OpenClaw prepares a complex tax return
Daniel Vassallo used OpenClaw to analyze years of financial records and guide a complex TurboTax return, taking over browser entry and planning an amendment when an estimated K-1 arrived.
x.com
Original prompt
Shared by the source author. See source notes for context.
read everything in chronological order and figure out my 2025 tax situation.
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source author
- Daniel Vassallo
- Original account
- Read the source ↗
- Published
- 1 Apr 2026
- Source link checked
- 15 Oct 2026
- Reported result
- Vassallo reports filing after about six hours of research and preparation, with a 137-page reviewed draft and estimated partnership income still requiring later confirmation.
- Human involvement
- Vassallo understood the finances, recategorized a transaction, entered forms himself and reviewed the final draft; an amendment could be needed for a missing K-1.
A linked account supports attribution. It does not mean we independently reproduced its result.
Report an issue with this case
Open correction emailOpens your email app with the case link included. Nothing is sent automatically.
How the agent did it
- Provide historical returns, business agreements and financial records for analysis.
- Let OpenClaw reconstruct basis, identify forms and prepare the business accounts.
- Work through TurboTax manually when browser automation proves too slow.
Daniel Vassallo gave OpenClaw prior tax returns, business agreements, bank records and other documents to reconstruct a complex tax situation. The agent traced entity changes and cost basis, compared earlier filings and produced a summary of figures, required forms and open questions.
The next day, Vassallo worked through TurboTax with the agent's help. OpenClaw analyzed business transactions and prepared a profit-and-loss statement. Vassallo corrected one categorization. An attempted browser-driven entry was too slow, so he took over clicking while the agent explained fields and supplied values.
A major difficulty was double counting the gain from an installment sale. Several attempts were needed to coordinate the partnership and installment-sale sections. Vassallo reports reviewing a 137-page draft with the agent and filing the return, but partnership income had to be estimated because the actual K-1 was missing. He expected a possible amendment.
The outcome is a reported filing after substantial human involvement, not an audited tax result. Vassallo knew his finances, checked expected totals and handled software entry. His savings and penalty calculations are personal claims. The pending K-1 and difficult form interactions remain material limitations.
Prompt
Original prompt
Source: X · Daniel Vassallo
Frequently asked questions
Did OpenClaw operate TurboTax throughout?
No. Browser automation was too slow, so Vassallo clicked through the software while OpenClaw guided entry.
Was all source paperwork final?
No. The actual partnership K-1 was not yet available, and the author expected he might need to amend.
What difficult error needed correction?
TurboTax double counted installment-sale gain across sections until the workflow was adjusted over several attempts.
More details
Remaining steps
- Review the draft, address double counting and retain the missing K-1 as an unresolved item.
About this account
OpenClaw analyzes records, reconstructs entity history and basis, prepares a P&L and guides and reviews TurboTax entries.
Firsthand primary account opened and read; original publication date checked in primary metadata or visible timestamp. Outcomes are author-reported, not independently reproduced. Self-reported filing, not an independently audited return; K-1 estimated and amendment possible. Original prompt excerpt reproduced exactly from the source; excerpt only, not the complete prompt.
Credit: Daniel Vassallo
Source published 1 Apr 2026
Source checked 15 Oct 2026
Tools and topics
OpenClaw · TurboTax · PDF returns · Bank transaction records
Money, bills · Reported deployment
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.