In one of the most candid sessions of the series, Martin and Kevin are joined by Alastair Skeoch of BPO Collection one of MEGA’s earliest enterprise customers for an unfiltered look at what voice AI deployment actually looks like in a regulated DCA. Alastair walks through the eighteen-month journey from first conversation to live operation, the compliance bar (ISO 27001, Cyber Essentials Plus, UK data residency) BPO required of MEGA, and why he chose to introduce AI into the ID&V wait-time rather than replacing the agent conversation.

‍
Martin then shares live performance data from BPO’s deployment: customer identification climbing from an 11% baseline to 76% by week 14, and 85% in the most recent week driven by CLI matching, reference-number optimisation and continuous refinement of name, voice and date-of-birth matching. The episode closes with a discussion of the next wave of use cases: out-of-hours support, post-call summarisation, direct-debit guarantee playback, and richer back-end integrations each one designed to free up human agents for the conversations they are uniquely good at.

Topics covered 

• From contract to go-live: how BPO Collection prepared for voice AI
• ID&V identification rates: from 11% to 85% through iterative optimisation
• CLI matching, name flexibility and the realities of multi-client data
• Why partnership beats vendor relationship in regulated AI deployments
• Future use cases: out-of-hours support, summarisation, agent assist
• Hidden wins: agent retention, reduced internal transfers, better customer routing