A role-level map of a wholesale bank's Capital Markets and Commercial Lending operation that decides, for every role, whether the work should stay onshore, move to a Global Capability Centre, be run by an agent, or some combination of the three.
The bank runs its Capital Markets and Commercial Lending work entirely onshore today — there is no GCC yet. AgentShift takes that operation, breaks it down to the role level, and scores every role for two things at once: how feasibly an AI agent can do the work, and where the work should sit.
Each role is placed on a five-way decision: keep it onshore, hand it to an agent, build it in a Global Capability Centre, run it as a GCC pod with agents, or wait. The result is a single picture a leadership team can act on — built from 102 roles across 10 function columns in the two business lines.
Numbers, headcounts and scores on this map are built from a wholesale-bank operating model. The interactive module lets you re-size them to your own organisation.
Standing up a GCC and adopting agents are usually treated as separate programmes. AgentShift answers both at once, role by role.
The Agent-Core set — high-volume, bounded operations work in trade processing, settlements and servicing.
Movable work that pays back in a GCC versus work an agent handles in place — so you never build a pod you will retire.
Front-office judgement, client-facing coverage and regulator-anchored roles, where people stay accountable.
Capacity released, net savings and a three-stage view — onshore cost, after a GCC move, after agents.
The decision sorts into a sequence: high-confidence operations first, judgement-heavy work approached later.
A regulatory exposure score on every role, so autonomy and location are set by risk, not by appetite.
The same four steps run across both business lines, so a Capital Markets map and a Commercial Lending map read the same way.
Break the operation into function columns, process areas and the job roles that staff them.
Rate each role on transaction intensity, decision complexity, regulatory exposure and tool readiness.
Place every role in one of three agent tiers and one of five feasibility zones.
Turn the map into a roadmap and a business case, lowest-risk roles first.
Each role's score places it in one of three tiers — the part an agent plays in the work. These are the colours used throughout the module.
High transaction intensity, low decision complexity. Settlements, confirmations, reconciliations and payments processing. People move to exceptions and oversight.
Agents take the sub-tasks; people own judgement and exceptions. Research, credit analysis, electronic execution and covenant monitoring.
High judgement or regulatory exposure. Trading, coverage, M&A advisory, underwriting and credit approval. Invest in people; use agents as tools.
The tier says what an agent can do. The feasibility zone says what to do with the work — the heart of a greenfield GCC strategy, where the whole estate starts onshore.
Stays where it is. Client-facing, judgement-heavy or regulator-anchored work.
Handed to an agent in place. Mature tooling, bounded task — never built in the GCC.
The right candidate, the market is not ready. Hold, then agentise on a trigger.
A GCC pod and agents together — the GCC as a transformation lab from day one.
Move to the GCC as labour arbitrage; standardise first, agentise over time.
The same engine drives the live module. Change the layer — Agent Impact, Decision, Readiness or Feasibility — and the treemap recolours. Change the transform — Savings, Capacity or Economics — and it reads out the value.
Set the estate size, the base cost and the business line to match your own organisation, then read the capacity released, the net saving and the onshore-to-GCC bridge straight from the map. The views above are taken from the live module.
Switch layers — Agent Impact, Transaction, Decision, Readiness, Feasibility. Drill from function column to process area to job role. Re-size the estate and base cost to model your own organisation, and read the onshore-to-GCC bridge and the three-stage economics.