Another Point of View
OpenAI’s Presence launch and ServiceNow’s $1B in AI bookings prove AI agents are eating enterprise support. But every one of those deals covers problems that arrive as text or data. The interaction that resists automation is the one where someone has to see a physical thing that’s broken.
Read more →Enterprise AI budgets keep climbing, but a new SAP survey shows AI is delivering insights and better conversations — not the cost and time savings CIOs promised. In field service, that gap has a name: first-time fix rate. The missing input isn’t more orchestration. It’s visual context.
Read more →SAP is freezing hiring to fund agentic AI. PwC just built an agentic front office with OpenAI. But there’s a category of support ticket no autonomous agent will ever close on its own — because the problem lives in the physical world, and the AI can’t see it.
Read more →OpenAI Presence and the TeamViewer–ServiceNow deal both bet on voice and chat agents. But the hardest support problems are physical — and no voice agent can see a leaking valve. The visual layer is the missing primitive in the AI support platform race.
Read more →OpenAI launched Presence and ServiceNow’s AI business crossed $1B on the same day — agentic support is officially a production category. But the entire stack is text and voice, and it still can’t see the customer’s problem.
Read more →A $2,000 ebike vanished behind a forged signature, and months of chatbot loops recovered one-tenth of its value. What the deflection-first AI support model gets fatally wrong.
Read more →The default escalation reflex—dispatching a technician—is the most expensive decision in support. Remote visual support flips the sequence: see the problem first, then dispatch if necessary.
Read more →Enterprise AI keeps landing on the desks of people who already have the most resources. The people physically touching broken things — drivers, technicians, installers — are still waiting. That has to change.
Read more →Salesforce’s pay-per-resolution pricing is a genuine improvement over deflection metrics — but it only works if your support problems are text-shaped. Here’s why Heathrow gets 90% and GE Appliances gets 25%, and what that means for companies with physical products.
Read more →AI is sweeping the easy, text-solvable support tickets at scale — and the industry is celebrating deflection rates. Nobody’s talking about what happens to the tickets that remain: harder, more visual, more physical, and increasingly handled by fewer people with the same old tools.
Read more →
