Key takeaways
- Automate the structure of engagement letters and forms; keep lawyer judgment on scope, conflicts and unusual fee terms.
- Templates must capture what the fee rules require, including the signed writing and closing statement for contingent fees[1].
- Boilerplate consent to AI use in an engagement letter is not enough for informed consent under ABA Formal Opinion 512[2].
- Test every template on closed matters and keep a version log before staff use it live.
Where automation saves time and where it adds risk
Engagement letters and routine forms are ideal for automation because most of each document is identical from matter to matter. The client's name, the matter description, the fee terms and a handful of options change; the rest does not. A good template turns a 30-minute drafting job into a few minutes of data entry and a careful read.
The risk is that automation makes it easy to send a document nobody really read. A scope paragraph that fits most divorces may not fit one with a business to value. A fee clause written for hourly work can slip into a flat-fee matter if the wrong option is ticked. The aim is not to remove lawyer review. It is to move review to the few places where it matters and make everything else consistent.
Sort your documents into three groups
| Document | Group | Why |
|---|---|---|
| Intake questionnaire | Automate fully | Structured questions with no legal judgment in the output |
| Non-engagement letter | Automate fully, with a few approved variants | The wording should be identical every time; the facts are minimal |
| Hourly engagement letter | Automate with lawyer review | Scope and exclusions need a human read |
| Flat-fee agreement | Automate with lawyer review | Defining what the flat fee covers is the judgment call |
| Contingent fee agreement | Automate with lawyer review | State rules on required terms and percentages vary |
| Conflict waiver | Keep largely manual | Informed consent depends on the specific conflict |
| Court forms | Automate with review | Forms change; the data must match the pleading |
| Closing statement | Automate with review | The math must tie to trust records |
Start with the first two groups. They deliver quick wins with little risk, and building them teaches you the template tool before you tackle documents where an error has consequences.
What the fee rules ask your templates to capture
The ABA Model Rules set the baseline most states follow. Rule 1.5(b) says the scope of the representation and the basis or rate of the fee and expenses must be communicated to the client, preferably in writing, before or within a reasonable time after the representation starts. Rule 1.5(c) requires a contingent fee agreement to be in a writing signed by the client, stating the method for calculating the fee, the percentages that apply at each stage, the expenses deducted from any recovery and whether they come off before or after the fee is calculated[1]. Your state may require more, so build the template from your state's rule, not the model.
- Scope: what the firm will do, and a separate list of what it will not do.
- Fee type: hourly, flat, contingent or hybrid, with only the matching clauses included.
- Rates or amounts for each person who may bill, or the flat fee and what it covers.
- Retainer and trust handling, including when funds move from trust to operating.
- Expenses: which ones, how they are billed and, for contingent matters, the before-or-after rule.
- Billing frequency, payment terms and what happens if invoices go unpaid.
- Termination by either side and file retention after the matter ends.
- Signature blocks for every client, with the date captured automatically by the signing tool.
Build a template in seven steps
Collect the variants
Gather the last 20 engagement letters the firm actually sent. Highlight every place they differ. Those differences are your variables and conditions.
Write a clean master
Agree one master version per practice area with the responsible lawyer. Remove clauses nobody can explain.
Name the variables
Use clear names such as client_full_name or matter_scope, and decide which come from intake and which are typed at drafting time.
Add conditions sparingly
Each condition (fee type, number of clients, trust retainer or not) should switch whole clauses in or out. Avoid conditions inside sentences, which produce awkward grammar.
Set review gates
Mark the scope paragraph, any custom fee terms and conflict language as lawyer-review fields that cannot be skipped before sending.
Test on closed matters
Generate the letter for five closed matters with different facts and compare each to what was actually sent.
Version and log
Give each template a version date, keep a short change log, and retire old versions so nobody uses them.
Engagement letters and AI tools
Some firms now add a clause about AI tools to their engagement letters. ABA Formal Opinion 512 warns that merely adding general, boilerplate provisions purporting to authorize the lawyer to use generative AI is not sufficient to obtain informed consent where consent is required, for example before entering client information into a self-learning tool[2]. If your firm needs consent for a particular use, explain that use, its risks and its benefits in plain language, separately from the standard terms.
Billing clauses need the same care. California's State Bar guidance says hourly billing must reflect time actually spent, recognizes that generative AI can reduce that time and suggests flat fees for more predictable work[3]. Florida Bar Opinion 24-1 also addresses billing alongside confidentiality and advertising when lawyers use generative AI[4]. If automation or AI makes a task faster, the fee terms in your template should reflect that honestly.
Who should draft, and what their time costs
Once templates exist, the person who generates documents can change. In May 2025 the median annual wage was $159,670 for US lawyers[5], $62,890 for paralegals and legal assistants[6] and $55,570 for legal secretaries and administrative assistants[7]. Moving the generation step to trained staff and keeping lawyer time for the review gates is usually where the real savings are.
Example: a solo estate planning lawyer
Example (fictional): a solo estate planning lawyer sends about 15 engagement letters a month, mostly flat-fee packages with three options. Before automation, she edits a Word file from a previous client, and twice in one quarter a letter goes out with the wrong package price. She builds one template with a single condition for the package and a review gate on scope. Her assistant generates each letter from intake data; she reads only the scope paragraph and the package line before it goes out for signature. Drafting time falls from about 25 minutes to five, and the wrong-price error stops because the price now comes from one field.
Quality checks before an automated letter goes out
- Names: every client and the matter name match the conflict check record exactly.
- Fee type: only one fee structure appears, and its clauses are complete.
- Scope and exclusions: read by the responsible lawyer for this matter, not assumed from the template.
- Placeholders: search the output for brackets, underscores and variable names before sending.
- Numbering and headings: no gaps or orphan headings where a clause dropped out.
- Signature blocks: one for each client who must sign, with the right names.
Keep a one-page checklist beside the template and require the sender to tick it. Most automation errors are caught in the last 60 seconds or not at all.
Mistakes that make automation backfire
- Automating a document nobody has agreed on. Fix the master first, or you will produce inconsistent letters faster.
- Too many conditions in one template. Past a handful, split it into separate templates by practice area or fee type.
- Variables that pull from the wrong contact role, such as the opposing party's address landing in the client block.
- Letting staff edit the generated document freely, so the template and the letters drift apart over time.
- No owner. Templates without a named owner go stale within a year.
Measure whether it worked
Pick two measures before you start and check them after 60 days: minutes from completed intake to engagement letter sent, and the number of letters that needed correction after sending. If drafting time falls but corrections rise, the review gates are in the wrong place. If neither moves, staff are probably still copying old letters, and the fix is training and a firm rule that letters come only from the template.
Keep templates current
Templates decay. Rates change, a new partner joins, a court revises a form, a rule is amended. Name one owner for each template set and schedule a review at least twice a year, plus an immediate review when your state amends its fee rules or a court changes a form. When you update a template, regenerate one test document and compare it with the previous version so you can see exactly what changed.
Frequently asked questions
Can I automate contingent fee agreements?
Yes, but keep a lawyer review gate. The agreement must meet your state's version of Rule 1.5(c), which in the model rule requires a writing signed by the client with the fee method, percentages and expense treatment[1].
Which documents should a small firm automate first?
Intake questionnaires and non-engagement letters. They are high volume, low judgment and teach you the template tool before you build fee agreements.
Do I need special software for document automation?
Most practice management systems include templates, and some firms use stand-alone tools. Choose the one your staff can maintain without outside help.
Is a standard AI clause in my engagement letter enough?
Not where informed consent is required. ABA Formal Opinion 512 says boilerplate provisions are not sufficient[2].
References
American Bar Association (2025). Rule 1.5: Fees, ABA Model Rules of Professional Conduct (archived copy) (opens in a new tab) โ
American Bar Association, Standing Committee on Ethics and Professional Responsibility (2024). Formal Opinion 512: Generative Artificial Intelligence Tools (archived copy) (opens in a new tab) โ
The State Bar of California, Standing Committee on Professional Responsibility and Conduct (2026). Practical Guidance for the Use of Generative Artificial Intelligence in the Practice of Law (opens in a new tab) โ
The Florida Bar (2024). Florida Bar Ethics Opinion 24-1 (opens in a new tab) โ
U.S. Bureau of Labor Statistics (2025). Lawyers: Occupational Outlook Handbook (opens in a new tab) โ
U.S. Bureau of Labor Statistics (2025). Paralegals and Legal Assistants: Occupational Outlook Handbook (opens in a new tab) โ
U.S. Bureau of Labor Statistics (2025). Secretaries and Administrative Assistants: Occupational Outlook Handbook (opens in a new tab) โ