deciding what risk is acceptable. The review doesn’t shrink, but the gap in the review does.

Coverage reported honestly.

A review that says everything is fine when it hasn’t looked at everything isn’t assurance: it’s a liability in writing.

Tektome reports coverage as a first-class output. The team can see, at any point in the review cycle, which parts of the package have been checked, which requirements have been cross-referenced, and which elements remain outstanding. An honest “not yet reviewed” is surfaced explicitly, not buried under a headline pass.

This is a direct application of the principle Dame Judith Hackitt’s 2018 Independent Review of Building Regulations and Fire Safety put at the centre of the dutyholder regime: that dutyholders must demonstrate what they did, when, and why, not merely assert it. A coverage report that shows gaps is more defensible than a sign-off that hides them.

What hasn’t been reviewed is shown. What has been reviewed is cited.

Confidence that reflects reality.

The output of a Tektome review isn’t a pass or a fail. It’s a graded confidence level, backed by evidence, across each part of the submission.

Where the package clearly satisfies a requirement, confidence is high and the citations are there to show why. Where there’s a conflict, or an ambiguity the platform can’t resolve alone, confidence is lower and the specific issue is surfaced for human judgment. Where something hasn’t been checked, that’s shown explicitly too.

This is calibrated confidence in practice. It doesn’t compress a complex picture into a binary verdict. It gives the team a precise view of where they stand and where they’re exposed at every point in the programme.

For a Head of Technical signing off a Gateway 2 submission under the Building Safety Act, the difference between “we checked everything, here’s the evidence” and “we think it’s fine” isn’t a question of comfort. It’s a question of personal liability.