BNY says it has trained 100% of its staff on Eliza, its Microsoft AI-powered system, putting the financial firm’s AI rollout firmly in the people-and-process phase rather than the shiny-demo phase. The programme has delivered more than 400,000 hours of training and created a tiered path from basic AI literacy to hands-on agentic AI work.
Microsoft AI Watch analysis
What happened
According to Microsoft’s customer account, BNY’s training is split into three levels: Explorer covers AI fundamentals, Builder introduces more advanced uses, and Pioneer is a smaller group of change agents who complete a 40-hour bootcamp focused on agentic AI. Employees can attend live or asynchronous classes, giving the programme a little more structure than “please watch this 12-minute video and transform the business by Friday”.
BNY says Pioneer cohorts have grown from 200 to 800 participants. Those staff are intended to help colleagues identify useful applications, experiment with workflows and spread lessons across teams. Saed Shonnar, BNY’s head of AI enablement, said the aim is to help people “rewire their workflows with AI”.
Why it matters
The useful detail here is not simply that a large company is using AI. It is that BNY is treating adoption as an organisational capability, with training, internal champions and executive participation running alongside the technology.
That matters because the gap between buying an AI system and getting useful work from it is usually made of people, habits and incentives. A workforce may have access to an AI tool while still lacking the confidence, judgement or practical examples needed to use it well.
Our read
This is a credible sign that enterprise AI is moving into the less glamorous work of implementation. Four hundred thousand training hours is substantial, but it is not proof that Eliza is improving revenue, lowering costs or producing better decisions. It is evidence of preparation, not a finished return on investment.
BNY’s approach is worth watching because it makes adoption visible. The firm is not only counting licences or enthusiastic executive remarks. It is building a network of trained users who can test where the system fits and where it does not. That is a more promising recipe than sprinkling chatbots over an organisation and hoping culture takes care of the rest.
What to watch
- Whether BNY publishes concrete productivity, cost or quality results from Eliza.
- Which workflows Pioneer teams change first, and whether those changes spread beyond early adopters.
- How BNY handles review, privacy and accountability when agentic AI enters regulated financial work.
- Whether the training programme remains effective as Microsoft’s AI tools and capabilities change.
Discussion spark: Should large firms measure AI adoption by training and usage, or refuse to call it a success until they can show audited gains in revenue, cost or decision quality?
Sources and evidence
- To innovate faster, BNY upskills 100% of staff on Eliza, powered by Microsoft AI – Microsoft (24 September 2026, 18:26 UTC)
not affiliated with or endorsed by Microsoft