Where does document review eat a firm's time?

A lot of professional work begins with reading. A data room for a transaction. A client's lease portfolio. Decades of site reports ahead of an environmental assessment. A box of policy documents before a renewal. Before anyone can advise, someone has to go through it all and pull out the handful of facts that matter.

That reading is usually given to junior staff, because it is time consuming, and checked by senior staff, because it matters. Both halves are slow. The juniors read every page to find the few that count. The seniors then re-read sections to be sure nothing was missed.

Why is the same set of documents read twice?

Because the first read produces a summary without a trail. A reviewer notes that a lease has a break clause, but not where it is, so the partner goes back to find it. A schedule mentions contamination in a site report, but the partner wants to see the exact wording. The second read happens because the first one could not be trusted on its own, or could not be checked quickly.

The same documents are often read again on the next piece of work for the same client, because the first review was stored as a memo rather than as structured facts anyone can search.

How would you estimate what review work costs?

Take a typical review job from the last few months. Count the documents or pages, and ask the team how long the first read and the check took, across every grade. Multiply each person's hours by their charge-out rate. Then count how many similar jobs you run in a year.

On fixed-fee work, those hours come straight out of margin. On hourly work, clients increasingly push back on large bills for reading, so the hours are often written off anyway. The unbillable hours calculator puts a yearly figure on it from your own rates, and the audit shows how it compares with the rest of your admin.

What does rebuilt document review look like?

The starting point is the output your team actually needs. For a lease review that might be a schedule of parties, dates, rent, breaks and repairing terms. For an environmental desk study, a list of past uses, incidents and recommendations by site. We define that output first, with your people, in your terms.

Then the process runs in four passes:

  1. Sort. Documents are identified by type, dated and grouped, and duplicates and unreadable files are set aside for a person.
  2. Extract. The defined facts are pulled from each document into your schedule, each one linked to its page.
  3. Flag. Anything outside what the team told us to expect, such as an unusual clause, a missing document or two sources that disagree, is highlighted.
  4. Check. A reviewer works through the flags and samples the rest, following links to the source wherever they need to. A named person signs off the finished schedule.

The finished schedule is saved as structured data, so the next piece of work for that client starts from it rather than from the documents again.

What do reviewers do once the reading is done for them?

Junior staff move from reading every page to checking extracted facts against their source and investigating flags, which is closer to the judgement they are training for. Senior staff review a schedule where every point can be verified in one click, so they stop re-reading.

Turnaround changes too. A review that used to hold up the advice for a week can be ready for checking much sooner, which matters when a client or a counterparty is waiting.

Which firms have the biggest document sets?

Law firms reviewing due diligence and disclosure bundles. Property and asset managers working through lease portfolios and service charge records. Environmental consultancies reading historical reports and regulator records. Surveying practices gathering title, planning and building information. Insurance brokers reading existing wordings and claims histories before placing cover.

Where the review is of contracts against your own standard positions, see contract review. Where the need is to find things across the firm's own past work, see knowledge retrieval.

What does a 30 day rebuild of document review involve?

Week one is spent with the people who do reviews now: we take one review type, agree the schedule it should produce and gather past examples with their source documents. We also check where documents can safely be processed under your confidentiality terms. Weeks two and three build the extraction and flagging on those examples and then run it on a live job, with your reviewers comparing it against their own reading. In week four the team is trained and that review type moves over. Support follows for 30 days. The stages are described on how we work.