AI Marketing
AI for Accountants: Practical Uses and Honest Limits
AI for accountants and small firms: six practical uses, which client data never goes into a chat window, and the data protection rules that still apply.

AI for accountants is useful today for the writing that surrounds the numbers: answering recurring questions, summarizing rule changes, building document checklists and creating client content. What it should not do is make decisions or receive client data without a contract.
In this profession the limits matter more than the tool. You handle payroll, tax IDs, bank details and sensitive personal situations. So this guide covers both: what to do, and what not to.
Demand is real: "ai for accountants" gets about 6,600 searches a month in the United States (Pengu Insights, 18 Nov 2026). A lot of firm owners are asking the same question.
Where AI for accountants pays off first
It pays off on tasks that repeat every week, are mostly text, and get reviewed by a person before they leave the firm.
In the AI workshops I teach at EAE Business School in Madrid, my advice is always the same: do not start with the hardest work (the ledger, the tax return), start with the most boring. The fortieth email explaining which documents a client needs. The note about a rule change. The spreadsheet formula nobody remembers.
That work has three advantages as a starting point: it happens often, it rarely needs personal data, and an error is easy to spot on review. Bookkeeping categorization and reconciliation can come later, once you know how the tool behaves and who checks its output.
Six practical uses for a small firm
- Standard replies to frequent questions (20 min per template). Write the question without client details and ask for a clear draft. Review it, save it, reuse it.
- Summaries of rule changes (15 min each). Paste the published text from the IRS or your state agency and ask for a plain-language client summary. Check every statement against the original.
- Document checklists per engagement (30 min each). New business setup, estate work, a move to another state. One checklist you reviewed saves rounds of back and forth.
- Spreadsheet formulas and macros (10 min per problem). Describe the sheet with invented data and ask for the formula. Test it before you trust it.
- Content for your website and newsletter (1 h a month). Deadlines, changes, common questions. That content is what helps you show up when someone searches for an accountant in your town.
- Document extraction (contracted tool only). Invoices or receipts turned into rows. This one does involve personal data, so it only happens inside a tool under contract, with every batch reviewed.
Can accountants put client data into ChatGPT?
Not on a personal account, and on a business account only under a contract that fits your security obligations.
In the United States, the FTC's guide to the Safeguards Rule lists tax preparation firms among the financial institutions it covers. It requires a written information security program, and it says your contracts with service providers must spell out your security expectations, with providers monitored and reassessed. An AI vendor that handles customer information for you is a service provider in practice.
If you serve clients in the EU, the GDPR adds its own rule in Article 28: use only processors that provide sufficient guarantees, under a contract that sets out the subject, duration, purpose, type of data and obligations. That part is EU-specific.
What the vendors say on their own pages, checked on 18 Nov 2026:
- OpenAI states on its business data page that, by default, it does not use data from ChatGPT Enterprise, ChatGPT Business, ChatGPT Edu or its API platform for training, and that it supports customers in signing a data processing addendum.
- Anthropic states in its privacy center that, by default, it does not use inputs or outputs from its commercial products, such as Claude for Work and the Anthropic API, to train its models.
Notice what that does not cover: free and personal accounts have their own terms. If your staff use their own logins, that is your first risk. This is not legal advice; check your setup with your compliance advisor.
What you should never hand over
Do not delegate anything you sign, file or bill as professional judgment.
- The tax position. AI can explain a rule. It does not know the client's full situation and it does not answer to them.
- The filing. No generated draft should reach a tax authority without a line-by-line review.
- Unchecked figures. Language models get numbers and dates wrong with total confidence. Every figure goes back to its source.
- The first conversation with a new client. People choose an accountant on trust, and that part is yours.
- Actions triggered by what a document says. If an assistant that can send emails or move files reads invoices and client messages, text inside those documents can try to give it instructions. Keep a person between what the model reads and what it does, especially anywhere money moves.
Which setup for which task
| Task | Personal data? | Where to do it | Review |
|---|---|---|---|
| Standard replies | No, once anonymized | Business plan or a data-free chat | Before saving the template |
| Rule summaries | No | Any assistant, with public text | Against the original |
| Document checklists | No | Any assistant | Once, by you |
| Spreadsheet formulas | No, with invented data | Any assistant | By testing it |
| Invoice extraction | Yes | Contracted tool only | Every batch |
| Website content | No | Any assistant | Before publishing |
If you are weighing assistants for a small team, Claude for small business walks through plans and setup.
How to start without a project
- Write a one-page AI policy (1 h). Approved tools, data that never goes in, who reviews. Add it to your written security program so it is not a separate document nobody reads.
- Pick one weekly task with no personal data (15 min). The one that repeats most.
- Run it for a month and time it (10 min a week). Before and after.
- Choose the next task with the number in front of you (30 min). If the first one saved nothing, the second will not either.
It is the same one-workflow method I describe in AI automation for small business, applied to a firm.
Next step
AI also changes how clients find you. When I tested ChatGPT on 18 Nov 2026 asking for a dental clinic and a divorce lawyer in Madrid, it recommended businesses with hundreds of Google reviews and cited ranking sites and directories. I have not tested accounting firms, but it is safe to plan as if the pattern holds. I cover how to get cited in LLM SEO.
My SEO + AI audit checks how your firm appears in Google and in AI assistants and gives you a list of fixes ranked by impact: €390 plus VAT, one-off. If you would rather train your team, my AI workshops run €97 to €149 per person.
Frequently asked questions
Can accountants use ChatGPT with client data?
Not on a personal account. If client data is involved, use a business plan with a data processing agreement, make sure it fits your written security program, and share only what the task needs. For most tasks the simplest route is to remove names, tax IDs and anything that identifies a client.
What can AI actually do for an accounting firm?
Draft replies to recurring client questions, summarize published rule changes in plain language, build document checklists, write spreadsheet formulas and produce content for your website or newsletter. Document extraction works too, inside a contracted tool with human review.
Will AI replace accountants?
It replaces tasks, not judgment. Language models still get numbers and dates wrong with total confidence, and clients pay for someone who answers for the result. The firms that benefit use AI on the repetitive writing and keep people on decisions.
Does the FTC Safeguards Rule apply to AI tools?
The FTC lists tax preparation firms among the financial institutions the rule covers, and it requires a written information security program and oversight of service providers. If an AI vendor handles customer information for you, treat it as a service provider and check the details with your advisor.
Where should a small firm start with AI?
With one weekly task that needs no client data, such as summarizing a rule change or drafting the standard reply to a frequent question. Time it for a month, then decide on the next one.
Sources

Christian Monge
Teaches AI workshops at EAE Business School, builds AI second brains with Obsidian and Claude Code, co-founded the competitive-intelligence SaaS Pengu Insights and publishes his Claude skills for free. Trains teams in English and Spanish from Madrid.
More about meSEO + AI audit
Want to know what your website is missing?
I review your site, your Google Business Profile and whether ChatGPT recommends you. Start with a free 90-second video, or go straight to the €390 audit.
See the SEO + AI auditKeep reading

AI Marketing
AI for Restaurants: Practical Uses and Honest Limits
AI for restaurants without the hype: menu descriptions, review replies, booking follow-ups, social posts and demand planning, plus the limits you should keep.

AI Marketing
AI Tools for Social Media: Create, Schedule, Measure
AI tools for social media for small businesses: what to use to create posts, schedule them and track competitors, with checked prices and a verdict for each.

AI Marketing
AI Video Ads for Small Businesses: A Practical Guide
How to make AI video ads for a small business: a step-by-step workflow, Reels ad specs, honest limits with faces and rights, and when platforms require a label.