Case Study · B2B Agent Workspace · ICICI Lombard

Rebuilding fragmented agent systems into one AI-powered IL NYSA platform.

IL agents had to jump across disconnected app and web systems to finish core work. We reframed that ecosystem as one platform for selling, managing, and growing insurance business, with AI assistance embedded into daily workflows and a clear financial case for consolidation.

Vision slides aligned with stakeholders

160 Cr

Estimated cross-sell opportunity

29%

Estimated cost reduction potential

Phase-1 handoff under active build

Role: Principal Design Lead
Team: Design + Product + Engineering + Business
Domain: Agent-facing Insurance Platform
Platform: Mobile + Web + Assistant Layer

01The business problem

The ecosystem worked, but only in fragments.

Multiple platforms supported IL agent work, but none worked end-to-end on its own. Agents had to switch tools to complete basic tasks, while teams rebuilt similar features in silos. The challenge was not inventing more features; it was consolidating the system with better UX and lower operational redundancy.

IN SCOPE

Unified journeys for agent work across mobile and web touchpoints
Core booking, renewal, and service workflows in one product language
AI-assisted prompts for growth, follow-up, and next-best action support

OUT OF SCOPE

No fragmented feature ownership repeated independently by channel teams
No assistant-led irreversible actions without explicit user control
No phase-1 attempt to solve every downstream legacy dependency at once

02The quantified case

Consolidation was a growth and cost decision, not just a UX cleanup.

Stakeholder decks established a direct commercial rationale: cross-sell and bundling upside, plus measurable reduction in redundant platform spend. That shifted IL NYSA from design proposal to strategic business program.

User Pain Reality

Multiple logins, no single business overview, inconsistent app/web UX, and poor resource discoverability were repeatedly surfaced in stakeholder interviews.

Growth Upside

Cross-sell potential was estimated at around 160 Cr, with bundling opportunities highlighted across motor and health portfolios.

Operating Cost Burden

Current fragmented platform development cost baseline was presented as 7,82,65,738, with duplicate team-level maintenance overhead.

Savings Hypothesis

Business estimate projected 29% cost reduction with IL NYSA, including an indicative reduction of 3,63,15,302 from the modeled cost stack.

03The platform strategy

One platform, role-aware journeys, shared capability backbone.

We unified booking, service, and growth operations across mobile and web. AI is positioned as an assistive layer inside governed workflows, not an isolated chatbot. That keeps execution accountable while making the system proactive and context-aware.

Agent Context

Role, tenure, pipeline, and customer workload are identified first.

Unified Task Layer

Booking, renewal, service, and insight tasks are orchestrated in one flow.

Insight + Suggestion

Assistant proposes actions, communication drafts, and prioritised nudges.

Control Gate

Business rules, permissions, and compliance checks validate every step.

Execution + Tracking

Actions complete with auditability and feed growth/performance loops.

04Design governance

Consolidation had to improve control, not just convenience.

A single platform can become a single point of failure if boundaries are vague. We built governance into product behavior: role visibility, permission logic, phase sequencing, and traceable decision paths became core UX quality criteria.

Role Governance

Parent, PoSP, and agent views are permissioned so the same platform still respects responsibility boundaries.

Operational Control

Every assistant recommendation stays reviewable and user-driven before final execution in critical workflows.

Platform Consistency

Shared patterns reduce duplicated feature implementation across disconnected systems and teams.

Phased Rollout

Vision is stable, while flows move through handoff and staged build to avoid broad-risk big-bang launches.

05Vision to delivery

Stakeholder truth is fixed; delivery flows are staged.

The IL Nysa Project deck established the final vision with business stakeholders. The concept proto translated it into end-to-end interaction hypotheses. The current Dev Handoff confirms phased implementation: homepage variants, UAM happy flows, select city + LOB patterns, web variants, and IMID login are now in active development.

06Outcome framework

Success is measured at business, experience, and system levels.

IL NYSA should reduce tool-switching effort, improve conversion and retention support quality, and lower duplicated platform overhead. The north star is clear: help agents act faster, track growth clearly, and make better decisions with AI-backed guidance across channels.

Next step

From pilot assistant to scaled AI service.

Next increments focus on closing remaining fragmented journeys, hardening role-based operations, and improving assistant quality from recommendation to reliable execution support.

See more work Start a conversation