GCC Intelligence Series

AgentShift Index

A Strategic Decision Enablement Framework for Agentic AI Transformation spanning both parent-carrier and GCC role universes (~800 archetypes across Life, P&C, Health, Reinsurance, and Specialty/Lloyd's) — mapping AI agent impact, jurisdictional anchoring, and a 5-way decision (Keep / Agent-Direct / Wait-then-Agent / Hybrid / GCC-Lift) into a Parent → GCC → Agent bridge.

The Core Framework

Every insurance role is scored on three dimensions: Transaction Intensity (T) — submissions, transactions, policy events, claims volume — Decision Complexity (D) — underwriting judgment, complex-liability adjudication, actuarial inference — and Regulatory Blast Radius (RBR) — Solvency II / IFRS 17 / NAIC / conduct (FCA Consumer Duty) / market-conduct exposure. The composite weights volume as opportunity and judgment as constraint, with RBR acting as a regulatory dampener.

Higher transaction + lower decision = greater agent impact. Each role then falls into one of three restructuring zones:

Agent-Core (8–10)— agent IS the worker, 60–80% FTE reduction

Agent-Augmented (5–7.9)— humans lead, agents on subtasks, 30–50% gain

Human-Core (0–4.9)— agents are peripheral tools, 10–20% uplift

Feasibility Classifier — the strategic question: for every role across the whole bank, where should the work actually live?

Keep Onshore— jurisdiction / client / regulator anchor (geo_lock≥7, client≥8, or RBR≥9)

Agent-Direct (skip GCC)— high impact + production-ready, automate at parent (impact≥7.5, readiness≥7)

Wait-then-Agent— hold onshore briefly, agentise at parent in 12–24 mo. Offshore setup costs don't pay back over that horizon.

Hybrid (GCC + Agent)— GCC as transformation lab, agents augment from day 1 (impact 4.5–7.5, D≥4)

GCC-Lift (classic)— labour arbitrage still wins, AI peripheral (T≥5, low geo-lock)

What’s Inside

8 Diagnostic Layers: Agent Impact, Transaction Intensity, Decision Complexity, Agent Readiness, Offshoring Potential, Feasibility (new in v5), Restructuring Mode, and Headcount Detail — each recolors and re-aggregates the treemap to reveal a different dimension of workforce exposure.

Two Role Universes: Toggle between Parent (~350 onshore archetypes across London / Zurich / Munich / Bermuda / Hartford / Des Moines / Dublin / Singapore / Hong Kong), GCC (~450 captive-center roles across Bengaluru / Hyderabad / Pune / Gurugram / Manila / Krakow / San José), or Both. The Location filter rewires every chart, bridge, and economic calculation.

3 Transformation Layers: Savings Map, Capacity Released, Upskill Potential — trigger a scorecard + dual bridge (People + Economics, USD $M/yr).

Feasibility Layer: A 5-way classifier per role — Keep Onshore, Agent-Direct, Wait-then-Agent, Hybrid, GCC-Lift — driven by geo-lock, client proximity, RBR, readiness, and impact. Adds a Parent → GCC bridge (Parent FTEs → eligible → blocked by jurisdiction → blocked by client-proximity → movable to GCC → of which agent-replaceable before move) and three-stage economics (Parent cost → GCC cost → Agent cost).

Economics Tab: Deployment-wave timeline (Now / Next / Later) plus Azure-stack bottom-up cost model.

Adjustable sizing: GCC headcount (75–10,000+), Parent FTEs (5,000–200,000+), and average base CTC (USD). All math recomputes live.

How to Read the Treemap

■

Rectangle = a role. Each tile represents one insurance role (e.g., FNOL Triage Agent, Bordereaux Technician, Open Market Underwriter).

▢

Area = workforce share. Larger tiles = more people in that role as % of total GCC.

■■■

Color = selected metric. Toggle layers to recolor: green (low impact) to red (high impact).

☜

Hover = full detail. Tooltip shows all scores, transformation model, upskill pathway, and target role.

▼

Filters = carrier type & department. Narrow to Life, P&C, Health, Reinsurance, or Specialty/Lloyd's carriers, or to a single function.

Start Exploring ↓

~800 roles: parent-carrier + GCC archetypes across Life, P&C, Health, Reinsurance, Specialty/Lloyd's, Composite Multi-line

7 carrier archetypes: Life & Annuities, P&C, Health, Reinsurance, Specialty/Lloyd's, Composite Multi-line

5-way Feasibility classifier: Keep / Agent-Direct / Wait-then-Agent / Hybrid / GCC-Lift

3-stage economics: Parent → GCC → Agent run-rate (USD $M/yr)

Configurable GCC headcount, Parent FTEs, and base CTC drive all bridge math

▶ Quick Glossary

Layer

Transform

Low

High

GCC Size

Parent FTEs

Base CTC (USD/yr)

Location

Carrier Type Department

Insurance Value-Chain Bridge — Submit → Underwrite → Bind → Service → Claim → Recover

The insurance value chain — where agentic AI lands hardest. Submit and Claim are the gravity wells; Underwrite is the regulated bottleneck. FTE share is computed on the currently filtered universe (respects Location / Carrier Type / Department).

Roles that don't map cleanly to a single value-chain stage (e.g., Risk & Compliance, Finance, Tech, HR) are excluded from this bridge but remain in the treemap above.

▶

AgentShift Index — Methodology & Layer Definitions

This index adapts Andrej Karpathy’s US Job Market Visualizer for the Indian GCC context. Karpathy scores occupations on: is the work fundamentally digital? For GCCs, nearly every role is already digital. The AgentShift Index replaces that axis with a transactional–decisional framework.

v5 extension — Strategic Decision Enablement Framework for Agentic AI Transformation: The Insurance v1 module answers "for every role across the global insurance value chain — Submit, Underwrite, Bind, Service, Claim, Recover — given AI's trajectory, where should the work actually live and when does it move?" Two universes are loaded (~450 GCC + ~350 parent-carrier archetypes spanning London, Zurich, Munich, Bermuda, Hartford, Des Moines, Dublin, NYC, Singapore, Hong Kong). Every role gets two new dimensions: Geo-Lock (jurisdictional anchor) and Client Proximity (face-to-face intensity). A 5-way feasibility classifier resolves into Keep Onshore · Agent-Direct · Wait-then-Agent · Hybrid · GCC-Lift. The Feasibility layer renders a Parent → GCC bridge and a three-stage economics bridge (Parent cost → GCC cost → Agent cost).

Agent Impact = (0.6 × T + 0.4 × (10 − D_adj)) × RBR_factor | D_adj = 10(D/10)^1.4 | RBR_factor = 1 − RBR/20

Higher transaction intensity + lower decision complexity = greater agent impact. A score of 9.5 (Invoice Processing) means agents are the primary worker. A score of 1.0 (Data Scientist) means agents are peripheral tools. Entry-level roles (0–3 yrs) average 8.0 impact; senior roles (7–12 yrs) average 3.9. The Transformation Model resolves into two parallel bridges: a People Bridge (in-scope FTEs → capacity released → upskilled & redeployed → net headcount) and an Economics Bridge (cost pool → released labour cost → agent platform cost → new run-rate, with net savings in USD $M/yr). Upskill (v3) is a composite of education headroom, experience plasticity, decision-complexity adjacency, and restructuring mode — not a function of agent impact.

Agent Impact Score

Primary index. Composite of T and inverse D. Rectangle area = workforce %. Color: green (low) to red (high).

0–1 Lloyd's Active Underwriter • 4–5 Reserving Actuary • 8–10 Bordereaux Technician / FNOL Triage

Transaction Intensity (T)

How repetitive, rule-based, volume-driven, and SOP-adherent the work is.

0–1 Lloyd's UW judgment • 4–5 Mid-market UW • 8–9 FNOL intake / submission triage • 10 Bordereaux ingestion

Decision Complexity (D)

How much judgment, ambiguity resolution, and stakeholder navigation. Color inverted: red = low (automatable), green = high.

0–1 Doc generation • 4–5 Reserving methodology • 8–10 Complex liability adjudication / specialty wordings

Agent Readiness

Tool maturity for this role. High impact + high readiness = immediate transformation candidate.

0–3 Early R&D • 4–6 Emerging • 7–9 Production-ready • 10 Off-the-shelf

Offshoring Potential

Combined: (a) parent→GCC migration upside + (b) GCC→agent displacement potential.

0–3 Fully offshored • 4–6 Partial • 7–10 Major opportunity

Feasibility

The strategic decision per role. A 5-way classifier driven by Agent Impact, Readiness, RBR, Geo-Lock, and Client Proximity.

Keep: geo/client/regulator anchor (geo_lock≥7 or client≥8 or RBR≥9)
Agent-Direct: impact≥7.5 + readiness≥7 + RBR≤6 — skip GCC
Wait-then-Agent: impact≥6, readiness 4–7 — hold onshore, agentise at parent in 12–24mo (offshore payback horizon too long)
Hybrid: impact 4.5–7.5, D≥4 — GCC as transformation lab
GCC-Lift: T≥5, low geo-lock — classic labour arbitrage

Geo-Lock & Client Proximity (v5)

Geo-Lock (0–10): jurisdictional / regulator / physical-presence anchoring. Client Proximity (0–10): face-to-face relationship intensity. Together these decide whether a role is offshorable regardless of its AI exposure.

GCC roles: typically 0–4 on both • Parent roles: hand-scored (Lloyd's UW 10/9, field adjusters 7/9, regulatory lawyers 9/5, cat modellers 7/3, bancassurance reps 9/9)

Restructuring Mode

Three zones based on Agent Impact Score.

8–10 Agent-Core: 60–80% FTE reduction
5–7.9 Agent-Augmented: 30–50% productivity gain
0–4.9 Human-Core: 10–20% efficiency uplift

Savings Map

Net savings per role = (capacity released × cost band) − agent deployment cost. Colors show where the money is.

Capacity Released

FTEs freed by automation. Tiles show FTE count and automation %. Drives the transformation waterfall.

Upskill Potential (v3 formula)

Composite 0–10. u = (0.35·edu + 0.35·exp + 0.30·d) × (0.4 + 0.06d) × mode_f × 1.15, clamped 0–10. Edu headroom: 12th=2, Bachelor's=5, Master's/CA=8. Exp plasticity: 0–3y=8, 3–7y=7, 7–12y=4, 12+y=2. Mode factor: Human-Core=1.0, Augmented=0.85, Agent-Core=0.6 (work itself is being eliminated). Decision-complexity term captures adjacency to higher-value roles. Each role retains its specific target role, certifications, trainings, and learning path.

Cost model: Salary bands are multipliers of a base CTC unit (1x). Plug in your organization’s base (e.g., 1x = ‎₹4 LPA) to convert all figures to absolute ‎₹. Agent cost ratios range from 0.15x (mature tools) to 0.45x (early-stage). Upskill affinity scores factor in education headroom, skill adjacency, and target role availability within the same department.
Applicable to GCCs with 75+ employees. Adjust the GCC Size input above to model your organization.

Conversation Guide — Insurance GCC AgentShift

Designed for meetings with insurance executives — Chief Underwriting Officer, Chief Claims Officer, Chief Actuary, COO, CFO — sponsoring the Insurance GCC. Three scenario playbooks below.

Scenario 1: "Where should we deploy agents first across the insurance value chain?"

Setup (2 min): Open at All Lines. Department treemap shows 16 insurance towers across Submit → Underwrite → Bind → Service → Claim → Recover. Tile SIZE = workforce share. COLOR = currently selected metric. Toggle through the layers to show how the view re-aggregates live.

Discovery (5 min): Ask "Which function has the largest operations footprint?" Click that department to drill in. Red/orange tiles surface the Agent-Core and Agent-Augmented role candidates.

Quantify (5 min): Switch to Savings Map. Enter the client's GCC headcount and average base CTC (USD). The Scorecard shows FTEs released, net FTE exit, and net savings in $M/yr. The People Bridge and Economics Bridge tell the same story visually.

Close (3 min): Switch to Upskill Potential to show redeployment is real, then open the Economics tab for a deployment-wave roadmap and Azure run-rate cost model.

Scenario 2: "We are a P&C carrier — show me our view"

Filter: Set Carrier Type = P&C / General. The treemap re-weights: Claims, Underwriting, Policy Admin and Cat & Loss Modelling grow; Investments and Actuarial (Life) shrink. This matches the carrier's actual operating model.

Key point: "Most consultants will tell you to automate FNOL. That is table stakes. The real opportunity is in submission triage + bordereaux + claims subrogation — where volume is 5–10x the FNOL pool and net savings per FTE is highest. The Bengaluru GCC already has the bordereaux ops — the question is whether you centralise more or skip to agent-direct at the parent."

Scenario 3: "Build the business case"

Three artefacts: (a) Capacity Released — FTEs freed (28-38% on a typical insurance GCC mix; lower in Underwriting + Actuarial, higher in Claims Ops + Policy Admin) and absorption via upskilling (60-75%). (b) Savings Map — net $M/yr after agent platform cost. (c) Economics tab — phased rollout (Now / Next / Later waves), Azure run-rate per agent-FTE, and one-time build by complexity tier.

Close: "This is a diagnostic view. The next step is a 4-week deep-dive where we map YOUR role inventory, validate scores against your process data, and produce a sequenced 18-month transformation roadmap with confirmed Azure unit economics."

Data Confidence (internal use)

• Source: KPMG Insurance Sector Taxonomy v1 — ~800 roles across Life, P&C, Health, Reinsurance, Specialty/Lloyd's, Composite Multi-line. Cross-referenced against Lloyd's market structure, NAIC Annual Statement schedules, Solvency II QRT taxonomies, and major-carrier published org charts.

• T/D scores: Rules-based from department context + role patterns. MEDIUM confidence. Validate per client.

• Workforce %: Benchmarked from NASSCOM Insurance GCC 2025, ANSR Insurance captive structure data, Swiss Re / Munich Re / AIG / MetLife / Prudential / Aviva published GCC headcount mixes. Directional only.

• Upskill paths: Based on KPMG workforce transformation frameworks. Target roles, certs, ramp times are indicative.