Another Point of View
AI agents now get identities, audit trails, and memory. But when a support case involves a physical problem, all that context still dies at the handoff to the field technician.
Read more →AI agents now have payment credentials and workforce identities, yet they still can’t see a blinking error light. Agency without perception is a liability.
Read more →BCG’s new AI-first field service blueprint covers org design, agents, data, and change management — everything except the physical world where field service actually happens. Salesforce’s own research shows why that omission matters: 95% AI adoption, rising turnover, and technicians still arriving onsite without context.
Read more →“Autonomous resolution” quietly assumes the problem lives in a database. When the problem is a furnace, a router, or a machine on a factory floor, resolution requires seeing the state, diagnosing from visual evidence, and guiding human hands. Here’s the realistic autonomy stack for physical problems.
Read more →Every AI layer you add makes the remaining human-handled support harder — by selection effect, context stripping, skill atrophy, and measurement fog. The residual queue is a concentrate of physical, visual, ambiguous problems. Managing it takes eyes, not more automation.
Read more →Salesforce, Microsoft, and Uber cut thousands of support agents and booked the headlines as savings. But the money doesn’t vanish — it migrates to AI vendor contracts, a pricier escalation layer, field service, and churn. A CFO-level map of where those dollars actually land.
Read more →The costliest support cases aren’t the complicated ones — they’re the ones where the customer can’t put what they’re seeing into words. Description failure drives misdiagnosis, repeat contacts, and wasted dispatches, and AI voice agents inherit the same ceiling. The fix: stop making customers translate the physical world and see it instead.
Read more →The field service software market is headed toward $13.8 billion, and nearly every dollar of it optimizes the trip instead of questioning it. Enterprises just boosted AI spending 110% with underwhelming results — because the real bottleneck isn’t data fragmentation, it’s that nobody has seen the problem before the truck rolls.
Read more →Microsoft just credited AI support automation with $500 million in savings while Uber and CBA cut hundreds of service roles. But every job AI eliminated handled problems that arrived as text. The expensive physical layer of support — truck rolls at $150 to $1,000 each — hasn’t gotten a dollar cheaper, because its bottleneck isn’t conversation. It’s visibility.
Read more →Microsoft, Uber, and CBA just confirmed AI-driven customer service cuts, and Wall Street cheered. But when everyone deploys the same AI support stack, easy-ticket parity is guaranteed. The real competitive edge shifts to how you handle the escalations the bot can’t solve.
Read more →
