Agentic Delivery Architecture

The unit of delivery is changing. The future shape is discovered, not forecast.

Traditional delivery split work across specialists. Agile brought them into a two-pizza team. AI changes the unit again: which skills must sit together, and where the boundaries between roles fall. What it does not hand you is a fixed new headcount ratio.

So we sell squads a specific way. Not everyone down to a two-person team, but the smallest safe and effective shape for each workflow, chosen from evidence and supported by an engineered system.

Book a call See the operating model

Do not predict the future organisation from today's job titles. Discover it from the work.

Emerging delivery shapes

There is no fixed squad size. Depending on the work, the right shape might be a full multidisciplinary squad, a smaller AI-amplified squad, a domain expert and an engineer, a single accountable professional orchestrating agents, or a deliberately larger team where judgement, customer care or operational risk require it.

START FROM the work shape is discovered, not fixed Multidisciplinary squad a larger human team, many skills together AI-amplified squad the same team, smaller, agents assisting Domain expert and engineer pod a Product Owner and a Forward Deployed Engineer One professional, many agents a single accountable owner orchestrating agents Larger team by design where judgement, care or operational risk require it
What holds true whatever the shape
Accountability A reusable platform Evidence over status Institutional learning

We discover the shape, we do not prescribe it

We do not move you to a two-person squad. We find the smallest safe and effective configuration for each workflow, and prove it before we scale it.

DISCOVER THE SHAPE from the work Observe how the work happens Identify what agents can do or support Define where human authority remains Test alternative team shapes Measure outcomes, risk, workload, skills Update workforce plans from evidence

We sell squads the DTA way

Most suppliers sell you a team. We sell you the right capability for the problem, and the evidence behind the choice.

Roles before bodies

We define the roles the outcome needs against SFIA, then assemble the smallest shape that can own the work. You buy the right capability for the problem, not seats.

The choice is evidenced

We pick the shape from how the work actually runs, and record why. Defensible when the board or the regulator asks, and revisable when the evidence changes.

Run inside our discipline

Every engagement runs inside DTA's operating model, with an accountable DTA lead, under the same decision and evidence discipline as our Accountable Bank demonstration.

One shared system. Many possible shapes.

The platform does not assume a team size. It supplies context, constraints, evidence, reusable knowledge and escalation to whatever shape the work needs, from a full team to a single accountable professional.

Accountability backbone Agentic paved road THE DELIVERY CELL Shape set by the work one professional, a pod, or a full team Agents amplify the work Agent contracts Named exec ownership Reference architecture Audit trail Mandates & permissions TDRs Trust ledger Human override Data & context Tools & connectors Code & deployment Evals & testing Observability Security & policy Cost & run-rate Reusable skills
Why this matters
Safe autonomyRepeatable deliverySupportable productionKnowledge stays with the client

This is the engineering view of a bigger shift

As AI becomes an active participant, every function evolves, not just delivery teams. That is a board-level conversation about how the whole enterprise changes, and it is the natural extension of this operating model. We can bring it to your leadership team.

Discuss the whole-enterprise view

Bring us a challenge worth solving.

Tell us the proposition, architecture decision, or transformation problem you are facing. We will tell you, honestly, whether and how we can help.

Discuss a challenge

Get the Observatory, our living models of banking, by email

Living models and Field Notes on the future of banking. Published when the model changes, not on a schedule.