The Crossing Report — Special Edition #7

ChatGPT for Professional Services Firms: A Practical 2026 Guide

Published March 16, 2026 · By The Crossing Report · 10 min read

Professional services firms can use ChatGPT for document drafting, client-ready summaries, meeting prep and follow-up, research synthesis, and proposal writing — tasks that currently eat 10–15 hours per professional per week. According to a 2025 LexisNexis survey, 70% of attorneys already use AI tools, but most use them personally and inconsistently, not as part of a firm-wide workflow. The gap isn't adoption — it's systematization. This guide covers exactly what ChatGPT can do for accounting, law, and consulting firms, the specific prompts that produce usable output, and the risks you need to manage before expanding use across your team.

Why Professional Services Firms Are Turning to ChatGPT

The 2025 LexisNexis survey finding deserves a second look: 70% of attorneys use AI tools — and most of them are using ChatGPT, not specialty legal AI. Not Harvey. Not Clio Duo. ChatGPT.

The same pattern shows up in accounting. CPA Trendlines reported that AI adoption in accounting firms jumped from 9% in 2024 to 41% in 2025. A significant share of that adoption is general-purpose AI like ChatGPT before firms moved to practice-specific tools.

What does that tell you? Firm owners and their professionals are already using ChatGPT — just not consistently, not at the firm level, and often without the right account tier or any guiding policy.

The gap isn't a technology problem. It's a systematization problem.

The adoption data

70% of attorneys use AI tools (LexisNexis 2025). AI adoption in accounting firms jumped from 9% to 41% in one year (CPA Trendlines). Most adoption is general-purpose ChatGPT, not practice-specific tools. The problem is not adoption — it is systematization.

What Professional Services Firms Actually Use ChatGPT For

Before getting into firm-type specifics, here are the five categories of work where ChatGPT consistently delivers for professional services firms. These are the tasks worth systematizing first.

•Document drafting and editing: First drafts of engagement letters, client agreements, internal policies, and standard correspondence. A realistic benchmark across firm types is 40–60% reduction in first-draft time.
•Client-ready summaries and memos: Converting complex source material — a tax return, a contract, a financial report, a deposition transcript — into a clear, client-facing summary. High-value and high-frequency in every firm type.
•Research synthesis (NOT primary research): Synthesizing and organizing information from sources you've already gathered. ChatGPT is useful for organizing and explaining what you already know; it is not reliable for finding or verifying legal precedent or tax authority.
•Meeting prep and follow-up notes: Drafting agendas and turning raw meeting notes into action-item summaries. When paired with an AI meeting recorder (Fathom, Otter.ai), this eliminates 2–3 hours of administrative work per professional per week.
•Proposal and pitch writing: Drafting service proposals, capability statements, RFP responses, and pitch decks. Professional services firms typically spend significant unbilled time on business development writing — ChatGPT shortens the drafting phase without reducing final quality.

ChatGPT for Accounting Firms — Specific Use Cases

Accounting firm adoption is ahead of most professional services sectors. Firms that are using ChatGPT effectively have settled on three workflows that produce consistent, high-quality output.

•Client-facing tax summaries from complex returns: Take a completed tax return and produce a plain-English explanation for the client — what changed from last year, what drove the outcome, and what they should do differently next year. A prompt template built around your firm's standard format produces a strong first draft in under two minutes.
•Engagement letter drafting: Standard engagement letters for recurring client work — audit, review, compilation, tax preparation — are ideal ChatGPT use cases. A well-built prompt template that accepts client name, service type, and key variables produces a draft that requires only review and signature setup.
•Turning financial reports into executive summaries: For clients who receive monthly or quarterly financial reports, a one-page executive summary translating the numbers into plain-language insight is a high-value deliverable that most firms currently produce manually. ChatGPT can draft this from the report in a few minutes.

For a deeper look at how accounting firms are approaching AI more broadly, see The AI Adoption Gap in Small Accounting and Law Firms.

Key takeaway

The three highest-ROI accounting ChatGPT workflows: client tax summaries, engagement letter templates, and financial report executive summaries. All reduce writing time by 40–60% on high-frequency, standardized tasks.

ChatGPT for Law Firms — Specific Use Cases

Law firm AI adoption is complicated by ethical obligations — but the ABA's Model Rule 1.1 comment on competence now explicitly identifies technology, including AI, as part of the competence standard. Prompt engineering is a basic professional competency for attorneys in 2026.

•First-draft contract review summaries: For a standard commercial contract — vendor agreement, service contract, NDA — ChatGPT can produce a structured summary of key terms, flagged provisions, and unusual clauses in minutes. The attorney still reads the contract and exercises judgment. But the first-pass summary reduces review time significantly for routine agreements.
•Client intake questionnaire drafting: Drafting or updating intake questionnaires for new practice areas or matter types is time-consuming work that ChatGPT handles well. Give it your existing questionnaire as a template, describe the practice area and the information you need, and it produces a solid first draft in seconds.
•Research memo formatting: When an attorney has already done the research — reviewed cases, pulled authority, synthesized findings — ChatGPT can convert rough notes into a structured research memo in the expected format. This is synthesis and formatting work, not research. The distinction matters: ChatGPT should never be your source of legal authority, only your editor for organizing it.

A note on data handling: before using ChatGPT for any client-adjacent work, review AI Data Security for Law Firms and your state bar's guidance on AI use and client confidentiality. See also our overview of AI Regulation and Compliance for Professional Services in 2026.

Key takeaway

The three highest-ROI law firm ChatGPT workflows: contract review summaries, intake questionnaire drafting, and research memo formatting. ChatGPT should never be used for legal research or case law — use Westlaw or LexisNexis for binding authority.

5 Prompts, Consulting Use Cases, and What ChatGPT Cannot Do

In our full Special Edition #7 analysis, we cover:

1.ChatGPT for consulting firms — proposal writing, slide narrative drafts, and meeting facilitation prep (the writing-heavy use cases where consultants see the most time savings)
2.5 copy-paste-ready prompts your firm can use this week — client summaries, engagement letters, meeting follow-ups, proposals, and contract review summaries
3.What ChatGPT cannot do for professional services firms — legal research, tax compliance checking, and the client confidentiality risks you need to manage before rolling out to staff
4.When to move from ChatGPT to a specialized tool — the signals that mean you've outgrown general-purpose AI

Premium Content

The Complete ChatGPT Guide for Professional Services Firms

Get the full consulting use cases, 5 copy-paste prompts, the limitations guide, and the rollout checklist for deploying ChatGPT firm-wide. Premium subscribers also get every future issue in full.

$19/month · Cancel anytime · First issue free

FAQ — ChatGPT for Professional Services Firms

Q: Is ChatGPT safe for client data?

A: Not the free consumer version. ChatGPT's free and Plus tiers use conversation history for model training by default, which means client details you paste in could potentially appear in other users' responses. For any work involving confidential client information, use ChatGPT Enterprise or ChatGPT Team — both include a data processing agreement that prohibits OpenAI from training on your inputs. The practical rule: if it would violate your client confidentiality agreement to email it to a stranger, don't paste it into a consumer AI tool.

Q: What's the difference between ChatGPT and Harvey?

A: ChatGPT is a general-purpose AI assistant built by OpenAI. Harvey is a legal AI tool built specifically for law firms, trained on legal data and designed for tasks like contract review, due diligence, and legal research synthesis. Harvey costs significantly more and targets mid-to-large law firms. For a 5–20 person law firm, ChatGPT Enterprise is usually the right starting point — it handles drafting, client summaries, and intake work without the premium legal AI price tag. Move to Harvey or Clio Duo when your team has maxed out what a general-purpose tool can do.

Q: Do I need a firm AI policy before using ChatGPT?

A: Yes — before any client-adjacent work touches an AI tool. You don't need a 50-page policy document. You need three things: (1) a rule about which tier of ChatGPT is approved for client work (enterprise only), (2) a list of what staff should never paste into any AI tool (SSNs, privileged communications, raw financial data), and (3) a disclosure position — do you tell clients when AI assisted on their work? Firms that skip this create liability exposure when a staff member uses a free consumer account on a client matter. A one-page policy takes an afternoon and eliminates that risk. See our guide to Engagement Letter Compliance in the AI Era for the relevant disclosure language.

Q: How do I train my staff on ChatGPT?

A: Don't start with a training program. Start with one workflow. Pick the highest-volume repetitive writing task your team does — engagement letters, client update emails, meeting summaries — and write one prompt template for it. Have every team member use that template for 30 days. After 30 days, you'll have proof that it works (time saved, quality maintained), which makes the next workflow adoption almost automatic. Formal training before real use rarely sticks. Real use followed by reflection does.

Your Next Step

Pick one prompt from the list above — whichever matches the most repetitive writing task your team does this week — and use it on one real piece of work. Don't evaluate it theoretically. Run it. Edit the output. See what you actually get.

The firms that are pulling ahead on AI didn't start with a strategy. They started with one task and built from there.

Get the prompts and the full guide with a premium subscription.