Accountants shouldn’t change how they work for software anymore. Software should adapt to them like Claude/OpenClaw. Their real value is judgment and how you approach the work.

Tax teams spend 2-4 hours per client manually entering K-1s, 1099s, W-2s into tax software. The worst part? K-1 footnotes. Dense, critical, completely manual. Most AI tools skip them. We don't. Combinely reads every detail - including horrible footnotes - and populates tax software automatically. Hours → Minutes. See it in action:

.@Combinely is building an AI coworker for tax teams. It uses firms’ existing tools to proactively handle inboxes, client questions, and return reviews— saving accountants 90 minutes every day. Congrats @Tom_inv & @ArthurGranacher! ycombinator.com/launches/Ou6…

LAWYERS AS VERIFIERS Use lawyers as AI verifiers. Here’s how. (1) First, tell the AI that it’s the named partner of the most capable law firm in your city, with expertise not just in corporate law, but tech startups, compliance, accounting, and the like. (2) Then, draft a memo to the AI describing your business goals in full. (3) Conclude by asking a list of specific questions. Make sure to add the catch-all question of “is there anything I may have overlooked.” Ask to include citations to specific sections of legal code, caselaw, or sample documents like SAFEs. (4) Next, run that prompt against several AIs (ChatGPT, Claude, Grok, Gemini, Perplexity) and assemble answers in a spreadsheet. Questions are rows, answers from each AI are in columns. (5) Now you have a survey of the space. Note in particular any new concepts or terms you’re unfamiliar with. You might rewrite the memo and resubmit it based on these initial answers. (6) Once you’re done, you have a draft game plan with citations. Next, ask each AI what legal docs you need, and put together first drafts of them individually with AI. These will have errors, but you will get the gist of what the final product looks like. (7) At this point you have the basic lay of the land. You can now go to each lawyer with the same memo you shared with the AI, ask them the same questions you asked the AIs, and compare their answers to the results of the AIs. (8) If the lawyer verifies that the AI produced the right answers, great. Then have the lawyer draft the final docs, possibly working from your AI-aided first drafts. Conversely, if a lawyer says the AI is wrong, that’s highly informative and they’ve added value. (9) This same strategy works with any service provider, from accountant to doctor. You use AI to structure your search, organize your thinking, and get initial answers to your questions. Then you run it by human specialists (lawyers, accountants, doctors) to verify whether the AI was right and complete the final step. (10) To be clear, this approach still requires skill on your end (in terms of prompting) and skill on their end (in terms of via verifying). Because AI doesn’t do it end-to-end, it does it middle-to-middle. But this approach does mean you no longer need to spend billable hours to learn the basics. And that means a scrappy startup can save money on lawyers.

Our new Sonar Resources page is live. See what others are building, read articles, and register for upcoming events.

the american mind cannot comprehend ⁦@Tom_inv⁩ ⁦@ArthurGranacher⁩ ⁦@LudvigSwanstrom⁩

Thanks to @perplexity_ai and the entire Sonar team for hosting an incredible event in SF with a packed house and amazing energy! We had the opportunity to showcase how Combinely leverages the Sonar API to build our AI coworker for accountants - tackling everything from answering complex client queries to producing and reviewing work.

Enjoyed this discussion with @t_blom and @dflieb!

Accountants aren't vanishing, they're upgrading: ↳Data interpretation over data entry ↳Client advisory over compliance checks ↳Business insights over bookkeeping ↳Strategic planning over transaction processing And 79% of new graduates prefer joining firms that actively use AI. The future belongs to those who treat AI as a partner - not a competitor.

🚀 @Combinely launched! AI Coworker for Accountants "AI-coworker that helps accountants review & produce work." 🌐 fondo.ai/430ZY0l Founded by @Tom_inv & @ArthurGranacher

🚀 @Combinely launched! AI Coworker for Accountants "AI-coworker that helps accountants review & produce work." 🌐 fondo.ai/4lWiLT9 Congrats @Tom_inv & @ArthurGranacher!!

🚀 @Combinely launched! AI Coworker for Accountants "AI-coworker that helps accountants review & produce work." 🌐 fondo.ai/430ZY0l Founded by @Tom_inv & @ArthurGranacher

Congrats to @Tom_inv and @ArthurGranacher on launching Combinely! These two are using AI to help accountants save time on 'quick' questions from clients that consume 30-40% of every working day. They're solving a problem they understand well thanks to Tom's accounting experience.

"Will our client data end up in ChatGPT?" This is the #1 question accounting firms ask us before adopting AI. But what we've noticed: as firms start using the software and see its value, security concerns often diminish - though larger firms with dedicated security teams maintain higher standards throughout. For accounting AI to succeed, It must be trustworthy as client information security isn't negotiable.

Most AI accounting tools miss the nuances within workflow automation. To deliver great accounting automation, we need to build deep knowledge at the firm, client, and industry levels.

AI will transform accounting – perfect example where there aren’t enough people to do the work, with 75% of CPAs retiring in the next decade

Most accounting AI startups integrate with QuickBooks first. We started with Emails. This approach gives us two key advantages. 1/ Emails contain the "why" behind every decision. Context matters before numbers. 2/ Email integration works globally with any accounting software. No regional limitations. When you capture context first, automation actually delivers on its promise.