Glean

OpenAI previews Private Safety Processing for ZDR

OpenAI reaffirmed Zero Data Retention (ZDR) for eligible API customers and previewed Private Safety Processing, a privacy-first safety capability that detects patterns across related interactions without exposing underlying prompts or model outputs to OpenAI staff. Under ZDR, customer prompts and model outputs are not retained and enterprise data won’t be used to train models unless customers explicitly opt in. Private Safety Processing runs on customer-controlled infrastructure or on OpenAI-hosted storage encrypted with customer-controlled keys (unavailable to OpenAI personnel), emitting narrow automated safety signals for enforcement rather than returning customer content. The feature is in trials with early customers; OpenAI plans a broader rollout and a technical white paper in September. The announcement is framed as competitive differentiation amid tensions with Anthropic’s new 30‑day session retention policy.

Frontier Model Costs Drive Demand for Model Routing

Rising costs of frontier LLMs and the growing power of open-weight models have made model routing a critical part of enterprise AI deployments. Glean, an enterprise AI company led by ex-Google engineer Arvind Jain, uses a three-tier routing approach (explicit selection, admin controls, automatic routing) to reduce costs and reserve advanced models for tasks that need them. Glean reports rapid commercial traction — a $7.2B valuation after a $150M Series F and $300M ARR — and says its Waldo agentic search model reduces latency by 50% and token usage by 25%. Enterprises are increasingly considering open-weight models and multi-provider strategies to control AI spend, while Glean uses real-world feedback and internal evals to continuously improve routing decisions.

Enterprise AI Go-to-Market Playbook: Lighthouse vs Landgrab

This Substack piece presents a go-to-market playbook for enterprise AI startups centered on a two-question diagnostic—how exposed is the deal signer and whether social proof travels in that market—to decide between two sales motions called “Lighthouse” and “Landgrab.” The author cites an essay by a16z partner Joe Schmidt and Julian Marx naming the split, argues that many founders default to the wrong motion, and links the framework to companion guides about market ownership and distribution. The playbook analyzes why most enterprise GenAI pilots fail to reach production, shows how startups (Sierra, Glean, Abridge, Rogo, Legora, Clay, Cursor) chose their GTM lanes, outlines common traps, provides a 7-point checklist, and ships a downloadable diagnostic (lighthouse-landgrab) as a Claude skill.