Tata Consultancy Services

A wealth analytics platform that lost advisors in its own information architecture

TCS built a B2B wealth analytics platform for a large Indian banking group — the daily working surface for ~800 relationship managers reviewing portfolios for retail and HNW clients. The platform had everything an advisor needed: returns analysis, tax forecasting, asset allocation planning, risk assessment. The data was right. The architecture was wrong.

I joined as Senior Product Designer to lead the redesign of investment dashboards and portfolio analytics workflows. The brief on day one was a number: advisors were averaging 25 minutes per portfolio review against a 12-minute SLA. That's not a UX problem the business could ignore — that's the bank's wealth advisory cost structure baked into a slow product.

Advisors weren't slow at thinking. They were slow at finding.

The fragmentation hypothesis emerged on day one of workflow interviews. Advisors weren't spending 25 minutes analysing portfolios — they were spending ~9 minutes per review hopping between four disconnected modules trying to reconcile numbers that should have already agreed. The analysis was fine. The architecture was forcing manual reconciliation that the system itself should have handled.

Four evidence streams. One unified diagnosis.

I anchored the research on four independent sources of evidence. None of them was sufficient on its own, but together they painted a clear picture: the analytical layer was strong; the navigational layer was costing the business an entire SLA's worth of advisor time per portfolio.

Competitive benchmarking — the gap was reachable

I audited three comparable B2B wealth analytics tools (Refinitiv Workspace, Morningstar Direct, Bloomberg PORT). Average review time in published case studies and user testimonials: 12–15 minutes. The gap between this platform and the category leaders was 10–13 minutes per review — almost entirely on architecture, not features. That established a realistic ceiling for the redesign and a defendable target metric for the business case.

Three architectures evaluated. Here's why we chose the one we shipped.

We optimised for the 82% who were slowing the business down — and accepted short-term pain from the top 18%.

The unified canvas was the right architectural call, but it had a real political cost: top-tier RMs (the 18% who generated 40% of revenue) had memorised the old module structure and used it efficiently. The redesign broke their muscle memory. For two weeks post-launch, they were the loudest complaint channel in the building.

How we de-risked the senior-RM transition

I built a parallel "legacy module view" toggle for the first 60 days post-launch — senior RMs could opt back into the old structure while learning the canvas at their own pace. Usage data after 60 days: 71% of senior RMs had voluntarily switched to the new canvas. The "legacy view" was retired in release 3, with no further escalations. The lesson: when you change a power user's workflow, give them a bridge, not a cliff.

Every architectural decision pointed back to one metric: minutes per review

Unified canvas — the hierarchy that replaced four modules

Reconciliation flags — the unsung hero feature

The single most-loved feature in user feedback after launch was the reconciliation flag — an inline warning when a portfolio's reported allocation didn't match the underlying holdings (typically due to corporate actions the data feed hadn't fully reflected). Before redesign, advisors caught these manually maybe 60% of the time. With the flag system: 100% catch rate, and the advisor doesn't have to think about it. This is a case of design quietly eliminating an entire failure mode the old architecture had been silently passing on.

A/B testing the dashboard improvements

I owned the experimentation strategy — running staged A/B tests on the new canvas vs. the old modules across the three regional clusters. Engagement on core dashboard features (returns drill-down, allocation rebalance, tax scenario modelling) lifted +31% in the test cohort. The "summary tab" feature requested by some stakeholders during scoping was killed by the data: in the cohort where we shipped it as a control, usage was 84% lower than the canvas itself. Evidence killed the band-aid.

Reusable component patterns — the engineering compounding

I architected reusable component patterns for financial workflows: portfolio cards, drill-down hierarchies, reconciliation states, tax-scenario inputs. These now sit in a documented Figma library with code-mapped tokens. The measurable effect: engineering clarification cycles per sprint dropped from an average of 8 to 3. Sprint velocity on adjacent features (risk module, client-facing reports) accelerated because designers and engineers were no longer reinventing patterns each sprint.

25 min → 14 min. Multiplied across 50,000 portfolios. The math is the story.

Measured outcomes

Per-portfolio review time: 25 min → 14 min (44% reduction). Measured via the same timed-task protocol as pre-redesign, n=30 reviews across the same three regional clusters. The result held steady across all three clusters, suggesting the improvement was structural, not site-specific.

Core dashboard engagement: +31% via A/B test across the three clusters. The summary-tab control variant saw 84% lower usage than the canvas — killing a stakeholder-requested feature before it shipped to the wider population.

Engineering clarification cycles: 8 → 3 per sprint on average after the documented component library shipped. Design QA rejection rate on sprint releases dropped from ~22% to ~7%.

Business framing — the leverage at scale

The per-portfolio time saving doesn't sound dramatic in isolation: 11 minutes per review. Multiply across the user base:

Time saved = (25 - 14) min × 50,000 portfolios/month = 550,000 min/month = ~9,167 advisor hours/month reclaimed

At a blended Indian RM cost of ~₹1,500/hour (publicly cited industry benchmark for B2B wealth advisory talent in 2023), that's ~₹1.37 Cr per month in advisor productivity recovered. Methodology: 9,167 hours × ₹1,500. The dollar figure here isn't precision — it's framing. The redesign translated an architectural fix into a business case the executive team could defend to their P&L.

AI-drafted portfolio commentary — targeting the bottom-quartile RMs

The 14-minute review time is strong, but the cohort distribution is informative. Top-quartile RMs are reviewing in 10–11 minutes; bottom-quartile RMs are at 16–18. The gap is now almost entirely in writing client-ready commentary — the qualitative narrative that wraps the numbers for the client meeting. Senior RMs write it fast; junior RMs labour over it.

Hypothesis: an AI-drafted commentary panel (LLM-generated first draft, advisor edits and approves) would compress the bottom-quartile time by another 4–5 minutes — collapsing the cohort gap and pulling everyone closer to the SLA. Instrumentation: commentary edit tracking is already in place; we know what advisors keep, change, and discard from any LLM draft. Risk: AI commentary in regulated wealth advisory needs careful guardrails — an issue I'd separately addressed in detail at Goldman Sachs.

A B2B client. A live user base. A sprint cadence. No room for redesign theatre.

TCS engagements run on hard, observable constraints: a client-set release cadence, a live user base that can't tolerate downtime, and a sprint-based delivery model where designs ship within the window or they don't ship at all. The redesign had to be defensible to advisors, the bank's product team, and TCS engineering leadership simultaneously.

The triangle: I picked Quality + Scope. Time was the lever I managed.

Cutting scope was off the table — an incomplete unified canvas would have created a worse architecture, not a better one. Cutting quality risked the bank's confidence in TCS as a delivery partner. Time was the lever: I shipped the canvas across two sprint cycles with a parallel "legacy module" toggle to absorb the senior-RM transition risk. Quality and scope held. Time stretched. That was the right call.

Three lessons from a B2B wealth analytics engagement