I'm Charlie Barmore, a CPA, CFE, and CVA who builds software. Using AI is the easy part now. The hard part is putting it to work on the problems your firm has today, and figuring out how to implement it so it sticks. That is what I help firm owners do.
41% of accounting firms now use AI. Only 14% have a defined strategy for it.
Wolters Kluwer 2025 Future Ready Accountant Report; Thomson Reuters 2025.
AI is not magic, and it is not one tool. The real gains show up first in the work that is repetitive, document heavy, and bound by deadlines. These are four places it already helps today, not someday. They are also the ones I rebuilt for my own returns.
Every engagement starts with the same grind: requesting documents, following up, following up again. Portals with smart reminders and completeness checks cut the back and forth, so the work starts sooner and busy season compresses less. The judgment of what you actually need from a given client stays with you.
Hand keying source documents into your software is the most cited time sink in a firm, and the least valuable use of a preparer. Document extraction reads W-2s, 1099s, bank statements, and receipts and populates the software for you. It is the most mature use of AI in a practice today, and it still needs a human on the messy scans and the complex K-1s, which is exactly where your time should go.
Before any analysis happens, someone spends hours assembling, indexing, and cross referencing the file. Agents can sort, bookmark, and tie out the source documents into a first pass, review ready set, so the preparer opens a workpaper that is already organized instead of a folder of PDFs. This is the part I rebuilt for my own returns: intake, file structure, document parsing, and a drafted workpaper, before I touch it. See what I built →
Bookkeeping scales linearly with clients and eats margin. AI categorization and reconciliation suggestions handle the repetitive pass and flag the anomalies, so your team reviews exceptions instead of every line. The exceptions and the judgment calls still get human eyes.
Every firm is at a different point, so I do not sell a boilerplate build. Some firms just want to start: an hour on their real files, or one session to bring the whole team along. Others are ready to map the whole thing and build it. Everything past the hourly rate is scoped and priced in writing before it begins, and there is no pressure to move up the ladder.
Screen-share sessions where we build on your actual files, not slides. Finish your CLAUDE.md setup properly, standardize how your whole team uses a shared Claude account, or go straight at the workflow eating your week.
One live session for your whole staff, so everyone hears the same answers at the same time. We set the firm-level foundations together and each person leaves with a personal quick win already in motion, the hesitant ones included, without anyone being singled out.
A structured look at your workflows, your data, and your software to find where AI helps, where it does not, and what has to be true for it to be safe.
I map your workflows end to end, then build and deploy the specific tools your firm needs: a standardized Claude playbook for all of your staff, and your top bottleneck solved for good.
Continuing support as your firm's needs change, from new tools to refinements to a standing line to someone who gets the work.
Every engagement is priced to your firm: where you are in your AI journey, your size, and the scope we agree on. Fees are fixed in writing before any work begins, so there are no surprises.
I had the same problems you do, so I built for them: a return-prep pipeline that onboards a client, reads the documents, and drafts the workpaper before I open it, and a workflow that takes a prospect from the discovery call to a signed engagement. Both keep the judgment, and the sign-off, with me.
Whether or not we ever work together, the groundwork is free. I wrote the four documents every firm should have in place before it puts AI anywhere near client data, and I give them away.
What tools are approved, what is never allowed near client data, and who reviews the output. The piece almost no firm has yet.
Free template →Your written information security program, structured to the FTC Safeguards Rule you are already subject to as a preparer.
Free template →Consumer chatbots that train on your inputs are not the same as tools with no-training terms and a business agreement. The difference is your clients' SSNs.
What you tell clients, including where IRC §7216 requires their consent before tax data reaches a third-party tool, and what you confirm about a vendor before you trust it.
Free templates →Take them, adapt them, use them. If you want help applying them and solving the bottlenecks underneath, that is the engagement. Get the templates →
I am a partner in one community and the AI practitioner for another. Both are places where firm owners and practitioners work through AI adoption together, out loud, on real client work.
A weekly community for accountants putting AI to work in real client work. My lane is the defensible side: IRC §7216, the FTC Safeguards Rule, Circular 230, and the state board rules that decide what you can actually do with client data.
ailabforaccountants.com →A community for accounting, bookkeeping, and tax firm owners, with coaching, CPE, and honest conversation about running a practice. I am the AI practitioner there, the person members bring their AI questions to.
thecollaborationroom.co →Longer answers to the bigger questions: what an AI-native firm is, what an agent-native practice is, and what MCP means for accountants.
“Charlie assisted our firm in implementing AI-based quality controls to ensure that we can continue providing the best service to our clients possible. He was a huge help in fielding our questions, assessing the situation and recommending the best course of action for us.”
“We brought Charlie on as an independent technical reviewer on an AI-assisted tax return review system that we are building in-house. What’s helped the most is that he doesn’t take anything at face value … he pushes back on unproven claims, names exactly what still needs evidence before it counts as done, and keeps us honest about what’s genuinely ready versus what still needs work.”
“Charlie was a big help, and his follow-up was valuable too. The simple changes we discussed have already made a difference.”
Email me to set up a call, or use the form to tell me a little about your firm and what is slowing it down. I follow up within one business day. The first conversation is free and there is no pressure to move forward.