# Forethought (Series B) Financial Model AI-powered customer support assistant ("Agatha") that helps support agents solve, triage, assist, and discover answers faster. - Canonical: https://finamodel.com/startups/forethought-series-b - Excel download: https://finamodel.com/startup-models/forethought-series-b.xlsx - Category: AI/ML - Model type: SaaS ARR / Valuation - Funding round: Series B - Funding: $17M - Founded: 2017 - Geography: US (San Francisco HQ implied by TC Disrupt SF win; target customers are US tech companies). - Customer: B2B ## About the company Forethought's Series B case centers on its enterprise customer-support automation platform: Solve, Triage, Assist, and Discover. The product learns from historical conversations to deflect requests, route tickets, help agents with knowledge, and identify gaps in the support operation. The land-and-expand motion begins with one support workflow and adds modules, agents, and channels. Its customer set includes Instacart, Marriott, Gusto, Carta, Asana, and Upwork, while its platform carries ISO 27001 and SOC 2 certifications for enterprise deployment. The model uses an ARR bridge that separates new contracts, module expansion, and churn. It forecasts support-team adoption, AI inference and integration cost, gross margin, sales productivity, customer-success and R&D headcount, cash burn, and runway, with resolution and deflection rates as operating drivers. ## What's included - 5-year monthly revenue build with stage-appropriate growth assumptions - Full P&L, headcount plan, and operating-expense schedule - Cash-flow statement, runway, and burn-rate tracking - Valuation via exit multiple with a DCF cross-check - Returns analysis with MOIC and IRR - Unit economics including CAC, LTV, payback, and cohort retention ## Product & value proposition - Product: "Agatha" - an AI assistant embedded in customer support workflows. Four stated capabilities: - **Solve** - auto-resolves incoming tickets. - **Triage** - classifies and routes tickets. - **Assist** - surfaces relevant knowledge-base articles to live agents. - **Discover** - surfaces knowledge gaps / patterns from ticket data. - Core positioning: Forethought vs. Google Search (slide 8) - proprietary enterprise knowledge search that returns contextually ranked internal articles (HelpCenter, Support Articles, Internal KB) rather than generic web results. Search results include recency stamps (e.g., "2 weeks", "2 months", "3 months"). - Mission: "Make everyone a genius at their job." ## Market - TAM stated as **$100B Opportunity**. No methodology, SAM, or SOM breakdown provided. - Narrative framing: AI as the "4th Industrial Revolution" (following steam, electricity, computing). Used as market-size justification, not a quantified addressable market. - No CAGR, analyst source, or segmentation in deck. ## Revenue model - Not explicitly stated in deck. Inferred from B2B SaaS context and enterprise customer list: - Per-seat or platform subscription sold to customer support teams; annual contract value likely scales with number of support agents or ticket volume. Rationale: standard for AI support tooling (Zendesk, Intercom comps). - Direct sales motion targeting COO / VP of Customer Support at mid-to-large companies. - No pricing tiers, ACV, or contract structure disclosed. ## Traction & metrics - **TC Disrupt SF Battlefield winner** - prize check of **$100,000**. Confirms early-stage. - **Named customers (6 logos)**: Carta, Instacart, MasterClass, Thumbtack, Gusto, Typeform. - **Target prospects named**: ESPN, TicketMaster. - No ARR, MRR, revenue growth rate, ticket deflection rate, agent CSAT lift, or customer count disclosed. - No retention, NRR, or churn figures in deck. ## Competition / moat - Slide 8 positions Forethought's search directly against Google Search - implying the moat is enterprise-specific, context-aware NLP that understands internal knowledge bases vs. public web search. - No explicit competitive landscape slide or named competitors (Zendesk AI, Intercom Fin, Salesforce Einstein, etc.). - Moat claims implicit: ML publications and infrastructure patents held by CEO Deon Nicholas; team pedigree from LinkedIn, Google, Facebook, Dropbox, Qualtrics, MIT, Harvard, Berkeley, Waterloo. ## Team & funding ask / use of funds - **CEO & Co-Founder: Deon Nicholas** - Forbes 30 Under 30; ML publications and infrastructure patents; 2x ACM ICPC World Finalist; alumni of Facebook, Dropbox, Palantir, Pure Storage; University of Waterloo. - Broader team: LinkedIn, Google, Facebook, Dropbox, Qualtrics; academic pedigree MIT, Harvard, Berkeley, Waterloo. - Use of funds implied by "What's Next": build brand, execute GTM, hire commercial relationships (COO / VP Customer Support buyers), target ESPN and TicketMaster as enterprise anchors. --- ## Recommended financial model - **Archetype + why**: B2B SaaS ARR model. Revenue is subscription-based, sold to enterprise customer support teams. Seats or platform fee scales with customer size. Standard SaaS ARR waterfall (new ARR, expansion ARR, churn, net ARR) is the right frame. - **Forecast horizon & granularity**: 5 years (Year 1–5); monthly for Year 1–2, annual for Years 3–5. Monthly needed to model early customer ramp and cash burn. - **Key drivers & assumptions**: - New logos per month: 1–2/month in Y1 ramping to 5–8/month by Y3. Rationale: 6 named customers at deck time; GTM being built. - Average ACV: $30K–$80K/year depending on company size. Rationale: mid-market SaaS AI tool; below Salesforce enterprise, above SMB point tools. - Net Revenue Retention (NRR): 110%–120%. Rationale: seat expansion as support teams grow; AI tools with proven deflection tend to expand. - Gross margin: 70%–75%. Rationale: SaaS with AI inference costs; below pure software but above 60%. - Sales cycle: 60–90 days for mid-market, 120–180 days for enterprise (ESPN/TicketMaster-type). - CAC (blended): $15K–$40K per customer. Rationale: direct sales to VP/COO; assume 1 AE closes 10–15 logos/year at $120K–$150K OTE. - Payback period: 18–24 months. Rationale: typical for mid-market SaaS. - Headcount: Sales + CS hiring paced to logo growth; Engineering flat in Y1 then +2–3 engineers/year. - Churn: 8%–12% gross annual logo churn. Rationale: early-stage with limited brand; improves as product matures. - **Scenarios (Base / Bull / Bear)**: - **Base**: GTM execution on track; 40–50 logos by end Y2; ACV $45K; NRR 115%. - **Bull**: ESPN/TicketMaster-type enterprise anchors land in Y1; ACV lifts to $80K+; NRR 125%; accelerated word-of-mouth in tech customer support segment. - **Bear**: Long sales cycles and brand-building delays; logo count misses by 30%; churn elevated at 15% as product-market fit narrows. - Flex variables: ACV, new logos/month, NRR, gross margin (if inference costs rise), headcount timing. - **Required sheets / outputs**: 1. Assumptions dashboard (all drivers, toggle Base/Bull/Bear) 2. ARR waterfall (new, expansion, churn, net new ARR by month) 3. P&L (Revenue, COGS, Gross Profit, S&M, R&D, G&A, EBITDA) 4. Headcount plan (by function, linked to opex) 5. Cash flow & runway (burn rate, months of runway) 6. Unit economics summary (CAC, LTV, LTV/CAC, payback) 7. Customer cohort table (logo count by cohort, expansion/churn by year) 8. KPI dashboard (ARR, MRR, logo count, NRR, gross margin, burn) ## Frequently asked questions ### Is the Forethought (Series B) financial model free? Yes. The Forethought (Series B) model is a free Excel download with live formulas. ### Can I change the assumptions? Yes. The workbook is editable and its live formulas recalculate when assumptions change.