Scaler's in-house CRM · Lead Details redesign

Redesigning Lead Details inside Scaler's ₹5 Cr CRM rebuild

First, a little context

About Scaler

Scaler is a technology education platform that helps working professionals build industry-relevant skills through structured programs, mentorship and career support.

Overview

Scaler replaced its external CRM with an in-house platform, then redesigned Lead Details so sales reps could understand a lead's intent and context before every call.

My role

I was a PD1, working alongside Kishan. I independently owned the CRM Lead Details redesign.

Outcomes

₹5 Cr/year

Vendor cost removed by the broader CRM initiative

7.5 → 8.2

CRM-wide CSAT from approximately 200 BDAs

100%

BDAs migrated by March 2025

Scaler needed more control over its sales CRM

The CRM sat at the centre of Scaler's sales workflow. Teams used it to understand prospects, prepare for conversations, record outcomes and move prospective learners towards enrolment.

Business Development Associates

Understand leads, make calls and update outcomes.

Sales Managers

Review pipelines, support reps and guide follow-ups.

By mid-2024, Scaler needed new capabilities from its CRM. The external vendor quoted ₹1 Cr for the upgrade, on top of the ₹5 Cr annual subscription. An in-house build was estimated at ₹80L to ₹1 Cr, making ownership the more sensible long-term choice.

₹6 CrPotential total

External CRM

₹1 CrEstimated build

In-house CRM

Hover or focus on either bar to see the cost context.

I was one of nine people who rebuilt the CRM. The broader initiative replaced the recurring ₹5 Cr vendor contract with an internally owned platform.

Moving from an external CRM to an in-house platform

Sales reps depended on the external CRM every day. We first recreated its core workflows so they could migrate without relearning their jobs.

The lead designer and I split the platform modules between us. We kept the workflows reps already knew. We then improved filters, lead-stage colours, automations and dark mode.

Although the migration succeeded,
the Lead Details page still slowed BDAs down

The existing page treated every call, form submission and system event equally. To understand a lead, a BDA had to scroll, open individual activities, read the details and mentally combine the signals.

Fun fact: I built this using Claude
Reconstruction of the original Lead Details page: finding one note meant scrolling the activity log, loading more, opening the call and reading to the bottom.

Four issues created this problem

Signal was buried in noise

Routine system events and high-value intent signals carried the same visual weight.

Notes were hard to recover

Reps had to reopen old call activities to understand what happened before.

AI insight was hard to scan

Background, objections and next steps appeared in dense blocks of text.

Actions were fragmented

Contact details and frequent actions were spread across the page.

What we heard from sales leaders and reps

During discovery, I interviewed approximately 15 AVPs and sales managers and spoke directly with around 10 BDAs. Sales leaders helped me map the workflows they observed, while the people doing the calling described the same workflow in their own words. Their concern was not missing data. It was the effort required to interpret it.

Below are the synthesised notes from those discussions. They represent recurring themes rather than verbatim quotes.

BDA notes

BDA note

Like the old CRM, this page has too many buttons. Can we simplify the actions we use most?

BDA note

We work fast. The right information needs to be easier to find.

BDA note

We use AI insights before calls, but the current blocks take too long to read and interpret.

Sales Leaders Notes

Sales Leader Note

Reps need clear contact details, quick context and actionable prompts in one place.

Sales Leader Note

Most of my reps use the CRM only to fetch contact details, not to understand the lead.

Sales Leader Note

Strong intent signals are present, but they are buried among routine activity.

“Create a single-screen interface tailored specifically for sales execution. Remove non-sales distractions and place useful context exactly where reps need it.”

Constraints I had to work with

01

Keep the existing taxonomy

Sales had just migrated and knew the existing terminology. The Sales Head had already defined the Sales, Marketing and Product categories. I kept this taxonomy and made intent easier to scan.

02

Move to production in one month

I was the sole designer and had one month to ship the redesign. I focused on the highest-friction workflow and reused the existing design system.

Redesigning the Lead Details page

Based on the interviews with sales leaders and BDAs, I followed these three design principles:

01

Prioritise action over information.

02

Show only what sales reps need.

03

Keep contact details, lead context and next steps within easy reach.

The redesigned Lead Details page, running live.

Exploring multiple iterations
before arriving at the final solution

Combining activity, AI and notes

The first version placed activity summaries in the centre and combined AI insights with notes on the right. It failed because the three-column layout became cramped on smaller BDA laptops and pushed important signals below the fold.

Breaking activity down by type

The next version used totals, rates and the latest event for each activity type. It still failed because every added activity increased scrolling, while routine events and stronger buying signals looked equally important.

Breaking down
the final design piece by piece

1. Activity Groups made lead intent easier to scan

  • Problem identified

    The chronological timeline gave routine system events and meaningful buying signals the same visual weight.

  • What I did

    I kept the latest five activities on top for recency. I then grouped the full history into Sales, Marketing and Product.

  • How the grouping was decided

    The three buckets came from the Sales Head, so the taxonomy was fixed before I started. My work was to make that structure carry intent rather than only sort events.

Recorded in production. The latest five activities stay on top, then Sales, Marketing and Product each carry their own group.

2. Hot Activities gave reps a reason to pick one lead first

  • Why I proposed it

    Nobody asked for this. Grouping told a BDA what a lead had done. It still did not tell them which lead in their set to call first.

  • What I did

    I worked with the sales and data teams to find the behaviours that correlated with higher lead-to-payment conversion. Only those behaviours are promoted to a Hot Activity and given a badge. Passing the entrance test, requesting a call back and downloading the placement report all qualified.

  • Why a badge

    The badge does two jobs. It helps a rep prioritise one lead over the rest of the set. It also gives the rep a visible win for the work that moved the lead.

  • Where it landed

    This was my own addition to the brief. Sales leaders backed it and asked for it in the build, so it shipped with the redesign.

Each Hot Activity carries a badge and a timestamp, so a rep can see the strongest signal without reading the full history.

3. Notes brought earlier conversations back into view

  • Problem identified

    Manually written notes were hidden inside individual call activities and were tedious to recover during a handoff.

  • What I did

    I moved notes into a persistent rail and styled them like the sticky notes BDAs already used, keeping earlier conversation context visible during the next call.

  • What it cost before

    Recovering a single note meant opening the activity, reading it and closing it again, roughly three interactions each time. In the redesign that context is already on screen, so the same information costs nothing to reach.

Notes sit in a persistent rail and keep their author and timestamp, so context survives a handoff.

4. AI Overview became a pre-call brief

  • Problem identified

    The existing AI summary presented call history as a dense block of text that was difficult to scan before a conversation.

  • What I did

    I turned it into a pre-call brief with four scannable elements: lead background and intent, transcript-derived objections, a call score and recommended next steps.

The AI Overview splits into a lead summary and transcript-derived pain points, each tagged by objection type.

5. Primary actions became easier to reach

  • Problem identified

    Contact details and frequent actions were spread across the page, adding friction before and after a call.

  • What I did

    I brought Call Lead, email, WhatsApp and process actions into the top-level hierarchy so BDAs could act without searching through the page.

Call lead, email, WhatsApp and the process menu all sit in the top-level hierarchy.

Going above and beyond
the brief and problems mentioned by stakeholders

6. A readable checkpoint before calling

BDAs were copying small phone numbers from the CRM, pasting them into WhatsApp and then using their phones to call.

Call Lead opened a drawer with the phone number and email in a larger format. It also surfaced AI context, giving the rep one final checkpoint before the call.

Call lead opens a drawer with readable contact details, the AI summary and the lead's pain points before the call starts.

7. Clear feedback after submitting a process form

After submitting a process form, BDAs refreshed the timeline to confirm that the system had registered their work.

I added a six-second confirmation showing the automations completed by the system. A production-ready Wolf of Wall Street moment acknowledged the rep and added levity to a repetitive workflow.

The confirmation lists what the system did, then closes itself after six seconds.

Impact after release

₹5 Cr

Annual vendor cost eliminated by the broader CRM initiative

7.5 8.2

CSAT across approximately 200 BDAs. Lead Details was the only major release in this window.

100%

Every BDA moved to the redesigned page by March 2025

Reviews shared by the team

Review image placeholder
Review image placeholder
Review image placeholder
Review image placeholder

What I would do
differently next time

  • Instrument the page before shipping it. We shipped this page without instrumenting it, and I treated measurement as the PM's responsibility. I would not make that trade again.
  • Agree the success measure before design starts. For this page it would most likely be the time a rep needs to reach a meaningful signal, and I would ask for that tracking inside the build scope, not after release.
  • Hold every surface to the standard I set for one. The AI calling work shipped with feedback capture attached, so the people relying on its output could judge the quality instead of assuming it. The rest of the page deserved the same loop.

Next Project

Scaler · Design Engineer · AI Build

How I designed and shipped an n8n learning simulator in 2 weeks

A guided simulator that records every learner decision, scores it against fixed rules and explains the result.

View full case study