Helping pharmaceutical field teams move from territory-level performance data to focused HCP actions — one workflow for insight, evidence and next-best action.
Field representatives manage large territories containing hundreds of healthcare professionals. To prepare effectively, they need to understand both the overall territory and the individual HCPs influencing its performance.
The Field Territory Insights Hub brought together AI-generated territory insights, market and prescription trends, patient-switching behaviour, HCP performance rankings, search and HCP-level intelligence, weekly engagement recommendations and conversational data exploration.
The product connected territory analysis with individual HCP actions — moving users from “what is changing?” to “who should I engage, and why?”
Field teams had access to extensive market, prescription and customer data, but translating that information into clear priorities was hard. The experience needed to help users:
…help field teams quickly interpret territory performance, identify priority HCPs and convert insights into confident engagement actions?
Two core needs framed the product:
The experience was then organised around a progressive journey — Territory overview → Key change → Supporting evidence → Priority HCP → Recommended action.
Open business questions on brand terminology, therapy-area comparisons, HCP naming conventions and future scalability were documented alongside the designs and considered during iteration.
The project file documented requirements and stakeholder questions rather than formal interview findings. The strongest UX insights came from analysing the field representative's daily workflow.
Reps manage many HCPs, so they need a territory-level view before investigating individual profiles.
Charts alone show change; AI summaries were used to explain the change, name the contributing HCPs and point to the next investigation.
Users needed to see why an insight or recommendation was generated — rationale, sources and dates sit alongside AI outputs.
Recommendation cards borrow the structure of the field team's existing CRM workflow, keeping familiarity while adding AI capability.
A suggestion may be wrong, mistimed or already done — users can explain, dismiss, snooze, complete, retrieve or ask for an alternative.
Narrative insights, charts and a wide leaderboard could not compete at once — progressive disclosure and overlays kept the surface calm.
The working file explored structure and components rather than a separate low-fi set. The wireframes below reconstruct the intent — four flows and four skeleton layouts for the surfaces that ended up in the UI.
Begin with a summary, then progressively access underlying detail.
Bypass the territory dashboard when the user already knows the HCP.
Dismissal includes reasons and snoozing; completed or dismissed items can be retrieved.
Complements the structured dashboard rather than replacing it.
Four hi-fi surfaces cover the territory dashboard, its scrolled analytical view, the integrated conversational assistant, and the weekly Suggestions list. All copy, HCPs and territory names are fictional; only the drug-class terminology is retained for realism.
The dashboard opens with concise territory insights covering class movement and share changes — an interpretation is offered before the detailed chart.
Selecting View Evidence opens a contextual overlay with the supporting chart and source dates instead of navigating to a separate report.
HCP references in insights, tooltips and leaderboard rows are interactive — a click opens the individual profile from anywhere in the flow.
Market correlation, patient-switching, share-vs-volume and HCP ranking each get their own module; tooltips reveal deeper metrics only on demand.
Rank and HCP name stay anchored while metric columns scroll horizontally — filters narrow the visible group, comparison never collapses columns.
Every card is stateful — explain, feedback, complete, dismiss with reason, snooze, retrieve, previous suggestion, generate an alternative.
A prompt entry point lives inside the territory experience with suggested starters and a usage disclaimer on generated answers.
Insight cards, KPI tiles, filters, tables, tooltips, overlays, recommendation states and feedback controls were built as one library to scale across brands and therapy areas.
The working file did not contain verified production metrics — the impact below is presented as product outcomes rather than business results.
Narrative insights established direction; charts and tables provided progressive levels of evidence.
An insight becomes credible when users can inspect its rationale, sources and underlying chart — not just a conclusion.
Dismiss, snooze, retrieve and previous-suggestion states gave users room to decide without fearing permanent consequences.
Expanded, completed, dismissed and alternate states aren't visual variations — each is a stage in the user's decision.
Vertical dashboard scrolling combined with a horizontally scrollable leaderboard preserved readability without hiding data.
Aligning recommendations with a known CRM pattern reduced the learning curve; AI explanations and alternatives added the new value.
Questions about brand labels, therapy comparisons and HCP naming exposed the importance of components that support future products, not just the first use case.