Professional services · 12 people

Less repetitive work for an accounting firm

Illustrative examples based on hypothetical organisations. These are not real clients; figures are assumptions, not measured results.

915 h/year · Potential time recovered

The organisation

A 12-person tax, payroll and accounting firm with two partners, eight specialists and two administrators. It uses practice software, email, WhatsApp and shared folders.

The challenge

Quarterly reporting and tax-return season overwhelm the team. Much of the day goes on repeated questions and entering clients’ receipts and invoices one by one. In this scenario, several staff also use personal ChatGPT accounts to draft emails containing client data: a risk to address first.

What we would examine

1. Questionnaire for partners, specialists and administration.
2. A 90-minute review interview with a specialist from each area.
3. A second round to measure weekly enquiry and document volumes.
4. Tool testing with work enquiries and documents free of personal data.

Processes and proposed improvements

• Enquiries: 150 a week by email and WhatsApp, at six minutes each: 690 hours a year. A response library and assistant would prepare answers and drafts. A specialist would review and send them in half the time: 345 hours recovered.
• Documents: 400 weekly receipts and invoices at two minutes each: around 613 hours a year. Automated reading and classification, validated by a specialist, would target a 75% reduction: around 460 hours recovered.
• Weekly newsletter: four hours summarising tax updates: 184 hours a year. Assisted summaries and drafts would reduce this to under two hours a week: around 110 hours recovered a year.

First, bring data use under control

We would propose business accounts, EU data hosting, a data-processing agreement and an internal policy defining what data may be shared. Each tool’s terms, professional confidentiality and GDPR obligations would need checking; a business account alone does not guarantee compliance.

What we would avoid

A website chatbot answering tax questions without supervision. Mistakes can have serious consequences: AI would prepare drafts, but a specialist would always review and send them.

Proposed 90-day roadmap

1. Month 1 · Accounts and policy. Address the data risk from personal accounts first.
2. Month 2 · Documents. Prepare automated reading before the next quarterly deadline.
3. Month 3 · Enquiries and newsletter. Build the response library and draft templates.

What the figures mean

The 915 hours a year represent about €20,100 of capacity at €22/hour. Tools are estimated at €180/month (€2,160/year), leaving about €18,000 in time value after licences, before the Study, implementation and maintenance. This is not guaranteed cash savings. The estimate of just over two months to recover the Study and first year of tools is indicative, excludes implementation and depends on achieving the projected time savings.