Key takeaways
- Automate documents that are frequent, structured and low in variation first: client care packs, routine letters and standard forms.
- Value the time saved at the right rate: much drafting is done by fee earners whose time costs about twice that of support staff.
- Client care packs are a strong first project because the Code of Conduct requires complaints and costs information at engagement.
- Every automated precedent needs an owner, a review date and version control, or errors spread to every matter.
What document automation is, and what it is not
Document automation means building a template once, with rules, so that a document can be produced from answers to a short questionnaire or from data already held in the matter record. The client's name, the property address, the fee estimate and the right optional clauses are inserted automatically, and paragraphs that do not apply are left out. Done well, a fee earner checks and adjusts a document instead of rebuilding it from the last similar file.
It is different from generative AI drafting. Rules-based automation is predictable: the same answers always produce the same document, and every word can be traced to an approved precedent. Generative tools write new text and can be useful for first drafts, but they need closer checking. Most firms gain more, sooner and with less risk, from rules-based automation of their standard documents than from experimenting with open-ended drafting.
Where automation saves real time
The best candidates are documents you produce often, that follow a clear structure and where the variations can be expressed as rules. Bespoke advice letters and complex agreements are poor first choices, because every one is different and the rules become unmanageable.
| Document | How often | How much it varies | Automation fit |
|---|---|---|---|
| Client care letter and terms of business | Every new matter | Low to medium: fee basis, practice area, team | Excellent |
| Conveyancing standard letters | Many per transaction | Low | Excellent |
| Simple wills and lasting powers of attorney | Frequent in private client | Medium: family structure, gifts, executors | Good, with careful rules |
| Employment settlement agreements | Regular | Medium | Good for the standard sections |
| Court forms and directions questionnaires | Regular in litigation | Low | Good |
| Bespoke advice letters | Every matter | High | Poor; use clause banks instead |
Delay is a serious client issue, not only an internal one. Delay or failure to progress made up 22% of complaint types brought to the Legal Ombudsman in 2025/26[1]. Automation will not fix a backlog by itself, but cutting the time spent on standard correspondence frees fee earners for the work that actually moves a matter on.
Putting a value on the time
Who does the drafting matters when you value the saving. Median gross hourly pay in 2026 was £31.50 for solicitors and lawyers and £14.95 for legal secretaries[2]. Pay is only part of the cost of an hour, but the ratio is a useful guide: time saved by a fee earner is worth roughly twice time saved by support staff, before counting the fees that hour could earn.
Source: Office for National Statistics, 2026[2]
Show the data
| Item | Value |
|---|---|
| Solicitors and lawyers | £31.5 |
| Legal secretaries | £14.95 |
Example: a fictional firm opens 40 new matters a month. Its client care letters take a fee earner about 30 minutes each to adapt from an old file, and an automated version cuts that to about 10 minutes of checking. That frees roughly 13 fee-earner hours a month from one template alone.
Run this calculation for your top five documents before choosing a tool. It tells you which template to build first and gives you a baseline to measure against three months later.
Start with the client care pack
The client care pack is the strongest first project for most firms, because it combines high volume with regulatory content that must not be missed. The Code of Conduct requires clients to be told in writing at the time of engagement about their right to complain, and to be told in writing if a complaint has not been resolved within 8 weeks[3]. Clients must also receive the best possible information about how their matter will be priced, at engagement and as it progresses[3].
For services covered by the SRA Transparency Rules, published price information must include the key stages of the matter and likely timescales for each stage[4]. An automated client care letter can draw on that same information, so the letter matches the website. When a fee earner rebuilds a letter by hand from an old file, an outdated complaints paragraph or a missing costs estimate is easy to overlook; when the paragraph is controlled centrally, it is right every time.
Building your first template
Collect real examples
Gather ten recent versions of the document from different fee earners and note every variation.
Agree one approved precedent
A partner signs off a single master version, deciding which variations are genuine options and which were just habits.
Write the rules
List each question the template needs answered, the possible answers and which clauses each answer switches on or off.
Map existing data
Link fields such as client name, matter type and fee earner to the case management record, so nobody retypes them.
Build and test
Produce the document for at least ten test scenarios, including unusual ones, and have someone outside the project check each output.
Pilot with one team
Use it live for a month in one team, collect corrections and fix the rules before wider release.
Release and retire the old versions
Remove old copies from shared folders so the automated version is the only route.
Designing the questionnaire
- Ask each question once and reuse the answer across every document in the pack.
- Use plain choices rather than free text wherever possible, so rules work reliably.
- Put the questions in the order a fee earner would naturally think about the matter.
- Show a short note under each question explaining what the answer changes.
- Allow a free-text override for genuinely unusual cases, but log when it is used.
If the questionnaire takes longer to complete than editing the old file, people will stop using it. Keep it to the questions that change the document and pull everything else from the matter record.
Governance that keeps precedents safe
Automation multiplies whatever is in the template. A correct clause appears in every matter, and so does an error. Each precedent therefore needs a named owner, normally a supervising solicitor in that practice area, a review date at least once a year, and a log of changes with reasons. When the law, the Code of Conduct or your prices change, the owner updates the template and the change takes effect everywhere at once, which is one of the strongest arguments for automating in the first place.
If you use generative tools alongside your templates, keep confidentiality in mind. The Law Society's generative AI primer, aimed particularly at small and medium-sized firms, flags the risk of misuse or disclosure of confidential information[5]. Do not paste client documents into tools that are not covered by your contracts and your firm's policy.
Where automation goes wrong
- Starting with the most complex document because it is the most painful, and never finishing it.
- Letting each fee earner keep a personal variant, so the firm ends up with several masters again.
- Building rules nobody but the original author understands, with no documentation.
- Skipping the test scenarios, so the first errors are found by clients.
- Leaving old templates in shared folders, so people bypass the automated version.
Most of these failures come from treating automation as an IT project rather than a precedent project. The people who know what the document should say, usually the supervising solicitors, need to be in the room when the rules are written, and they need protected time to test the results. A tool chosen without them tends to produce templates that look right in a demo and wrong on a real file.
Example: a conveyancing team's first quarter
Example (fictional): a six-person residential conveyancing team automates three things in its first quarter: the client care pack, the standard letters sent at instruction, exchange and completion, and the completion statement covering letter. Each template pulls the property, parties and lender from the matter. The supervising solicitor owns all three and reviews them each January and whenever the firm changes its published prices.
After three months the team tracks two numbers: the minutes spent preparing each pack, and the corrections clients or other parties point out. Both fall, and the team moves on to lease extension letters. For the technology around these templates, see conveyancing technology, and for identity checks at the start of the matter, read digital ID verification providers. Hosting questions are covered in cloud vs on-premise legal software.
Frequently asked questions
What documents should a law firm automate first?
Frequent, structured, low-variation documents: client care letters and terms of business, standard conveyancing letters and routine forms. Leave bespoke advice letters until later, or use clause banks for them.
Is document automation the same as AI drafting?
No. Rules-based automation produces the same approved text from the same answers every time. Generative AI writes new text, which can help with first drafts but needs closer checking.
Who should own an automated precedent?
A supervising solicitor in the relevant practice area, with a review date at least yearly and a change log. Ownership matters because an error in a template spreads to every matter.
How do we measure whether automation is working?
Track minutes spent per document before and after, the number of corrections needed, and how many documents are still produced outside the template.
References
Legal Ombudsman (2026). Legal Ombudsman 2025/26 annual complaints data and insight (opens in a new tab) ↑
Office for National Statistics (2026). Earnings and hours worked, occupation by four-digit SOC: ASHE Table 14 (2026 provisional) (opens in a new tab) ↑
Solicitors Regulation Authority (2025). SRA Code of Conduct for Solicitors, RELs, RFLs and RSLs (opens in a new tab) ↑
Solicitors Regulation Authority (2025). SRA Transparency Rules (opens in a new tab) ↑
The Law Society (2026). Generative AI: the essentials (opens in a new tab) ↑