Wedge

The beachhead targets complex individual returns featuring multiple K-1s and Schedule Cs for regional CPA firms. This niche presents the highest volume-to-complexity ratio and causes severe seasonal bottlenecks. After dominating high-complexity 1040s, the product expands horizontally into corporate returns and partnership filings to capture the entire preparation workflow.

Timing

Context-window expansions in large language models now allow ingestion of entire tax document packages simultaneously. Structured extraction via vision-language models accurately pulls numbers from non-standardized forms like complex K-1s or brokerage statements, replacing manual transcription.

Why This ICP

Independent and regional CPA firms feel the talent shortage most acutely because they cannot compete with Big Four salaries. They have high seasonal volume and immediate willingness to pay for tools that replace unavailable headcount.

Size Of Prize

There are approximately 90,000 small-to-midsize accounting firms in the US spending an average of $30,000 annually on seasonal contractor labor for tax data entry, creating a total addressable prize of roughly $2.7B.

Gap Narrative

CPA firms face a severe shortage of qualified tax preparers while dealing with increasing tax code complexity. They need a system that ingests messy client documents, maps them directly to tax schedules, and drafts the return for senior CPA review without requiring a junior accountant to manually key in the data.

Defensibility

Defensibility compounds through workflow lock-in and localized tax-mapping intelligence. The system learns a specific firm's categorization preferences for recurring clients, raising switching costs because a new tool requires re-teaching those nuances. The core extraction technology is a commodity, meaning the actual moat relies on deep integrations into incumbent tax software systems like CCH or Thomson Reuters.

Why This Thesis

Service-as-Software fits perfectly because tax preparation requires a highly standardized output derived from semi-structured inputs. Firms want the actual labor of document classification and schedule calculation completed so senior partners simply review and sign.

Hardest Part

Mapping unstructured user uploads like blurry receipts and non-standard tax forms to exact IRS tax code logic without hallucinating deductions or triggering compliance penalties.

Min Viable Scope

Build exclusively for single-state W-2 employees with basic 1099 freelance income. Completely exclude corporate returns, complex real estate depreciation, K-1 partnership parsing, and multi-state compliance.

Cold Start Problem

Models lack reasoning capabilities over niche tax edge-cases without large volumes of historically verified returns and their source documents. Break this by partnering with a regional CPA firm to run parallel shadow processing on their current-year client filings.

Time To First Value

10 minutes to process uploaded documents and generate a draft federal return, gated by the user gathering and uploading their complete document set.

Data Moat Available

true

Technical Difficulty

High

SAM

~$1B-2B US mid-market accounting firms

SOM

~$20M-50M

TAM

~100k US accounting firms × ~$50k/yr ≈ ~$5B

Growth Rate

~12-18%/yr, driven by worsening CPA pipeline shortages and rising seasonal contract rates

Paid Comparable Spend

~$40k-100k/yr per firm on outsourced seasonal labor, offshore prep teams, and legacy tax software licenses

Kill Thresholds

• Senior CPA review time exceeds 45 minutes per mid-complexity return
• Data extraction error rate exceeds 5 percent on standard W-2 and 1099 forms
• Zero paid pilot conversions at the $10,000 tier after 30 days of active usage
• Security or compliance objection rate exceeds 20 percent of sales pipeline

Leading Metrics

• Time from raw document upload to CPA-ready draft
• Percentage of AI-mapped tax fields accepted without human override
• Number of returns processed per active user per day
• Exception rate requiring fallback to manual document entry

What Proves Right

Mid-market accounting firms replace at least one seasonal offshore contractor with the software during tax season, paying upwards of $20,000 annually. Users auto-generate completed tax forms from raw client document dumps with less than 15 percent of fields requiring manual correction by a senior CPA. First-season cohorts commit to next year's prep with upfront annual contracts within 90 days of tax day.

What Proves Wrong

CPAs spend as much time reviewing and correcting the AI outputs as they spend doing the preparation manually from scratch. The system fails to correctly parse unstandardized client documents, forcing admins to manually enter data before processing begins. Firms refuse to upload actual client PII due to compliance fears, trapping the platform in pilot purgatory.