Opportunities
Audit Graph
Connected through 7 “incumbent in” links and 1 “applies thesis” link.
Structure
Demand side
Opportunities
Opportunities
Connected through 7 “incumbent in” links and 1 “applies thesis” link.
Structure
Demand side
The gap
Wedge
The initial beachhead is revenue transaction vouching for software and manufacturing client audits. This niche features highly standardized document triangles, such as invoices, proofs of delivery, and bank receipts, and represents the largest block of manual testing hours. After dominating revenue testing, the system expands horizontally into expense tracing, payroll verification, and ultimately full trial balance tie-outs.
Timing
Context windows of over one million tokens and advanced multimodal extraction capabilities now allow models to ingest entire unstructured evidence directories and accurately match them to massive ERP ledger exports. Two years ago, attempting to cross-reference multi-page PDFs against CSV ledgers at this scale resulted in unacceptably high hallucination rates.
Why This ICP
Regional CPA firms face an acute, well-documented shortage of junior accountants and shrinking margins from offshoring costs. They act quickly to adopt solutions that execute substantive testing because it immediately relieves critical capacity bottlenecks during peak busy seasons.
Size Of Prize
There are approximately 46,000 public accounting firms and 15,000 mid-market internal audit departments in the US. Capturing 20,000 of these target entities at an annual software-as-labor spend of $30,000 per year yields a $600M addressable prize for automated substantive testing.
Gap Narrative
Auditors spend hundreds of hours manually tracing ledger transactions to unstructured source documents like invoices and bank statements. Current audit software acts as a static repository for workpapers, failing to actively map or verify the relationships between financial assertions and underlying evidence. This leaves a gap for a system that autonomously links line items to verified extracted data across disparate client files.
Defensibility
Defensibility stems from deep workflow lock-in and a compounding client-specific data moat. As the product maps the idiosyncrasies of specific client ERPs and document formats year over year, the accounting firm experiences near-zero setup time for recurring annual audits, making switching to a competitor prohibitively expensive in lost efficiency.
Why This Thesis
The Service-as-Software model maps directly to the public accounting business model. Firms require completed testing matrices and tied-out workpapers rather than another SaaS dashboard to manage, meaning the product directly replaces the specific unit of manual labor they lack.
Overview
Build difficulty
Hardest Part
Reliably parsing unstructured financial evidence like layered PDFs and complex contracts to deterministically map them to specific ledger line items with zero hallucination. If the linkage confidence is flawed the entire graph loses audit-grade trust.
Min Viable Scope
The v1 maps Accounts Payable invoices directly to bank feed transactions for single-entity SaaS companies. Deliberately leave out Accounts Receivable, multi-currency consolidation, payroll reconciliation, and real-time API ingestion in favor of batch flat-file uploads.
Cold Start Problem
The matching models require thousands of verified evidence-to-ledger pairs to achieve baseline accuracy across diverse vendor formats. The first move is acting as a tech-enabled service for a single regional audit firm to manually verify mappings and build the initial training corpus.
Time To First Value
1 to 2 weeks of historical data ingestion and initial graph generation
Data Moat Available
true
Technical Difficulty
High
Build profile
Sized prize
IllustrativeIllustrative targets and order-of-magnitude estimates — not an achieved track record. This Thing is concept-stage; real figures come from live data once operating.
SAM
~4k-6k US mid-market and regional audit firms x ~$40k-60k/yr ≈ ~$160M-360M
SOM
~$10M-25M
TAM
~40k-50k global audit firms x ~$40k-60k/yr per firm ≈ ~$1.6B-3B
Growth Rate
~12-15%/yr, driven by acute CPA pipeline shortages and expanded regulatory mandates for deep-tier fraud detection
Paid Comparable Spend
~$150k-300k/yr per firm spent on junior auditor billable hours for manual substantive testing and legacy data extraction tools
Market sizing
How you know
Kill Thresholds
Leading Metrics
What Proves Right
Mid-market audit firms connect client ERP and banking systems to Audit Graph within three days of onboarding. Engagement teams eliminate 60 percent of manual substantive testing hours per audit and rely entirely on the graph for deep-tier fraud detection. Pilot programs convert to 40,000 dollar annual contracts within 30 days.
What Proves Wrong
Partners refuse to sign off on automated graph mappings and revert to manual Excel sampling to satisfy regulatory peer reviews. Client IT teams block financial data ingestion over security liabilities, stretching deployments beyond 90 days. Graph compute and traversal costs exceed 30 percent of the contract value, destroying gross margins.
Win conditions