Magnit Pulse
AI
UX DESIGN
Making the VMS disappear.
Enterprise hiring managers spend hours a week approving timesheets, extending contracts, and reviewing candidates, inside systems they weren't designed to use. Magnit Pulse is the intelligence and action layer that moves all of it into Microsoft Teams and Slack. The VMS stays as the system of record. Managers never have to open it.
"If I could do everything in Teams, game changer." Hiring manager, Fortune 500 telecom
Role
Senior Director of Design, acting as strategic design lead. I owned the end-to-end design vision for Pulse from the core UX principle through prototyping the MVP and future state, working alongside product and engineering leadership from concept through live product and beyond.
Scope Product strategy · Research · UX architecture · Interaction design · Prototyping · Design system · Future state vision
Platform Microsoft Teams (live) · Slack (in development) · AI agent layer (designed)
Nobody had drawn the whole thing.
Before this project, research at Magnit was siloed. Each product manager handled it their own way for the products they owned. We had deep domain expertise sitting in people's heads and almost nothing shared, which meant every team understood its own slice and nobody could see across them.
I built a centralized research practice to fix that: personas, interviews, and evidence consolidated into one place that product and engineering could work from.
The first thing I did with it was map the entire contingent workforce lifecycle across every persona we serve. Client sponsor, supplier, client manager, MSP, and worker. Requisition through offboarding. Every handoff between lanes.
Seen from inside any one product, each of those steps looks reasonable. Seen all at once, a pattern appears that no individual product team was positioned to notice: the client manager touches this system constantly, at almost every stage, and is the only persona who has no dedicated place to work. Suppliers have a mobile app. Workers have a mobile app. MSPs live in the platform all day and know it well. The hiring manager, who makes most of the actual workforce decisions, logs into a system built for procurement specialists, does one task, and leaves.
That's the opportunity Pulse was built on, and it only became visible once the research stopped being organized by product.
Act one: Maggi.
Our first AI capability inside the VMS was Maggi, an assistant that let users ask questions and take action in natural language rather than navigating to the right screen.
Maggi worked. It proved that conversational interaction fit this domain, that people would trust it for routine tasks, and that the underlying data could support it.
It also revealed the limit. Maggi was still inside the VMS, which meant a hiring manager had to log in to use the thing designed to save them from logging in.
The insight wasn't that we needed a better assistant. It was that the container was wrong.
Don't port. Redefine.
The easy version of this project was taking the VMS approval screen and wrapping it in a Teams tab. That's porting, and it would have failed. It would have carried all the friction of the original interface into a context where people expect things to feel native.
The harder, and right, question was: if this workflow didn't exist in the VMS at all, how would a hiring manager want it to work inside Teams?
That reframe changed everything. Not just the visual design, but the information model, the interaction patterns, and the scope of what belongs in the app versus what should hand off elsewhere. It meant starting from a manager's actual day, not from a feature list.
A manager's day in two panels.
The home screen is the core design decision of the MVP. Everything a manager needs to get through their day, and nothing they don't.
Two sections: Action Items and Upcoming Events. Action Items surfaces what needs a decision today: timecard approvals, expense adjustments, headcount approvals, request financials. Upcoming Events gives a rolling seven-day view of what's coming, including interviews, engagement starts and ends, and contract expirations.
What I deliberately left out is as important as what's there. No dashboard metrics. No navigation to the full VMS. No feature discovery. The home screen has one job: help a manager prioritize and act.
776,000 timecards in 100 days.
That's the approval volume we were designing for across the client base. At that scale, every point of friction compounds. A manager approving 200 timecards a month who has to log into the VMS for each one isn't just inconvenienced, they're delayed, and so is every worker waiting to get paid.
The design question wasn't "how do we show a timecard in Teams." It was: what does a manager actually need to see to approve with confidence? That answer is much smaller than a full VMS timecard view, and getting to it required pulling apart the information hierarchy and rebuilding it around the decision rather than the data model.
The result: a card that surfaces the signal, a detail view for the cases that need it, and a one-tap approval path for everything in between.
Hiring someone shouldn't start with a 30-field form.
Request creation is the highest-stakes workflow in the VMS. It's also one of the most common reasons managers abandon the system and call their MSP rep instead. The form is long, the terminology is internal, and nothing guides you through it.
The design challenge was real. We couldn't remove the data, because compliance, rate validation, and sourcing routing all depend on it. But we could change the model entirely. Instead of presenting a form, Pulse walks managers through a short guided conversation.
A few contextual questions: what type of worker, backfill or net new, role, location. Then the system does the work, routing through Sourcematch, surfacing rate intelligence from Pay Intel, and generating a summary the manager can review and confirm.
The principle: reduce cognitive load without reducing data quality. In practice, that meant opinionated defaults, surfacing intelligence at the moment of decision, and making every question feel like it's already working on an answer.
The best interface for routine work is no interface.
Most approval tasks are predictable. A timecard comes in, it's within tolerance, it gets approved, every time. A manager doing this 200 times a month isn't making 200 decisions. They're doing 200 confirmations, and the cognitive overhead is real even when the action is automatic.
The next phase of Pulse takes this seriously. The intelligent approval agent lets a manager define their own policy: approve automatically when hours are within range, flag anything that exceeds a threshold or hits an anomaly.
From that point, the manager only sees what actually needs their attention. A variance, an unusual rate, a worker whose hours jumped unexpectedly.
This isn't automation for its own sake. It's the logical end of the design principle: surface only what requires a human decision, and get everything else out of the way.
This feature is designed and prototyped, currently in development.
The bigger picture.
Pulse started as a Teams app. But the design foundation was always bigger than one platform.
The architecture underneath, Magnit's open API platform, means the same intelligence that powers the Teams experience can run anywhere. Slack, AI agents embedded in SAP or Workday, or purpose-built internal tools.
As design owner, my job wasn't just to design screens for Teams. It was to establish what the right experience model looks like when workforce management lives inside the tools people already use, whatever those tools turn out to be. That model scales. The screens are the first expression of it.
Where it stands.
Magnit Pulse launched publicly in June 2026. It's live in Microsoft Teams and available via API, with Slack in development and the intelligent approval agent prototyped and in build. Pilot clients are onboarding now.
The thesis held up outside the design team. From the launch announcement:
"Most VMS platforms are primarily designed for procurement specialists, creating friction for hiring managers who make most workforce decisions. Magnit Pulse shifts the paradigm from complex software navigation to seamless user experiences." Chandra Dhandapani, Chief Executive Officer, Magnit Global
That's the argument the journey map produced, restated as company strategy.