Scope Creep Is Free Legal Work, and a Vague Engagement Letter Is How It Starts.

Every hour of work that falls outside a clearly defined scope is an hour the firm either eats or fights the client over. Florida's fee rules are specific about what has to be in writing, and a thin engagement letter is where both the malpractice risk and the fee dispute begin. Defining scope at intake, and flagging the moment a matter drifts past it, is cheaper than either.

I have come to believe that engagement letter quality is the single most under-engineered workflow in Florida law firms. Partners spend real money on case management platforms — Clio, MyCase, PracticePanther — and then let a paralegal paste a two-paragraph retainer into a Word template and call it documentation. Industry write-off studies on legal billing realization regularly put unbilled or written-off time at 6 to 9 percent of recorded hours, and a meaningful share of that traces back to scope ambiguity nobody resolved at intake. When that matter runs into billing friction eight months later, the letter is exhibit A for the client and a liability for the firm.

What the Rules Require Before the Letter Goes Out

Rule 4-1.5 of the Florida Rules of Professional Conduct governs fees and costs. The reasonableness factors it enumerates are not suggestions; they frame the standard against which a disputed fee gets evaluated.

For contingency arrangements, the rule requires a written agreement that states the method by which the fee is to be determined, including the percentage or percentages that accrue to the lawyer in the event of settlement, trial, or appeal, along with litigation and other expenses to be deducted, and whether such expenses are deducted before or after the contingent fee is calculated. The closing statement requirement under Rule 4-1.5(f) is equally specific.

Rule 4-1.2 addresses scope of representation. The client and lawyer may limit the scope of representation if the limitation is reasonable under the circumstances and the client gives informed consent. "Informed consent" is not a signature on a boilerplate retainer. It is documented agreement to a defined boundary, and that boundary has to be specific enough that both parties know when they have crossed it.

Most engagement letters drafted under time pressure at intake satisfy neither rule at the level that matters when a fee dispute reaches The Florida Bar's Attorney Consumer Assistance Program. Fee-related grievances are a recurring ACAP complaint category year after year, which is reason enough to treat the engagement letter as a compliance artifact rather than a formality.

Intake as the Upstream Failure Point

The typical engagement letter workflow runs like this: a new matter opens in Clio or a comparable platform, a staff member selects a template by practice area, fills in the client name and fee amount, and sends it for signature. The scope section, if it exists at all, reads something like "representation in connection with the above-referenced matter." That phrase is operationally useless.

When the client calls three months in asking about a related contract dispute, or requests that the firm handle a probate matter that surfaced during a real estate transaction, the firm has no written record of what was and was not included. The associate who handled the call added a task. The partner signed off. Nobody generated a scope amendment.

The work happened. The bill arrived. The client disputed it.

This is how scope creep becomes free legal work, and it starts at intake, not at the billing stage.

Drafting the Engagement Letter from Matter Type

An AI agent integrated into the intake workflow operates differently from a template library. The distinction is functional: a template library gives the user a starting document; an agent assembles a document from structured inputs and includes the rule-specific clauses the matter type requires.

In practice, the agent receives matter type (residential real estate closing, Chapter 7 bankruptcy, commercial contract dispute), jurisdiction, fee structure, and any known scope limitations flagged during the intake interview. From those inputs, it produces a draft engagement letter that includes the Rule 4-1.5 required language for the fee structure selected, a defined scope section specific to the matter type, an explicit exclusions clause listing what the representation does not cover, and a process for handling out-of-scope requests.

For a contingency matter, the draft includes the Rule 4-1.5(f) closing statement framework. For a flat-fee transactional matter, it includes the specific deliverables that define completion. For an hourly litigation engagement, it defines the stage of proceedings covered and flags what happens if the matter proceeds to appeal.

The output is not a finished letter. It is a complete first draft that a reviewing attorney approves, edits, and sends. The agent compresses the drafting time; the attorney retains the professional judgment call. That distinction matters under Rule 4-5.3 and Florida Bar Ethics Opinion 24-1, which together establish that a lawyer's supervisory duties over nonlawyer assistance extend to generative AI tools used in the practice. Any automated drafting workflow that does not keep a licensed attorney in the approval chain creates its own bar exposure.

The Drafting-Error Risk, Addressed Directly

The honest counterargument to any AI-drafting workflow is that the agent itself can produce a non-compliant fee clause, a hallucinated provision, or a jurisdiction-mismatched paragraph. The reviewing attorney could miss it. The client signs. Now the firm has a defective engagement letter generated by a system rather than a tired paralegal, and the malpractice posture is arguably worse because the firm cannot claim the error was an oversight on a one-off document.

This risk is real and it is the reason the workflow has to be structured around three controls, not one. First, the agent draws clause language from a maintained library of attorney-approved Florida-specific provisions rather than generating fee terms from scratch. Second, the reviewing attorney works from a diff against the prior approved template for that matter type, so any deviation from known-good language is highlighted rather than buried in prose. Third, the approval action itself is logged with the attorney's identity and timestamp, which is what Rule 4-5.3 and Opinion 24-1 supervision look like in practice.

A firm running this stack is in a better position than a firm relying on a paralegal pasting into a Word template, because the failure modes are visible and the supervisory record exists. A firm running an unsupervised agent against a generic clause library is in a worse position than either. The control structure is the entire argument.

Scope Monitoring After the Letter Is Signed

Drafting a well-defined engagement letter solves the documentation problem at intake. It does not solve the operational problem of a matter that drifts past its defined boundary over a six-month timeline.

An agent embedded in the matter management layer can flag this in real time. Time entries, task descriptions, and client communications get compared against the scope definition in the engagement letter. The hard part is that scope definitions are written in natural language and time narratives are written in natural language, so the comparison is a semantic matching problem rather than a task-code lookup. This is precisely where an LLM-based agent outperforms the rules-based flagging built into legacy practice management platforms, which can only match against predefined activity codes.

When a time entry describes work that falls outside the defined scope categories, the agent surfaces a flag to the responsible attorney before the entry hits the pre-bill.

This is a win-win for the firm and the client. The firm catches unbilled scope expansion before pre-bill. The client gets either a scope amendment or a documented courtesy write-off. Neither path ends in an ACAP filing.

Clio, MyCase, and PracticePanther each publish developer APIs that expose matter, time entry, and document endpoints. The integration pattern described here is feasible against those APIs today; it is not yet a packaged product line in the way conflict-checking or e-signature has become. A firm evaluating this should expect to build or contract for the integration, not buy it off the shelf.

The Fee Dispute That Never Gets Filed

Fee disputes under Rule 4-1.5 do not begin with bad faith on either side in most cases. They begin with a client who did not understand what was included and a firm that cannot point to clear written terms that show the client agreed to something specific. The engagement letter is the document that either resolves that dispute before it starts or escalates it into a bar complaint.

ACAP's process allows complainants to raise concerns about attorneys for fee-related conduct. The Bar's own guidance on fee disputes emphasizes written documentation of the basis for the fee. A firm that can produce a signed engagement letter with defined scope, a documented scope amendment when the matter expanded, and a pre-bill review log showing the attorney approved all out-of-scope entries is in a materially different position than a firm that produces a two-paragraph retainer and a stack of time entries.

The agent-generated letter functions as both a fee protection instrument and a record that survives bar review. Those two outcomes are not separable; they come from the same document produced at the same moment in the intake workflow.

Pressure-Testing a Vendor Offering Engagement Letter Automation

If you are evaluating a platform that promises to automate engagement letter drafting or scope monitoring for your Florida practice, four questions will tell you whether it is built for your actual risk profile:

1. Does the system's output include jurisdiction-specific language tied to Florida Bar Rules 4-1.5 and 4-1.2, or does it produce generic legal-industry templates that require substantial attorney revision to meet Florida's written-fee-agreement requirements?

2. Where does the attorney approval step sit in the workflow, and is it logged in a way that satisfies Rule 4-5.3 and Ethics Opinion 24-1 supervisory obligations if a bar complaint references the document?

3. How does the scope monitoring function handle ambiguous time entries — entries that could fall inside or outside the defined scope depending on context — and does it surface those for attorney review rather than auto-categorizing them?

4. If the firm uses Clio, MyCase, or another existing matter management platform, does the integration write the scope definition back into the matter record in a way that the platform's native reporting can surface, or does it create a parallel data silo that falls out of sync with the live matter?

A vendor that cannot answer all four questions with specificity is selling a product built for a different industry's compliance environment, and a Florida firm's bar exposure is not generic.