Conversation beside the account

The challenge

Trustbridge was meant to help Singtel account managers sell. The people it was for already had the harder job: portfolios in the thousands, no reliable way to see who was worth a call today, and customer context split across Salesforce, billing, and whatever sat in a spreadsheet. Outreach was inconsistent. Care tickets ate the hours that should have gone to sales.

The other product problem was the project. When I joined, there were decks, workshops, data work, vendor conversations, and an incomplete product bible, but not one MVP. Service design was not consistently feeding delivery. Security and data were moving in parallel. Business owners had been through the room more than once without seeing something they could use.

I was brought in as a product designer and, within the first two weeks, asked to operate as product owner: own the thing, put it in one place, make workshops produce requirements, and sequence what could actually be built.

What we did

I led the product definition and, after a later pivot, the UX architecture and UI direction, working closely with Ishwar and the wider product, data, and engineering group. Temus built from that handoff. I did not invent Trustbridge. The original concept and early screens were already there.

• Took product ownership of a fragmented AI initiative
• Narrowed the MVP to a lead engine, not a replacement CRM
• Mapped the UX before anyone redrew the screens
• Designed the two-panel pivot and handed it to Temus for build

A lead engine, not another CRM

The work kept trying to become a platform. More fields, more automation, a new system of record. Salesforce was already the CRM. Trustbridge needed a boundary or it would never ship.

The shape we held to was a loop: clean enough account context, a way to see who to focus on and why, focused outreach, then a response that writes back into the account instead of dying in an inbox. Rules first, because the data was not ready to pretend at machine learning. Humans in the loop wherever confidence was weak.

If a lead was not qualified enough to save into Salesforce, it stayed in Trustbridge. That sentence did more for the MVP than another workshop.

UX before UI

The inherited project had screens. What it lacked was a flow. We aligned on requirements, then a UX map, then UI, because drawing another dashboard would have frozen the wrong product.

The map had to answer the ugly questions in public: what is safe to automate, what happens to a bounce or a care reply, whether top leads refresh or sit until acted on, whether Trustbridge writes to Salesforce or asks first. Those decisions are the product. The screens come after.

Keep the account in view

Earlier UI work, from other designers, did not survive a later pivot. I took the product definition from the design side and rebuilt the architecture around a split the account manager can actually use: an AI conversation on the left, the account, signals, and actions on the right, visible at the same time.

The point of the assistant is not a chat that hides the work. Priority, why-now signals, source, and the next action have to stay in view while the AM talks. Reply, reach out, fix a contact. That two-panel model is what went to Temus, and what they built.

What it moved

This is in UAT, not production, and there is no conversion number I will put on this page. What changed is that a fragmented initiative became a product with a boundary, a flow, and an interface a vendor could build. The direction that is in testing is the pivoted one, conversation beside the account, not the screens that came before it.