AI Summary
→ WHAT IT COVERS Keith Paris, CEO of Lightfield (a16z's $47M Series A investment), explains how he pivoted from Tome's 25 million AI presentation users to building an AI-native CRM. Lightfield constructs a "business world model" from emails, calls, and meetings, replacing rigid database schemas with intelligence-driven customer relationship records. → KEY INSIGHTS - **Pivot trigger — product conviction over metrics:** Keith abandoned Tome despite 2 million monthly users because the founding team could not envision high-stakes professionals, such as investment bankers or consultants, using it indispensably. Founders should treat personal conviction in the product as a leading indicator, not a lagging one, before growth metrics force the decision. - **Intelligence over schema architecture:** Traditional CRMs fail because rigid relational schemas lock in bad data models from day one — wrong stages and fields cannot be retroactively filled. Lightfield stores raw activity logs chronologically, then infers structured fields on demand, meaning schema changes require no historical re-entry and onboarding compresses to a five-minute email sync. - **Negative pricing as early validation:** Lightfield offered free office space to 10 startups in exchange for using a four-month-old CRM. Hourly Slack feedback from those users — despite a slow, incomplete product — signaled genuine product-market fit. Founders should measure engagement intensity, not satisfaction scores, when validating early-stage systems of record. - **Hybrid pricing model for AI CRMs:** Pure seat pricing caused 10x consumption imbalance across users; pure credit pricing froze engagement entirely. Lightfield resolved this with a fixed platform-plus-seat fee covering core CRM capture, then consumption pricing for pipeline generation, workflow automations, and intelligence forecasting — separating predictable operational costs from variable ROI-generating activities. - **Expansion-first prioritization in competitive markets:** Rather than optimizing for initial land, Lightfield evaluates each account's three-year expansion potential and prioritizes engineering resources toward fastest-growing customers. Early-stage Silicon Valley logos function as marketing assets, not revenue targets, enabling entry into sectors like healthcare and manufacturing through referenceability rather than direct sales. → NOTABLE MOMENT When Lightfield connected to Power's clinical trial marketplace — modeling both pharmaceutical companies and patients with complex illnesses — automated matching against FDA and clinicaltrials.gov data helped an Alzheimer's patient locate frontier treatment options within days, a workflow impossible inside conventional CRM schema constraints. 💼 SPONSORS None detected 🏷️ AI-Native CRM, Startup Pivots, Business World Model, SaaS Pricing Strategy, Systems of Record