Tech Mahindra and AWS have launched an Agentic Process Transformation Centre of Excellence aimed at moving enterprise AI agents beyond pilots and into operational business processes. The first joint deployment, a collections tool used by UK financial-services outsourcer Target Group, is reported to have delivered anticipated efficiency gains of about 40% in the areas where it has been rolled out.
AWS AI Watch analysis
What happened
The centre combines Tech Mahindra’s process and industry expertise with AWS cloud and AI services. The partners say it will develop and scale sector-specific agentic AI solutions across areas including telecoms, healthcare, banking, retail and manufacturing.
The first named solution is Collections Guru, an agentic system co-developed by Tech Mahindra and AWS for Target Group’s arrears-management operations. Target Group’s chief operating officer, Katie Pender, described the results as early and said the organisation was seeing anticipated efficiency gains of around 40% in the areas using the tool.
Why it matters
This is an attempt to solve the unglamorous part of enterprise AI: getting systems from an impressive demonstration into a workflow where money, decisions and customer relationships are involved. A centre that combines cloud infrastructure with process expertise could make deployment easier for companies that have the data but not the specialist team to turn it into a working agent.
The 40% figure is an early operational claim from the companies involved, not an independent benchmark. The announcement does not set out the baseline, sample size or precise measurement method, so the useful question is whether the result survives wider deployment rather than whether the launch has found a suitably glossy number.
Our read
The interesting development is not another promise that agents will transform everything before lunch. It is the partnership’s focus on repeatable business processes, with collections as the first test case. That is where agentic AI will either become useful or discover that real organisations contain exceptions, regulations and customers who do not behave like tidy flowcharts.
For buyers, the sensible takeaway is to ask for the operating evidence behind the headline: what the agent can decide, where people intervene, how outcomes are measured and what happens when the system is wrong.
What to watch
- Whether Tech Mahindra and AWS publish results from more customers and longer deployment periods.
- How much of Collections Guru’s work is autonomous and where human approval remains required.
- Whether the 40% efficiency estimate translates into measurable customer, recovery or cost outcomes.
- Whether the centre produces repeatable tools across other regulated industries, rather than a procession of bespoke pilots.
Discussion spark: Should companies trust agentic AI with collections and other sensitive workflows once early efficiency gains look promising, or should independent measurement and longer trials come first?
Sources and evidence
- How Tech Mahindra & AWS Operationalise Agentic AI at Scale – Technology Magazine (24 September 2026, 16:09 UTC)
not affiliated with or endorsed by Amazon Web Services (AWS)