Forrester Just Told Enterprises Delaying an AI Platform Decision Is 'An Existential Risk' — a16z's Numbers Say Why
Forrester’s Q3 2026 Wave on AI platforms, covering fifteen vendors from AWS to UiPath, opens with a claim built to get a VP’s attention: delaying an AI platform decision is now “an existential risk” for enterprises that don’t convert AI into completed work across core processes. Analysts Mike Gualtieri and Rowan Curran argue agentic AI has redefined what the category even means — the competitive bar has moved from data-science workbenches to platforms that “understand context, navigate workflows, and complete work.” Their recommendation isn’t to pick a winner: it’s a portfolio strategy, matching different vendors’ strengths (workflow automation, RPA, enterprise SaaS, cloud infrastructure) to different use cases rather than betting the org on one monolithic platform. Frontier model labs are deliberately excluded — Forrester is treating them as a separate market with its own forthcoming evaluation.
The “agentic AI redefined the category” framing isn’t just analyst language — there’s fresh data behind it. Andreessen Horowitz published benchmark numbers the same week showing computer-use agents have crossed from demo to production viability for narrow, high-volume work: on the OSWorld-Verified benchmark, the leading model (Claude Fable 5) now completes 85% of tasks, above the roughly 72% human baseline and up sharply from about 42% a year ago. On cost, a16z put fully-loaded computer-use agents at $6–8/hour against $30–45/hour for US back-office staff, with the strongest fit in protocol-following work — CRM updates, IT tickets, portal interactions — not open-ended judgment calls.
For a revenue org still treating “AI platform” as a single line item on next year’s budget, the two pieces together are the actual argument: Forrester’s Wave says the category has shifted to agentic, and a16z’s numbers show that shift is already cost-competitive with a back-office headcount line. The vendor selection question Forrester is pushing enterprises toward isn’t “which platform” — it’s “which workflows are ready to hand off, and to which combination of platforms,” a portfolio call that has to be made now, not once the market settles.