Bobbi Virtual AI went live on the South Yorkshire Police website on 1 October, answering non-emergency queries that would otherwise add to a 101 line already taking more than 1,000 calls a day. It follows an earlier AI system that began triaging those calls in August. Neither tool reports a crime or replaces a human officer, but their arrival sits inside a much larger shift already under way across UK call centres, where chatbots now handle a significant share of customer contact and employers are actively planning headcount reductions.
What has actually launched
South Yorkshire Police introduced Bobbi Virtual AI on Thursday, 1 October, as a chat tool on its website designed to deal with non-emergency concerns. The force says it receives more than 1,000 calls to its 101 non-emergency number every day, with some callers waiting 20 minutes or more to speak to staff during busy periods.
Chief Superintendent Cherie Buttle, head of the Force Control Room, said the tool would “be a useful tool to help reduce calls to 101, save members of the public time, and provide concise and reliable answers to frequently asked questions.” South Yorkshire Police has been explicit that Bobbi cannot be used to report a crime and is not an alternative to calling 999; it runs alongside the existing 101 line, online reporting forms and station front desks rather than replacing any of them.
A distinguishing feature, according to the force, is that Bobbi draws only on closed-source information supplied directly by South Yorkshire Police, rather than searching the open internet the way general-purpose AI assistants do. The force says this is intended to keep answers accurate to local procedures and to keep any data shared in a chat confidential to the force, rather than exposed to a third-party model trained on public web data.
Bobbi is not a South Yorkshire invention. The same system has already been deployed by Hampshire and Isle of Wight Constabulary, Thames Valley Police and Humberside Police, and South Yorkshire is simply the latest force to adopt it.
A second, separate AI system is already triaging the calls that do get through
Bobbi is not South Yorkshire Police’s first AI deployment on its non-emergency line. From 10am on Tuesday, 18 August, the force began using a natural language processing (NLP) system, developed jointly by the Home Office, the National Police Chiefs’ Council and Vodafone, to triage incoming 101 calls before they reach a human call handler.
Rather than placing every caller into an identical queue, this system listens to why someone is calling and matches the request against a set of likely reasons, redirecting callers whose issue is not actually a police matter toward the right organisation, such as a local council, NHS 111 or the 105 power-cut helpline. South Yorkshire Police has said the volume of what it calls “ridiculous calls”, citing examples such as delayed pizza deliveries and slow pub service, wastes time that could otherwise go toward policing. The force has stated the system went through Home Office testing intended to reduce the risk of callers being misdirected because of accent or phrasing, following trials across 15 forces nationally.
Together, the two systems represent different stages of the same call: the NLP tool sits ahead of the queue, sorting calls before a human is involved, while Bobbi sits on the website as a self-service option a caller might choose instead of dialling in at all.
Is this part of AI replacing call centre jobs more broadly?
South Yorkshire Police’s own framing, that Bobbi is “an alternative to, not a replacement for” 101 calls, sits against a far larger pattern of AI deployment in customer-facing roles across the UK economy, where the distinction between alternative and replacement is less settled.
UK contact centres employ an estimated 1.3 million people. Industry analysis suggests AI chatbots already handle a substantial share of first-line customer service queries in sectors such as banking, retail and telecommunications, with some estimates putting that figure at 40 to 60% of queries in those industries.
The job impact is not hypothetical in energy retail, at least. Centrica, the owner of British Gas, confirmed plans to cut 1,300 call centre and back-office roles, a 14% reduction over two years, citing a 20% decline in phone call volumes alongside a shift toward AI-driven self-service and digital channels. Internationally, Reuters reporting on India’s business process outsourcing sector, which employs around 1.65 million people in call centres, payroll and data handling, cited investment bank Jefferies’ prediction that AI adoption could cut call centre revenue in that sector by as much as 50% over five years.
Survey data points the same direction for the UK specifically. The Chartered Institute of Personnel and Development’s autumn 2025 Labour Market Outlook, surveying more than 2,000 UK employers, found 17% expected to cut headcount due to AI within 12 months. Separate research from Gravitee, based on a September 2025 survey of technology leaders at 250 large UK companies, found 65% of executives said they would reduce headcount before the end of 2026, and most large companies expected to have between 16 and 20 AI agents operating within their organisation by the end of this year.
Not every projection points to net job loss. The Tony Blair Institute estimates that AI could ultimately displace between 1 million and 3 million UK jobs, with peak annual losses of 60,000 to 275,000 positions, but describes these figures as modest relative to the roughly 450,000 jobs lost annually across the UK economy over the past decade through ordinary churn. Its central scenario projects 545,000 net job losses alongside GDP gains of 3.1%; a worst-case, full-displacement scenario would see up to 1.5 million jobs eliminated with no offsetting GDP gain.
West Midlands Police has already run a trial that frames the technology differently, as an aid to safety rather than a staffing tool. Its AI call-answering system, nicknamed Andi-Esra, prioritised nearly 800 vulnerable callers, including victims of domestic violence, people at risk of suicide and people reporting missing children, and answered around 17,000 calls over a two-month trial, with Police and Crime Commissioner Simon Foster noting that some of those calls might previously have gone unanswered if a human handler was not immediately free.
The practical reading for South Yorkshire
For residents, Bobbi and the NLP triage system do not currently change the basic options for contacting South Yorkshire Police: 999 for emergencies, 101 for non-emergencies, and online or in-person reporting routes remain intact. What is changing is what happens before a human answers, with software increasingly used to filter, redirect or pre-empt contact that does not need to reach a person at all.
Whether that constitutes job replacement, as it plainly has done in sectors such as energy retail, or remains confined to call deflection, as South Yorkshire Police currently describes it, is likely to depend on how call volumes and staffing levels at the force move over the coming months, figures which have not yet been published.
Sources: The Star, “South Yorkshire Police introduce AI assistant to deal with non-emergency calls” (1 October 2026); Yorkshire Live/Examiner Live, “AI ‘officer’ now answering South Yorkshire Police’s non-emergency queries” (1 October 2026); The Yorkshire Press, “South Yorkshire Police Launches AI Assistant Bobbi To Help Cut 101 Call Times”; Hello Rayo, “AI chatbot named ‘Bobbi’ launched by South Yorkshire Police to handle non-emergency reports”; Time News, “South Yorkshire Police Launches Bobbi AI Chatbot to Cut 101 Wait Times”; Yahoo News/BBC, “Police force launches AI tool to reduce 101 calls”; ITV News Calendar, “South Yorkshire Police among forces using AI on its 101 service to screen ‘ridiculous calls’” (18 August 2026); Natpoint, “Tech News: AI To Triage UK Police 101 Calls”; West Midlands PCC, “PCC praise for police virtual call handling assistant”; CIO Bulletin, “Why British Gas AI Chatbots are Replacing 1300 Jobs Despite Massive Public Backlash?” (25 July 2026); ITBrief UK, “AI agents set to replace thousands of UK jobs by end of 2026”; Dispatch Times, “How Many UK Jobs Has AI Already Replaced? The 2025–2026 Data Explained”; Reuters (via Yahoo News), “Meet the AI chatbots replacing India’s call-center workers”; CAIO Weekly, “Which UK Jobs Are Most Affected by AI in 2026?”