How a 15-person accounting firm went from 'AI is overhyped' to saving 22 hours per week in six months.
When Sarah Chen first heard about using AI in her accounting firm, she rolled her eyes.
"I've been doing this for 18 years," she told me. "My team knows what they're doing. We don't need a robot to do our jobs."
Six months later, her 15-person firm saves 22 hours per week with AI — and Sarah is the biggest advocate in her professional network. This is her story.
Month 1: The Reluctant Experiment
Sarah's office manager, Marcus, kept bringing up AI tools. "Just let me try it for the client status emails," he said. "If it doesn't work, I'll stop."
Sarah agreed — mostly to get Marcus to stop talking about it.
The task: Every Friday, Marcus spent 3 hours drafting personalized status update emails to 45 clients. Each email summarized their current engagement status, upcoming deadlines, and any items needing attention.
The AI approach: Marcus built a prompt template in Claude that took client data and generated a first draft. He still reviewed and edited each email, but the drafting time dropped from 3 hours to 45 minutes.
Week 1 savings: 2.25 hours.
Sarah was unimpressed. "So you saved two hours on emails. That's not going to change the business."
Month 2: The Snowball
Marcus documented his email workflow and showed it to two other team members. They started adapting the approach:
Tax preparer Lisa used AI to generate first-draft client communications explaining tax situations in plain English. Previously, she spent 20+ minutes on each complex letter. Now it took 5 minutes of editing.
Senior accountant Dev started using AI to draft engagement letters and proposal responses. He cut his admin time by 6 hours per week.
Month 2 total savings: 11 hours/week across three people.
Sarah noticed the team leaving earlier on Fridays. "What's going on?" she asked Marcus.
"We're finishing our work faster," he said. "Same quality. Less time."
Month 3: Sarah Gets Curious
The turning point came when Sarah had to prepare for a client meeting — a complex situation involving a business restructuring. Normally, she'd spend 2 hours reviewing documents and preparing her talking points.
"Show me how to use that AI thing," she told Marcus.
Together, they fed the relevant documents into Claude and asked it to summarize the key financial considerations, identify potential tax implications, and draft a meeting agenda with discussion points.
Time to prepare: 35 minutes instead of 2 hours.
"The summary was actually good," Sarah admitted. "I still had to add my professional judgment and check the tax code references, but it did 70% of the legwork."
Month 4: The System Emerges
By month four, the team had a shared prompt library with 23 templates covering:
- Client status emails (5 variants)
- Tax situation explanations (8 templates by complexity)
- Engagement letters (3 types)
- Meeting preparation (4 formats)
- Internal process documentation
- Training materials for new hires
They created a simple rule: Any task that gets a prompt template must also get an SOP. This meant every AI workflow was documented and shareable.
Month 4 total savings: 18 hours/week across 8 team members.
Month 5: The Revenue Impact
Something unexpected happened. Because the team was spending less time on admin work, they had more time for client-facing activities.
Lisa started proactively reaching out to clients about tax planning opportunities she'd previously been "too busy" to pursue. Dev took on three new clients that he would have turned away before.
Sarah calculated that the additional billable hours and new client revenue generated roughly $12,000 in incremental monthly revenue — directly attributable to the time freed up by AI.
"Okay," Sarah told Marcus. "I'm a believer."
Month 6: The Culture Shift
The firm now has a monthly "AI Hour" — a 60-minute session where team members share new prompts, troubleshoot issues, and evaluate new tools.
Every new hire goes through a 2-hour AI orientation. The prompt library has grown to 40+ templates. Two team members have earned Basecamp AI certifications.
Month 6 total savings: 22 hours/week across 12 team members.
Estimated annual impact: $180,000+ in time savings and incremental revenue.
What Sarah Would Tell Her Past Self
I asked Sarah what advice she'd give to the version of herself from six months ago. She didn't hesitate:
"Stop thinking of AI as replacing your team. It's a tool that makes your team faster and better at what they already do. You're not hiring a robot — you're giving your best people superpowers."
She paused, then added: "And listen to Marcus sooner."
The Playbook
Sarah's firm followed a pattern that works for any small business:
The firms that struggle with AI adoption skip steps 2 and 5. They use the tools but don't build the system. Six months later, they've forgotten half their prompts and can't prove the ROI.
Sarah's firm can prove every dollar. And that's why they're doubling down.
Names have been changed to protect client confidentiality. Metrics are based on actual tracked results with client permission.