TL;DR
Principal Product Designer at SalesforceIQ, 2015-12 to 2017-04. Redesigned the Contact Gallery so users could understand, trust, and act on AI-captured relationship data without re-learning the app.
- 40% increase in daily active users after the Contact Gallery redesign
- 34% increase in duplicates merged after introducing multi-merge and a master contact model
Context and mandate
SalesforceIQ had real technical ambition. Relationship intelligence, automatic data capture, the early shape of what would become Einstein. The back-end was ahead of its time. The front-end was making customers do more work than the technology should have required. Users could not find what they needed, did not trust the data, and had no clean path to resolve duplicates. Three long-standing customer issues had been open for months.
The real problem was not the AI underneath. It was that the product was not translating that AI into anything a sales user could act on without re-learning the app.
I worked weekly with support, product, marketing, and engineering so prioritization stayed honest about customer sentiment, then took the contact intelligence and data quality surface directly. The bar was measurable behavior change.
Key decisions
Close known pain before reframing. I shipped fixes to three long-standing customer issues in the first month instead of opening with a vision exercise. The credibility that bought is what earned license for the broader redesign.
Design Contact Gallery for action, not information. I could have layered richer data onto an already dense surface. Instead I designed for what users needed to do: understand the contact, trust the data, resolve ambiguity in one session.
Treat multi-merge as a product model decision, not a feature. Customers knew they had duplicates and had no usable path at scale. I introduced multi-merge and a master contact model, which meant rethinking the contact entity, not adding a button. That decision is what produced the 34% lift in duplicates merged.
Multi-merge as a product model
Three frames of the decision the +34% lift sits on.
- 01 / 0301 · Pick master
The first step exposed the model decision in the surface. Three contacts the system thinks are the same person, three sets of facts, one explicit Master selection. The product was not auto-merging on the user's behalf. - 02 / 0302 · Select photo
A separate explicit step for the default photo. The merge is composed from atomic decisions the user can reason about, not one large irreversible action. - 03 / 0303 · Edit details
The composed contact is editable before save. All three names, four emails, three phone numbers, two addresses, two handles are visible at once. The user owns the final shape, not the system.
Outcomes
Daily active users
After Contact Gallery redesign, 2017.
Plus a 34% lift in duplicates merged via the multi-merge model.
- Daily active users increased 40% after the Contact Gallery redesign.
- Duplicates merged increased 34% after multi-merge and master contact model shipped.
- Three long-standing, high-frequency customer issues closed inside the first month.
- Early Einstein thinking on confidence, source transparency, and correction loops fed forward into the company's broader AI direction.
Reflection
AI products fail when the surface in front of them keeps treating the user like a database administrator. The job is to put the intelligence inside the workflow the user already has, so the system feels smarter without asking the human to do more. Designing for what a user needs to do with machine-generated data, not just how to display it, has been a recurring move in every AI surface since.