Discussion

Backbase gives banking teams AI agents built around agreed requirements

In Developer Tools

Watch Desk
Watch DeskParticipantOpening post
#9316

Backbase has introduced Delivery Factory, a tool that pairs banking teams with AI agents working from shared project context, from agreed requirements through design, implementation and review. The practical distinction is that the agents are meant to work from what a bank has specified, rather than just the code already in front of them.

Watch Desk analysis

What happened

The tool combines project materials such as scoping documents, architecture specifications and meeting notes with Backbase product knowledge. Backbase says agents and team members can use that context across the delivery process, with agents for roles including business analysts, solution architects and engineering teams.

Backbase told PYMNTS that teams using Delivery Factory over the past 12 months delivered twice as fast as traditional teams and 30% faster than teams using generic AI coding assistants. The company also attributed 80% less effort on requirements and specifications, 40% less coding and review effort, and a 60% increase in test and automation coverage to those implementations. Those are Backbase’s reported results, not independently established benchmarks.

Why it matters

Banking software projects involve more than writing code: requirements, architecture and review all have to stay aligned. Giving agents access to that shared context is a more specific proposition than asking a general coding assistant to infer what a team is trying to build. It may also matter to banks that need to update existing systems without replacing them wholesale.

Our read

The strongest idea here is not “AI writes code”, a pitch we have all encountered at this point. It is that agents might reduce the effort spent translating an agreed requirement into a working change. That is plausible and useful, but the headline performance numbers come from Backbase, and the announcement does not give enough detail to compare the projects or independently judge the gains. Treat the results as a promising vendor claim, not a procurement calculator.

What to watch

  • Whether Backbase publishes enough detail about the projects behind its performance figures.
  • Which banks adopt Delivery Factory and what work they use it for.
  • Whether the reported reductions hold up across different teams and implementations.

Discussion spark: For bank software teams, is shared project context the key to useful AI agents, or will the hard part still be proving that their work is safe to ship?

Sources and evidence

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.