Can AI do pre-bid due diligence?

01 The honest short answer

Yes, and no. AI can do the document-heavy part of due diligence - reading, extracting and structuring evidence - quickly and consistently. It cannot do the judgement part: deciding what a finding means for a bid, weighing it against everything else, and putting a name and a professional liability behind the conclusion.

So the useful question is not "can AI replace diligence" but "which parts of diligence is AI good at". On a pre-bid timetable, the days lost to stitching documents together are exactly the part AI compresses, leaving the judgement to the person who is on the hook for the number.

02 What AI is genuinely good at

Four things, all of which dominate the early days of a diligence exercise. Reading a large volume of unstructured documents in minutes rather than days. Extracting specific facts - a planning condition, a title covenant, a flood zone, a survey figure - and putting them in one place. Applying the same checklist to every site, so that sites in a bidding round are genuinely comparable rather than each assessed differently. And surfacing what is missing, which is often the most valuable output.

Done well, this is where the hours-not-weeks claim comes from: a structured first pass across many disciplines at once, while the bid window is still open.

03 What AI does not replace

It does not replace professional judgement and experience - knowing which caveat in an appendix actually changes the bid. It does not carry accountability: a chartered professional signs off a position and stands behind it, an algorithm does not. It does not hold the local and site-specific knowledge a good consultant brings, nor the appointment, insurance and duty of care that come with an instruction.

The honest framing is that AI is a force multiplier for experts, not a substitute for them. It changes how fast they reach the decisions, not who makes them.

04 Can ChatGPT do a risk assessment?

A general-purpose chatbot can summarise a single document you paste into it, and that can be genuinely useful. But it is the wrong tool to base a bid on. It can hallucinate detail that is not in the document. It does not reliably cite the source page, so you cannot check it. It has no structured model of the diligence disciplines, so it does not know what it has not been shown. And it is not built around a deal's data room or the audit trail you need to defend a decision in committee.

Purpose-built diligence tooling differs on exactly those points: it works from the data room, it cites every finding, and it covers the disciplines deliberately rather than ad hoc.

05 Will AI replace consultants?

Not the judgement, the sign-off or the relationship. What AI changes is the economics of the grunt work. The days a senior analyst or a panel of consultants spends assembling the picture - finding the documents, reading them, transcribing the facts - collapse towards hours. That is a real saving against the cost of instructing a full panel, and it makes it viable to assess more sites to the same standard inside a bidding window.

The experts are still needed. They just spend less of their time on assembly and more on the calls that need a human.

06 Where Plumb fits

Plumb is AI-assisted, human-in-the-loop pre-bid due diligence. It reads the documents in a deal's data room across ten diligence disciplines - planning, title, environmental, flood, ground, transport, utilities, biodiversity net gain, commercial and survey - and turns them into one structured assessment, with every finding cited back to the source document it came from.

That keeps both halves of the answer above. The analyst still decides; Plumb does the reading and structuring that gets them to a faster, more consistent and more defensible position, in hours rather than weeks.

07 Frequently asked questions

Can AI do due diligence?

Yes for the document-heavy parts: reading, extracting and structuring evidence quickly and consistently across a large set of reports. No for the parts that need professional judgement, accountability and sign-off. The realistic model is AI-assisted due diligence with a human in the loop, where AI compresses the reading and a qualified person owns the decision and the risk.

Can ChatGPT do a risk assessment or due diligence?

A general-purpose chatbot can summarise a document you paste in, but it is not the right tool for a bid. It can hallucinate, it does not reliably cite the source page, it has no structured model of the diligence disciplines, and it is not built around a data room or an audit trail. Purpose-built diligence tooling is designed for exactly those requirements.

Will AI replace due diligence consultants?

It is unlikely to replace their judgement, professional liability or client relationships. What it changes is the economics of the manual work: the days spent stitching documents together compress to hours, which frees experts to spend their time on the judgement calls that actually need them.

Is AI-assisted due diligence accurate enough to rely on?

It depends on whether the output is checkable. The standard to insist on is that every finding is cited back to the source document, so a person can verify it rather than take it on trust. Plumb is built around that principle: structured findings, each linked to the page it came from, so the read is easy to defend in committee.