Porter Prime: The Gemstone Journey
Porter's partner-loyalty program had stopped doing its one job — telling partners what winning looks like. I used AI to explore 12 narrative directions, authored the two finalists, and the concept the team shipped turned tier status into something you can see.
the number that matters
Prime support tickets −80% (team-reported)
- role
- Concept author of the adopted design direction
- territory
- AI-assisted design · Loyalty · Visual narrative
- status
- quest complete
A loyalty program nobody could read
Prime is Porter's tier program for its driver-partners — the mechanism that's supposed to reward the most reliable partners and shape behaviour across the marketplace. It had stopped doing its one job: sending a clear signal about what it takes to win. Ask a partner what Prime was worth and you'd get a shrug. Tiers blurred together; performance scores lagged by up to a day, so the program felt untrustworthy; and a program nobody understands doesn't shape behaviour — it's just a badge.
The reader we were designing for doesn't read: busy, pragmatic, low reading comfort, many languages, moved by real things — earnings, time, status. The design question, sharply put:
How does an abstract tier system practically shout "here is your path to better earnings" — to someone who won't read a paragraph to find out?
Twelve narratives, two concepts, one metaphor
Rather than jumping to screens, I used AI to rapidly explore twelve narrative framings of what a tier journey could mean to a partner. The team narrowed them; I then authored the two strongest as complete visual concepts:
- The Gemstone Journey ("Porter Ascent") — partners evolve like minerals under pressure: a raw geode (unassuming outside, veins of potential) → polished quartz and silver → a faceted gold gemstone. Each tier specified fully: palette progression, textures (rough stone → brushed silver → polished gold), badge forms, crystalline progress iconography.
- The Energy narrative ("Porter Momentum") — the alternative: spark → current → radiance.
The gemstone won because it does the real work: it turns abstract status into a tangible object you're earning. The visual gap between a rough geode and a faceted gem says "you are not yet Gold" without a single word of copy — which is exactly what a low-reading-comfort, multi-language audience needs. (The seed of it came from a gem interaction in the Opal app I'd been carrying around as a reference — proof that a magpie reference library eventually pays.)
Five rounds with partners, one stubborn problem
Concept is half the job; comprehension is the other half. We ran five rounds of usability testing (Apr–Jul 2025) — deliberately starting in black-and-white wireframes so colour couldn't rescue a structure partners didn't understand:
- Rounds 1–2 — honest failure. Nobody could tell the three tiers existed, let alone the thresholds. "Same old thing." The premise itself was on the table.
- Round 3 — the breakthrough. Stop telling partners about tiers; segregate the tiers and their names directly on the progress bar and benefit cards. Status, target, and reward became visible in one glance. Comprehension landed for the first time.
- Round 4 — polish. Cross-tier benefit discovery; light vs. dark tested functionally equal (dark read "more premium"). One holdout: the Gold-benefit copy still confused.
- Round 5 — the copy fix. A focused concept test (29 partners, 3 cities) landed the words partners themselves use — "back-to-back orders," "order only for you" — not our jargon.
That discipline — B&W first to isolate comprehension, then iterating the one unresolved issue across rounds instead of declaring victory at the first pretty screen — is the part of this project I'd defend in any design review.
What shipped
Prime 2.0 turns the gemstone concept into a legible, real-time coach: a three-tier program (Gold · Silver · Regular) with a progress bar that plots tiers and thresholds so a partner sees the exact gap to the next status; a performance summary against the metrics that matter per vehicle type (in testing, partners often skipped the header and went straight here — familiar wins); near-real-time scores (the old one-day lag cut to ~10 minutes) with a "Last Updated At" timestamp — radical transparency about our own freshness; personalized nudges; and the gemstone system itself, engineered so tier names ship as a font and only the gem ships as an asset — a deliberate call that made vernacular translation across languages dramatically easier.
Impact — team-reported, credited honestly
Prime 2.0 launched 11 Sep 2025 and is live with 100% of partners across seven city-vehicle segments. From Porter's official launch announcement (where I'm credited in the design & research shoutout) — these are team-reported program numbers, not personally attributed impact:
- Prime support tickets down ~80% PAN-India (≈3,000–3,500 → ~700 weekly). A conservative analyst read isolating just Trucks — stripping out concurrent interventions — still shows a 59% drop (870 → 356 weekly), which is the number I'd defend hardest.
- Gold-criterion awareness +17pp in 2W (60→77%) and +11pp in Trucks (64→75%) — comprehension was the design goal, and it moved.
- ~85% of partners explore the Prime page after a homepage visit (up from 76%); ~60% of Gold/Silver partners check it 5+ times a week.
- 30+ partner interviews post-launch confirmed the score is finally trusted.
The honest read: as deep behaviour change, the jury's still out. On comprehension and trust — the design's actual job — it clearly worked.
What I carry forward
Visibility drives behaviour. Partners didn't need fluffier incentives; they needed an accurate, real-time mirror of their work and what it earns. And the AI meta-point for anyone hiring: the model generated twelve directions fast, but the craft was choosing the metaphor that made an abstract system legible to someone glancing at a phone between orders. AI accelerated the exploring; the judgment stayed human.