Inverclyde Council's £4,500 AI chatbot deployment presents a textbook case of misaligned success metrics masking implementation reality. Despite Clyde AI handling nearly 8,000 questions since launch and demonstrating strong engagement patterns—evidenced by high follow-up conversation rates across 640+ topic areas—overall call volumes to the customer service centre have risen rather than fallen. Council officers attributed this counterintuitive outcome to external demand spikes from the Greenock Central Project, yet this explanation warrants scrutiny. The chatbot's genuine performance indicators (2,000+ questions handled in June alone, positive user feedback, broad topic coverage) suggest the technology itself functions adequately, raising the question: are organisations deploying AI deflection tools without first establishing baseline demand patterns or accounting for how external events distort their effectiveness metrics?
The Inverclyde case exposes a critical gap between chatbot implementation and contact centre strategy. A £4,500 investment generating 8,000 question interactions represents reasonable throughput, but if those interactions don't correlate with reduced agent workload or call abandonment, the ROI calculation becomes murky. The council's acknowledgment that call volume increases are "masking business as usual" suggests they lack visibility into which queries the chatbot actually deflects versus which ones it merely documents before customers escalate to agents anyway. For teams evaluating or defending AI chatbot investments to stakeholders, this distinction matters enormously—engagement metrics and conversation counts tell a different story than contact centre efficiency gains. The real test will arrive once the Greenock Central Project concludes; if call volumes remain elevated, it signals the chatbot is functioning as a supplementary channel rather than a deflection mechanism, fundamentally changing how its success should be measured and justified.
Council call numbers rise despite rollout of AI chatbot designed to reduce enquiries Greenock Telegraph