200 due diligence questions in 1 hour 30: transforming production capacity without compromising legal quality

In short.
A data room opening on Friday, two hundred due diligence questions, three partners of whom only one is available: that is the typical situation of a firm on a deal. The sourced draft produced in an hour and a half does not replace the lawyer; it moves their work from compiling to making judgement calls.

The call comes in on a Wednesday at 11 am. Your biggest client is launching an acquisition. The data room opens on Friday. There are 200 questions in the due diligence questionnaire. The preliminary report is expected on Tuesday.

You have three partners available. One is in court on a dispute on Thursday and Friday. Another is finalising a deal that was due to close this week. That leaves you one.

You accept. Because it is your client. Because saying no is not an option. And because you know (you hope) that something will free up.

This scenario is not exceptional in firms that handle M&A, restructurings or complex transactions. The tension between production capacity and quality requirements is constant. And it is always resolved the same way: by working faster, longer, with fewer people, and by taking a quality risk that nobody names.

What “working faster” really costs

The risk of producing under pressure is not the obvious mistake. A good lawyer spots the obvious mistake even when tired. The real risk is the invisible mistake, the one that cannot be seen in the answer itself but emerges on cross-reading.

An answer on the subcontracting policy that does not exactly match clause 8.3 of the framework agreement, because the associate used the 2022 version instead of the 2024 amendment. A cited certification whose renewal date is three weeks away, which nobody checked in the rush. A claim about the target’s ESG policy that relies on the public annual report, not on the internal documents in the data room.

These mistakes do not come from a lack of competence. They come from a production process that asks a talented lawyer to do two incompatible things at once: go fast and be rigorous.

The solution is not to choose between the two. It is to separate the tasks.

Before and after: the same deal, two processes

The traditional process: 200 questions, 4 days

Day 1. The associate receives the questionnaire. They go through it, identify the questions and mentally sort them into categories. They start looking for the source documents in the data room, in the firm’s archives, in internal databases. They answer the simplest questions. The more complex ones wait.

Day 2. They carry on producing. They send messages to the partner in charge to approve two sensitive points. The partner replies in the evening. The associate adjusts.

Day 3. Review. They spot two inconsistencies and correct them. Two documents are missing: they send a follow-up request to the seller through the data room’s Q&A system.

Day 4. The partner reviews the whole thing. They change five wordings, add two legal nuances and approve. The document goes out.

Total time: 4 days. Quality: good, provided nobody fell ill in the meantime, the data room was not updated halfway through the process, and the client did not call with three extra questions on Friday evening.

The augmented process: 200 questions, 1 hour 30 + review

Hour 1. The questionnaire is loaded into Optivalue.ai together with the data room documents and the firm’s relevant internal policies. The platform produces a first draft of the answers by drawing directly on the source documents. Every answer is sourced: document, page, date. Certifications with less than 6 months of validity left are flagged automatically. Questions with no identified source document are marked as to be completed.

Hour 1:30. The associate reviews the draft. They are no longer searching: they are assessing. They correct two wordings that are technically accurate but legally imprecise in this context. They add a nuance on the liability clause that only someone who knows the matter can put into words. They complete the three questions with no identified source.

Hour 2. The partner approves in 30 minutes. They do not have to read 200 answers from scratch: they review a document that is already structured, with the points of attention flagged. They focus on the five answers with strategic stakes. They approve the rest.

Total time: 2 hours. Quality: structurally higher, because every answer is sourced, the certifications are checked, and the partner spent their 30 minutes on what requires their level of expertise, not on compiling.

The lawyer’s role does not shrink. It shifts.

This is the point on which the firms most reluctant about AI most need a clear answer.

Optivalue.ai does not produce legal advice. It produces a structured, sourced and organised draft, from your own documents. Nothing goes out until the lawyer has read, assessed and approved it. Responsibility remains entirely theirs. The signature remains yours.

What changes: the balance between the time the lawyer spends searching and the time they spend thinking.

In the traditional process, 70% of a lawyer’s time on a due diligence questionnaire is spent on tasks that AI can perform: identifying source documents, extracting the relevant information, structuring the answers, checking validity dates. Only 30% goes on what only a lawyer can do: assessing the legal relevance of a wording in a specific context, identifying a nuance that the source document does not state explicitly, anticipating the questions the buyer will ask on reading the answer.

The augmented process reverses that ratio. The lawyer spends 70% of their time on what justifies their level of expertise and their fees, and 30% on approving and refining a draft that is already structured.

This does not devalue the profession. It revalues what is most irreplaceable about it.

What it changes for the firm

The transformation is not only qualitative. It is economic and strategic.

On capacity. A firm that cuts the production time of a 200-question questionnaire by 70% can handle more mandates with the same team, or accept mandates it would have had to turn down for lack of resources. The capacity constraint that sets the growth ceiling of many firms is structurally eased.

On quality. Systematic sourcing eliminates the most frequent class of errors: the answer that is right in principle but rests on the wrong document. In a 200-question questionnaire produced under pressure, this class of errors generally accounts for 5 to 10 answers. Five to ten answers the buyer can use in negotiation.

On the firm’s memory. Every questionnaire handled through Optivalue.ai feeds the firm’s knowledge base. Answers approved by the partners become references for the next mandates. Institutional memory builds up, instead of dying in the inboxes of associates who leave.

The non-negotiable condition: the private instance

For this process to be compatible with the firm’s professional conduct obligations, one condition is non-negotiable: a dedicated private instance.

Optivalue.ai runs in a dedicated private instance per firm, hosted in France, with no data sharing with other customers and no reuse for model training. Professional secrecy is preserved structurally: not on the strength of a promise, but on the strength of an architecture.

To go further on this point: Professional secrecy and AI: what every partner should demand before deploying a tool in their firm

Optivalue.ai produces a draft of 200 due diligence questions in 1 hour 30, sourced document by document, in a private instance dedicated to your firm. Your partners approve. Responsibility remains theirs. Request a personalised demo →