30-Second Take Week 33 produced two hard findings that look unrelated. Inside the enterprise: 95% of organisations stalled or cancelled an AI project in the past year. The named blocker isn’t model quality. It’s that nobody can state, in machine-readable form, what is actually true about the business. The organisations that have governed their context…
THE 30-SECOND TAKE The AI answer layer became a marketplace this week. ChatGPT started serving multi-product carousel ads built straight from retailer feeds, Google Maps began taking food orders and hotel bookings inside the answer, and Cloudflare gave AI agents identities, spending limits and payment rails. At the same time, three independent studies made machine…
As of March 2026, most organisations that invested in AI over the last two years have nothing measurable to show for it. Not underperformance. Zero impact on profit and loss. MIT’s NANDA initiative studied 300 public AI deployments and interviewed 150 executives. Ninety-five percent of enterprise AI pilots delivered zero measurable P&L impact. The same…
As of February 15, 2026, three major AI developments are reshaping the industry: (1) DeepSeek’s R1 open-source reasoning model released January 2026 proved Chinese firms can match frontier capabilities, (2) OpenAI’s operator agents now autonomously browse and interact with websites, and (3) context engineering has emerged as the critical skill for AI workflow optimization. These…
by Pål Erik Waagbø | Week 6, 2026 THE 30-SECOND TAKE This week, three things converged that haven’t overlapped before. Models are now building models (OpenAI’s Codex debugging its own training runs). Agents are managing agents (Anthropic’s Opus 4.6 with agent teams, McKinsey running 25,000 AI agents among 60,000 total “workforce”). Infrastructure became the actual…
The collapse of traditional measurement when AI agents control commerce transactions Attribution vacuum is the collapse of traditional marketing measurement when AI agents intermediate commerce transactions. It happens because AI agents — ChatGPT, Perplexity, and emerging commerce agents — handle product searches, evaluations, and purchases without visiting tracked websites, making pixel-based attribution structurally impossible. The…
THE 30-SECOND TAKE ChatGPT launched advertising this week at $60 CPM with no conversion data. Read that again. Advertisers are paying 3x Meta rates for reach they can’t measure. This isn’t a temporary gap, it’s the new reality. The measurement system that built modern marketing assumed humans visited websites we could instrument. AI agents don’t….
A framework by Pål Erik Waagbø that identifies what data infrastructure organisations need before content optimisation can produce AI citations at scale. Last updated: June 2025 | By Pål Erik Waagbø The AEO Data Readiness Model (ADRM) is a three-layer strategic framework developed by Pål Erik Waagbø that determines whether an organisation has the data…
THE 30-SECOND TAKE AI is entering the physical world, and the race isn’t for the smartest model anymore. It’s for who owns your context. Apple just signed a multi-year deal to power Siri with Google’s Gemini while simultaneously racing to put AI in a wearable pin. The shift is clear: competitive advantage has moved from…