CN

SmartSupply AI

Centrica PLC — Demo Access

Sign in to continue

Need access? Email kemphil@amazon.co.uk

Access restricted to amazon.co.uk, amazon.com, and mckinsey.com email addresses.

AI Cost & Successor-Agent Pipeline

-8.3%

£0.0165

LLM cost / contact

Target: £0.011 · Baseline: £0.018

-5.1%

£0.21

Cost-to-serve / contact

Target range: £0.12–£0.30

-120ms

870ms

Avg. inference latency

Across active model fleet

gap

£0.0055

Gap to LLM target

Closes as successors promote

Inference Cost & Model Mix

Model mix by vendor (share %)

Cost per contact by vendor (£)

Cost Records — drill into spikes

Vendor Agent type Cost / contact Latency Share Status

Successor-Agent Board

Projected Savings Ranking

Ranked by projected annual saving
Candidate Vendor Projected saving Quality gate Status Action

Cost Spike Detail