Crunched

What do they actually do

Crunched is an AI assistant that works inside Excel and Google Sheets. Users can ask it in plain English to build models (e.g., DCFs, market models, LBOs), debug existing workbooks, and extract numbers from PDFs or the web. It previews planned changes in‑sheet for review, then writes formulas, links, formatting and charts directly into the workbook; edits are reversible via an undo/restore flow (site, terms, YC, demo post).

Early adoption is among Excel power users in consulting, investment banking and private equity, with enterprise features for secure deployment and admin controls. The company states customer data is not used to train models, uses AES‑256 encryption in transit and at rest, supports EU data residency, lists ISO 27001 and GDPR compliance, and notes SOC 2 Type 2 is under monitoring. Its terms require customers to validate outputs before relying on them (enterprise, security, terms, YC).

Who are their target customer(s)

• Management consultants (senior consultants/managers): They build and iterate multi‑sheet client models and must catch logic or formula errors before delivery. They need faster model building, formatting, and end‑to‑end error checks inside the workbook (site, YC).
• Investment banking analysts/associates: They produce DCFs, LBOs and valuations under tight deadlines and pull figures from IMs/annual reports into linked templates; manual data entry and hidden errors derail deal timelines (site, YC).
• Private equity analysts/associates: During diligence they need auditable, consistent models that survive many edits and reconcile to sources, and worry about hidden formula bugs across large workbooks (YC).
• Corporate FP&A and finance managers: They spend recurring cycles compiling reports, reconciling inputs and hunting anomalies, and want to automate repetitive reporting without introducing errors (site).
• Accounting teams / controllers: They need correct consolidations, formulas and audit trails and are wary of tools that change live workbooks without easy recovery; preview/approve and undo/restore flows reduce that risk (terms, site).

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

• First 10: Founder‑led, high‑touch pilots with consulting, IB and PE teams (initially in Europe), installing into real workbooks and running live model builds/debugging to show time‑savings and caught errors, using in‑sheet preview/approve and undo/restore to build trust (site, YC).
• First 50: Turn pilots into a repeatable play: land adjacent teams via champion referrals and targeted outbound, offer short paid pilots with tailored templates (DCF/LBO/FP&A) and a clear security package (SSO, EU residency) to ease procurement; publish case studies to shorten sales cycles (enterprise, security).
• First 100: Scale into the U.S. and larger accounts with a small enterprise sales team, standardized onboarding, and completed compliance (e.g., SOC 2 Type 2 and contractual terms). Add a self‑serve trial and template library for mid‑market FP&A/accounting while routing larger deals through sales and security review (YC, security).

What is the rough total addressable market

Top-down context:

The broader market is teams that build and audit analyses in Excel and Google Sheets. Crunched focuses first on finance/consulting power users who need reliable in‑sheet modeling, debugging, and data extraction.

Bottom-up calculation:

If Crunched serves roughly 120,000 power users across consulting, investment banking, private equity, FP&A and accounting at $1,200 per user per year, the initial beachhead TAM is about $144M annually.

Assumptions:

• Target geographies are the US and Europe, with an estimated 600,000 professionals in consulting/IB/PE/FP&A/accounting roles and ~20% being power users who would adopt.
• Average contract value approximated as $100 per user per month ($1,200 per year) for enterprise functionality.
• TAM reflects the initial wedge; expansion to broader finance/ops roles or deeper automation would increase the ceiling.

Who are some of their notable competitors

• Microsoft Copilot (Excel): Microsoft’s native AI in Excel can generate formulas/charts, answer natural‑language prompts, and pull data, bundled with Microsoft 365 for enterprise customers (Copilot in Excel).
• Google Duet AI (Sheets): Duet AI features for Workspace let users ask Sheets to analyze data, build charts, and source from Drive/Gmail, competing for teams standardized on Google Sheets (Duet AI).
• Coefficient: Add‑in for Sheets and Excel focused on live data connectors, scheduled refresh, and an AI copilot for formulas/pivots/charts; popular with FP&A and RevOps (product, AI copilot).
• Endex: Excel‑native AI agent for financial modeling with an emphasis on auditability and enterprise deployments—directly overlapping finance/consulting use cases (Endex).
• Formula Bot / spreadsheet GPT add‑ins: Lightweight tools that convert plain English to formulas, explain formulas, and generate VBA/Apps Script; useful for quick tasks but typically lighter on large‑model validation (FormulaBot, GPTExcel, GPT for Excel).