New playbook: when your own AI model pays. Read
Westpac (banking) reported this on 4 Aug 2026 in company self-report (source). It belongs to the agents, traceability and evals layer.
Specialist AI agents on AWS Bedrock AgentCore inside home-loan and credit-card applications: agents that extract payslip data and verify it against policy, and an account-scrutiny agent for transactions and bank statements, working ahead of human bankers.
Processes 32,000+ payslips a week and 1.5M+ transactions in production; saving 150,000+ staff hours that Westpac says it reinvests in customers; time to put a new agent capability into production fell from ~6 months to about 4–6 weeks.
The number covers lending document verification workflow; the hours are reinvested staff time; a headcount reduction is not reported.
“…our bankers spend less time wrestling with documents and much more time helping our customers,” Westpac said.
Each agent has one narrow task, a named owner and an action log.
Read the guide to agents, traceability and evals.
$10M
EBITDA uplift from two agent systems, Partners Group says
Two agentic systems built on a cleaned proprietary policy-level dataset: one runs new-client policy processing, the other automates human-intensive policy checking; delivered by a seven-person team with Version 1 (source).
2–3 minutes
per ocean or air freight quote, down from one to three days
Hundreds of in-house AI agents embedded across the shipment lifecycle (quoting, orders, appointments, missed-pickup checks) and now ocean and air quoting, ocean booking and customs brokerage, built and run under a lean waste-removal discipline by 450+ in-house engineers and data scientists (source).
80%
less manual submission triage time
An agentic intake platform that reads broker submissions, digitises them and triages risks into a decision-ready workflow for underwriters across more than 20 countries (source).