Decisional AI

What do they actually do

Decisional turns existing spreadsheets into autonomous agents that complete recurring, spreadsheet-based work without code. Users describe the job in plain English and connect the agent to their spreadsheet, documents, and apps so it can run tasks end-to-end in the tools they already use (homepage, examples).

In practice, a user loads a spreadsheet, writes a job description, and links data sources (Decisional advertises 200+ connectors). The agent ingests documents (PDFs, Excel, Word, OCR), applies the sheet’s logic with runtime reasoning, updates rows, produces outputs (e.g., reconciliations, memos), and leaves citations/audit logs for review (examples, MarkTechPost, Homebase summary). The company also signals enterprise readiness with features like a usage dashboard, implying active pilots or customers (Decisional LinkedIn).

Who are their target customer(s)

• FP&A and financial analysts at scaling startups: They spend hours each period copying, reconciling, and sanity‑checking spreadsheets and need a reproducible, auditable way to finish those tasks without building custom code (examples, homepage).
• Revenue operations and billing teams: They stitch together data from spreadsheets, billing tools, and documents to generate quotes/invoices and track payments, leading to missed invoices and slow collections (examples, homepage).
• Accounting / AR teams: They lose time matching bank/ledger rows, chasing missing receipts, and producing audit trails for approvals and compliance (examples, Homebase).
• Investment analysts, VC/PE associates, finance-focused ops: They compile due diligence and investment memos from many documents, manually extracting facts and struggling to cite sources or reproduce conclusions (examples, MarkTechPost).
• Small ops/people teams and customer-success ops at growth-stage companies: Their bespoke spreadsheet automations (macros, scripts, manual steps) are brittle and break with scale or team changes, requiring engineering to fix (homepage, blog).

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

• First 10: Run hands-on pilots via YC/investor intros and outbound to FP&A, billing, accounting, and investment teams; wire agents into real spreadsheets, deliver a working run and audit report, and use a short pilot contract with clear success metrics to convert.
• First 50: Productize early wins into 3–5 ready templates (e.g., monthly close, AR follow-up, invoice generation, investment memo) and drive targeted outbound, niche finance newsletters, and short webinars with time‑limited trials. Publish tight case studies and offer a standard one‑week setup + 30‑day pilot package.
• First 100: Launch a self-serve template library with simple connectors and stand up a partner/reseller channel (bookkeepers, fractional CFOs, integrators). Invest in onboarding automation and referral incentives to convert trials and scale paid rollouts.

What is the rough total addressable market

Top-down context:

Decisional sells into the broader automation/hyperautomation space, which is already in the low tens of billions of dollars: RPA was about $18.18B in 2024, while broader hyperautomation estimates for 2024 range roughly from ~$13B to ~$46B depending on scope (Fortune Business Insights, Mordor, GMInsights).

Bottom-up calculation:

Using the $18.18B RPA estimate and an indicative ~34% finance/accounting share implies a ~$6.2B finance/ops automation SAM in 2024 (18.18B × 0.34 ≈ 6.18B) (Fortune Business Insights, SNS Insider segmentation example). Applying the same share to a ~$46.4B hyperautomation estimate implies ~=$15.8B (GMInsights).

Assumptions:

• Finance/accounting capture ~34% of function-level automation spend; actual share varies by source and definition.
• Spreadsheet- and document-heavy workflows are a subset of finance/ops automation and materially smaller than the whole SAM but high-value.
• Estimates use 2024 figures and high-level market definitions; numbers are directional, not precise.

Who are some of their notable competitors

• Microsoft Excel (Copilot / AI in Excel): Built-in AI for formula generation, analysis, and workflow helpers inside Excel reduces the need for separate tools for teams already in Excel (Microsoft).
• Google Sheets (Gemini / AI in Sheets): Gemini adds smart fill, AI functions, and natural-language analysis directly in Sheets, enabling prompt-driven work without separate agent platforms (Google).
• Zapier (and similar no-/low-code platforms): Rule/trigger-based automations connect Sheets/Excel to apps and handle many repetitive steps, overlapping with integration use cases but not agentic reasoning (Zapier).
• fileAI (file.ai): Focuses on OCR, document ingestion, and structured extraction from PDFs/spreadsheets to feed downstream automations—competing on the data-prep/document-intelligence layer (fileAI).
• SheetAI (and similar add-ons like Numerous.ai): Add-ons that put LLM functions into Sheets for formula generation and bulk cell transforms; closer to AI-in-cells than full multi-app, auditable agents (SheetAI, Numerous.ai).