How CFOs are navigating growth, pricing, and forecasting in an AI world (AI column)

I read an interesting newsletter from a16z Foundation on fintechIt discussed growth, pricing and prediction in our new AI world. Respected gentlemen from respected companies shared their views (especially Databricks) Dave Conte, CFO at Databricks – a platform for enterprise-grade analytics and AI solutions Maciej Mylik, Finance at ElevenLabs – research and deployment company for AI audio technologies Hanson Hermsmeier, VP Corporate Finance in Together AI – a cloud platform for AI developers and researchers Matthieu Hafemeister, co-founder of Concourse, is a company that creates AI agents for corporate finance teams Noah Barr, CFO in Ambient.ai – computer vision company for automated security

These gentlemen shared their ideas on topics 1. Rethinking Pricing: Moving from Subscription to Consumption and Outcome n AI is driving a shift towards results-based pricing. Databricks uses a model where "pricing and revenue recognition are entirely based on output, as opposed to input-based consumption patterns." ElevenLabs applies dynamic pricing – reducing unit prices while increasing customer obligations to stimulate large contracts. In general, it looks like models with revshare or discount model for wholesale consumers. 2. ARR (Annual Recurring Revenue) requires rethinking Traditional ARR metrics do not reflect the reality of use-based pricing models. Companies are implementing hybrid metrics: ElevenLabs tracks “ARR plus annualized use” to correctly account for corporate customer quota exceedances (conditionally, consumers exceed their quotas in subscriptions, which means it is better to take into account not subscription for the year, but consumption taking into account excesses.) Databricks uses its own data platform to understand and predict true ARR based on consumption. 3. Pressure on gross margin and cost management Building on foundation models introduces significant variable costs scalable using AI. The marginal cost of an additional user is no longer zero and varies by user. The key challenges are Infrastructure spending monitoring - ElevenLabs tracks cost growth relative to usage Managing fixed GPU costs – Together AI tracks GPU downtime as efficiency losses New Cost Types – Ambient.ai includes the “human in the loop” command 4. Assessing ROI in the AI world As some AI functions commoditize, investments in the future become critical. Companies are focusing on Long-term differentiation through research projects Creating complex product layers to increase switching costs (switch cost) Predictive analytics to measure the impact of opportunities on customer growth 5. Using AI for Advanced Financial Forecasting Planning for 12 For months, AI remains a challenge due to constant innovation. Decisions include Databricks uses its own platform to predict consumer and product consumption patterns Apply advanced analytics and machine learning instead of traditional tools like Excel Recognition that reliable revenue forecasting for AI is not yet fully resolved

In general, the creators of products with AI for large users have to solve issues of growth, pricing and prediction in a changed AI world.

#AI #Management #ML #Software #Engineering