Media Watcher - a full revamp of the case investigation experience
Media Watcher is an adverse-media and entity-monitoring product for compliance and analyst teams. I introduced, led, and designed a whole new line of features and a ground-up revamp of the single-case experience - from a thorough product audit through to the redesigned case page - working closely with product and engineering.
- Role
- Product Designer · Lead
- Scope
- Audit · New feature line · Case-page revamp
- Tools
- Figma
- Industry
- Media monitoring / RegTech
- Platform
- Enterprise web
- Users
- Compliance analysts
- Team
- Product · Engineering
- Type
- Enterprise SaaS
01 Overview
Media Watcher tracks adverse media and entities at scale for compliance and investigation teams. The single-case experience had grown powerful but uneven - its strongest asset was buried, core investigation tools were missing or shallow, and analysts couldn't fully trust or act on what they saw. I led a full revamp: a thorough product audit, a new line of features, and a redesigned case page, all built with product and engineering.
02 The challenge
A case could hold hundreds of mentions, but the tools to investigate them hadn't kept pace.
- Sentiment was automatic-only and couldn't be corrected, so case metrics weren't fully trusted.
- Duplicate, syndicated stories repeated in the feed and inflated the counts.
- Search and filtering were too shallow to pinpoint anything in a large case.
- Monitoring was a single on/off toggle - no configurable or spike alerts.
- No case status, owner, or audit trail, so it couldn't be the system of record for an investigation.
- No collaboration, tagging, or noise controls to triage a big mention set.
- The product's signature entity analysis was hidden behind a "View Details" tab.
03 My role
I introduced, led, and designed the new feature line end to end, and led the redesign of the whole case experience - partnering closely with product and engineering throughout.
- Ran a thorough product audit and built a prioritized improvement backlog.
- Defined and designed a new line of features across investigation, workflow, risk, and intelligence.
- Redesigned the single-case page into a clear, structured flow.
- Scoped an MVP and an advanced version for every idea so engineering could ship in stages.
04 Research & audit
I worked from evidence outward - auditing the existing case experience end to end and turning every gap into a scored, prioritized backlog.
- Logged every gap against the real single-case workflow.
- Scored each improvement for impact, effort, and confidence on a 1-5 scale.
- Prioritized into Critical, High, Useful, and Nice-to-have tiers.
- Pulled out high-impact, low-effort quick wins to ship first.
05 The new feature line
The new features fall into six areas, each designed with an MVP and a more advanced version.
Trustworthy data
Editable sentiment with logging, duplicate and syndication clustering, source-credibility scoring, and noise controls - remove a mention, mute a source.
Investigation power
Full-text and boolean search with saved filters, custom tags with bulk actions, and an AI case assistant that answers questions grounded in the case's entities.
Risk & triage
A case-level risk score with explainable drivers, and period-over-period trend deltas so analysts can read momentum at a glance.
Workflow & trust
Case status, an owner and an audit trail; internal comments and notes; and role-based access for sensitive cases.
Reporting & sharing
Multi-format export and a curated report builder, read-only executive dashboards, and scheduled digests.
Intelligence
A worldwide real-time trends engine, AI predictive trend forecasting, and an entity relationship graph that reveals hidden connections.
06 Redesigned case page
I restructured the single-case page into a clear top-to-bottom flow, so an analyst can answer "how bad is this, and what do I do?" in seconds - and so the product's hidden differentiator sits up front.
- Case header - status, owner, and primary actions, always in reach.
- Risk banner - a composite risk score with its drivers, for instant triage.
- AI summary - what's happening and what changed since the last view.
- Key findings & entities - top adverse entities and keywords surfaced first, no hidden tabs.
- Evidence feed - editable sentiment, tags, bulk actions, and collapsed duplicates.
- Analytics - trend, sentiment, geography, and theme clusters as support.
- Run & prove it - comments, assignment, decisions, and an audit trail.
07 Before & after
The revamp moved the case experience from a reactive viewer to an investigation system of record. Drag the handle to compare the old case page with the redesign.
- Sentiment was automatic-only and couldn't be corrected.
- Duplicate, syndicated stories inflated the counts.
- Search and filters were too shallow for large cases.
- Monitoring was a single on/off toggle.
- No case status, owner, or audit trail.
- Entity analysis was hidden behind a "View Details" tab.
- Export was PDF-only and whole-case.
- Editable sentiment with logging, so the metrics are trusted.
- Duplicates collapse into one expandable cluster.
- Full-text and boolean search with saved filters.
- Configurable spike and anomaly alerts.
- Case status, an owner, and a full audit trail.
- Risk, findings, and entities surfaced up front.
- Multi-format export and a curated report builder.
08 Process
Audit
Mapped the case experience end to end and logged every gap.
Backlog
Scored 35+ ideas by impact, effort, and confidence.
Prioritize
Sorted into tiers and pulled quick wins to ship first.
Redesign
Restructured the case page and designed the features.
Build
Scoped MVPs with engineering and shipped in stages.
09 Outcome
The work turned Media Watcher's case experience from a reactive viewer into an investigation system of record, with a clear, prioritized roadmap behind it. Trust-first fixes and quick wins shipped first, with the larger feature line designed and ready to build.
- A redesigned single-case page that surfaces risk, findings, and entities up front.
- A new feature line across six areas, each with a staged MVP and advanced version.
- High-impact quick wins shipped first for immediate value.
10 Learnings
- Trust in the data had to come before flashy AI - editable sentiment, de-duplication, and credibility scoring made every other number believable.
- Scoring every idea for impact, effort, and confidence kept the roadmap honest and shippable.
- Surfacing the product's hidden differentiator, its entity analysis, was one of the highest-ROI moves in the whole revamp.