Thames Valley Police AI assistant Bobbi handles 200 calls a day

TL;DR:

  • Thames Valley Police and Hampshire & Isle of Wight Constabulary’s AI assistant, Bobbi, handles around 200 non-emergency conversations daily.
  • About 45% are resolved without human intervention, freeing an estimated 3,200-plus staff hours.
  • Every day it flags at least one high-harm offence, and on average two cases of violence against women and girls, for escalation.

A UK police force is putting hard numbers behind an AI deployment that goes beyond a pilot. Bobbi, a 24/7 digital assistant launched in November 2025 by Thames Valley Police and Hampshire & Isle of Wight Constabulary, now handles roughly 200 non-emergency conversations a day, with about 45% resolved without a human stepping in.

Freeing staff, flagging harm

The stated payoff is capacity. Bobbi is estimated to free more than 3,200 staff hours, letting contact-centre staff focus on emergencies and sensitive cases while unresolved queries are routed to human-operated channels. Mike Lattanzio, chief digital and information officer for the forces’ joint operations unit, called it “a pioneering moment in policing” and a way to “rethink how we manage non-emergency demand.”

More striking is the safeguarding role. Every day, the system identifies at least one high-harm offence needing escalation and, on average, two cases of violence against women and girls for immediate action. In one instance, a young person unable to speak aloud used Bobbi discreetly to seek help over a high-risk domestic incident; it detected the risk, escalated to a supervisor and handed off to a live operator who dispatched officers.

The deployment fits a wider UK pattern of public bodies chasing time savings from AI — echoing Lancashire council’s claim that AI is giving social workers time back and Whitehall’s focus on back-office automation. What sets Bobbi apart is that the metrics extend past efficiency into risk detection, where the cost of an AI error is measured in safety rather than time.

Looking forward

Lattanzio argues Bobbi lowers a common barrier — the instinct not to “be a burden” — by offering a contact route that did not exist before. The open question is oversight: as more forces let AI triage public contact, the reliability of its risk-flagging, and the human checks around it, will decide whether the safeguarding gains hold up under scrutiny.