Anthropic, Atlassian, and Scale Landed on the Same AI Playbook Independently

Three companies at SaaStr AI 2026 — Anthropic, Atlassian, and Scale Venture Partners — independently landed on the same answer to the “replace or augment” question, and it wasn’t the one AI-native pitches usually make. Anthropic’s Eleanor Dorfman described threading Claude through the sales tools reps already use — Salesforce, Slack, Gong, Clay, Ironclad — rather than replacing them, a move she credits with pushing self-serve enterprise adoption to 54% of new logos in 2026 and returning 70% of AE time from admin work back to selling, via five specific “daily skills”: morning brief, call prep, follow-up, competitive intel, asset creation. Atlassian’s Sharif Mansour described a similar path from the product side: ship chat-based AI first to discover which workflows actually deserve a dedicated UI, under the rule “anything a human can do, an agent can do.” Scale’s Rory O’Driscoll added the capital-markets argument for why this converges: foundation-model companies are running $688B in annual capex against $110B in revenue, implying five-plus more years of investment-mode spending — so defensibility has to come from proprietary data and workflow integration, not the model itself, which he called a “new LAMP stack.” Recapped by SaaStr’s Jason Lemkin.

The pattern matches what’s showing up in adoption data elsewhere. A separate roundtable of revenue leaders from Renaissance Learning, ServiceNow, and Databricks reached the same structural conclusion — one panelist’s line, “process and people before technology,” and the finding that the biggest AI gains came from previously neglected segments (long-tail accounts, churned customers) rather than core pipeline, is the adoption-side mirror of Dorfman’s stack-integration approach. And 11x’s own data on GTM teams plateauing after initial AI-agent deployment — 87% of sales orgs already use AI for prospecting, yet most stop scaling after the pilot — is what happens when the “daily skills” layer Dorfman describes never gets built past the first workflow.

For sellers, the takeaway isn’t which vendor wins the platform argument. It’s that “we use AI” is now table stakes, and the actual differentiator — for a seller’s own tools and for what to ask a prospect’s stack — is whether AI is threaded into five or six specific, named workflows, or bolted onto one and left there.