Rotherham NHS Foundation Trust cuts IT helpdesk calls 28% with AI agent

TL;DR:

• The Rotherham NHS Foundation Trust has cut IT helpdesk call volumes by 28% in roughly three months after deploying an AI-powered autonomous agent through Netcall’s Liberty Converse on 28 January 2026.
• The agent resolves routine queries in real time, raises pre-categorised IT service tickets where escalation is needed, and routes complex issues to the right specialist; 41% of all IT queries are now handled through self-service or AI.
• A phase-two rollout targeting out-of-hours support is planned, aimed at reducing on-call staffing costs and avoidable disruption for IT teams.

The Rotherham deployment is one of the cleaner concrete UK NHS AI metrics published this quarter, and it lands at a useful moment. The Ada Lovelace Institute has just warned UK policymakers that AI productivity claims need more methodological scrutiny ([our coverage]); this is the kind of single-trust, time-bounded number that the Ada Lovelace critique argues should be read with care, even when (as here) the figure looks credible.

A concrete number in a noisy debate

The 28% headline figure is precise enough to be useful — the trust gives the start date (28 January 2026), the vendor (Netcall), the platform (Liberty Converse), and a corroborating second metric (41% of queries via self-service or AI). Christine Hazlehurst, Rotherham’s head of IT service management and support services, was honest about the constraint that pushed the deployment: rising IT support demand “without the ability to increase headcount”. The Ada Lovelace test would ask: what is being measured here that is not measured? Service quality from the user’s side, equity of access across staff groups, lifetime cost of the deployment, and the opportunity cost of the team’s attention during onboarding are all absent from public reporting. That does not invalidate the call-volume reduction — but it does mean a single trust’s number is data, not yet evidence of a system-wide pattern.

What makes the case study more interesting than typical vendor success stories is the operational logic. Netcall’s Liberty Converse was already used elsewhere in the NHS for patient engagement; redeploying agentic AI for staff-facing internal support is a structurally different use case with different risks. John Clarke, Netcall’s head of client solutions for health, called this “an important shift” — and for once the vendor framing is plausibly correct. UK NHS trusts typically procure separately for staff-facing and patient-facing tooling; if Rotherham’s pattern spreads, that procurement boundary may move.

Looking forward

Watch for similar trust-level case studies from peer trusts and from the NHS England transformation directorate. For UK SMEs in IT service management, agentic AI and NHS-adjacent software, the practical procurement signal is that buy-side conversations are now anchored on “show me your call-deflection number” rather than “show me your model accuracy”. For trust IT directors, the phase-two out-of-hours target is the one to track — if it lands at a similar 28%-class deflection, the staffing-cost case becomes the dominant argument, with all the workforce-relations implications that brings.