TaxGPT

What do they actually do

TaxGPT is a web-based assistant used by tax professionals and finance teams to speed up research, drafting, document analysis, and return review. Users can ask complex tax questions and get answers with citations, generate first drafts of memos and client communications, upload returns/notices for analysis, and run an automated review on prepared returns that flags errors and potential savings. The product supports U.S. federal, all 50 states and territories, and Canada, with international handled mainly through treaties connected to the U.S./Canada Tax Research, Writer, Matrix, Agent Andrew, Document Analysis pages and help docs. Document ingestion currently accepts PDFs/images with stated limits (examples: up to 10 files and a 100‑page bundle) document feature and docs.

Firms typically schedule a short demo before activation, then onboard users and clients into a shared workspace so answers and drafts reflect client-specific context. Outputs (cited answers, memos, client letters, and return review reports) can be exported and edited, with firms keeping professional judgment “in the loop” demo/FAQ and FAQs. TaxGPT says it stores data on U.S.-based, SOC 2 Type II–compliant infrastructure and does not train models on customer data FAQs. The company reports seed funding of $4.6M and says it has been adopted by thousands of professionals (claims include “10,000 CPAs, tax lawyers, and enrolled agents serving 250,000 clients” and “30,000+ users”) seed blog and homepage.

Who are their target customer(s)

• Small CPA firm owner / principal: Spends many hours on research and drafting memos or IRS-response letters across many clients; needs faster, reliably cited output to hit deadlines and manage workload.
• Mid-size firm quality-control or review manager: Must check many returns for errors and missed savings; current line-by-line review is slow and inconsistent across teams.
• Tax attorney / senior specialist: Needs authoritative citations and cross-jurisdiction comparisons for opinions; cannot risk hallucinated authority or inconsistent sourcing.
• In-house tax or finance team at a company: Small teams juggle multi-state filings, treaties, and rule changes; they need repeatable workflows and alerts rather than ad‑hoc research.
• Solo practitioner / enrolled agent: Receives many client documents and notices in varied formats; spends billable time on extraction and drafting instead of higher-value advisory work.

How would they acquire their first 10, 50, and 100 customers

• First 10: Run founder-led, paid pilots with small firms and solo practitioners using the required live demo as the conversion point; provide hands-on onboarding, gather feedback, and secure testimonials to refine workflows and error handling demo/FAQ support.
• First 50: Host weekly webinars around common pain points (research, document analysis, return review), amplify LinkedIn and guide content to drive trials, and offer referral incentives to early pilot firms Tax Research guide homepage.
• First 100: Standardize a one‑hour demo-to-trial playbook, hire an SDR/implementation lead to scale onboarding, and pursue integrations/embeds with tax prep and practice‑management vendors for channel reach embed/lead gen seed blog.

What is the rough total addressable market

Top-down context:

In the U.S., there are about 653,000 actively licensed CPAs and roughly 850,000 individuals with current PTINs who prepare returns, indicating a large base of professionals who purchase tax research and workflow tools NASBA IRS PTIN stats. The broader accounting software market is a multi‑billion‑dollar category globally, suggesting ample spend available for specialized AI assistants in tax workflows market context.

Bottom-up calculation:

If TaxGPT targets 300,000 professional seats across firms and in‑house teams in the U.S./Canada at an average of $100/month per seat, annual TAM would be about $360M. Adding an enterprise tier (e.g., 25,000 seats at $200/month) would add roughly $60M, for a combined ~$420M in annual software spend in the core markets.

Assumptions:

• Targetable buyers include a subset of PTIN holders, CPAs, EAs, tax attorneys, and in‑house professionals; we model 300k seats as an attainable serviceable pool.
• Average pricing approximated at $100/month/seat for standard and $200/month/seat for enterprise, reflecting research+automation value.
• Scope focused on U.S./Canada buyer base aligned with current coverage; does not include broader international expansion.

Who are some of their notable competitors

• Thomson Reuters (CoCounsel Tax / Checkpoint): Incumbent research vendor with AI features tied to Checkpoint/CoCounsel; strong curated sources and enterprise controls, with higher cost and a research‑first orientation for smaller firms.
• Blue J: AI tax research focused on defensible, citation-backed analysis and multi‑jurisdiction interpretation; competes on research quality more than on return‑prep automation workflows.
• Bloomberg Tax: Legacy tax research platform adding AI on top of a large database of codes, rulings, and analysis; valued for source authority and monitoring rather than embedded firm automation.
• Filed: Automates preparation and review workflows (document ingestion and “review‑ready” returns); overlaps with TaxGPT’s document analysis and review use cases.
• CPA Pilot: Startup AI copilot for small/mid‑market CPA firms offering citation‑backed research, drafting, and workflow automation; a direct functional competitor to TaxGPT for firms wanting an all‑in‑one assistant.