BANKING EXPERIENCE. A VIEW ON WHAT COMES NEXT.
Turning trust
into digital
growth.
At BCP, I led nine direct reports to grow monthly digital sales from $24M to $55M. Here’s the work behind that result, and how I’d apply it at Ria.
Explore the work ↓
DIGITAL SALES11 MONTHSThe outcome.
Growth came from connecting distribution, trust and first use. The decisions are below.
Three stories from BCP. One thread: make digital work for people who trust a person.
I’ve worked on
this transition.
Physical touchpoints, wallet distribution and the last mile of a financial journey. Explore the decisions behind the outcomes.
BCP / OMNICHANNEL GROWTH
The first digital sale starts with trust.
I connected Agentes BCP, wallet partnerships, WhatsApp, guided first use and the bank’s own digital channels around one goal: increase digital sales.
Use a trusted visit as the starting point, then help the customer complete the first digital journey.
My role & measurement
I led the team of nine accountable for the digital-sales program. Partnerships with Agentes BCP put offers on receipts after utility payments. Wallet partnerships, WhatsApp, teleoperators, push notifications, trusted deep links and media supported first use.
The result is reported monthly digital sales across the program. These journey stages do not represent measured stage conversions or a channel-by-channel attribution model.
Explore the related BCP case studyA utility-payment visit at a neighborhood Agente BCP creates a familiar moment to introduce a relevant offer.
servicios
A utility payment.
A trusted starting point.
The customer comes to a neighborhood agent for a task they already know.
Reconstructed interfaces, not historical screenshots. No personal data, banking connection or real transaction. Stages describe a journey, not measured conversion rates.
WALLET DISTRIBUTION
$0 → $20Mmonthly loan disbursementswithin six months
SIGNING EXPERIENCE
7 days → 4 minsigning timeacross a separate initiative
CUSTOMER EXPERIENCE
49 → 91NPS in the signing programreported before / after
Carry the trust.
Keep the promise.
The challenge I recognize: help customers move between channels without making them start over, question the offer or lose the confidence that brought them there.
02 / FIRST-HAND RIA TEARDOWN
One transfer scenario.
Four surfaces. Two totals.
I tested a $100 US → Peru quote with bank funding and mobile-wallet payout, in ChatGPT and Ria’s website.
About 11.91 PEN more for the recipient per $100.
A separate fee waiver reduces what the sender pays.
Ria card
inside ChatGPT
ChatGPT’s
written answer
Ria public
handoff page
Ria secure
account quote
What worked
Amount, destination, currency, funding and payout survived the public handoff.
What needs explaining
The price message and fee-waiver eligibility. Different account eligibility may explain some differences.
Open the test notes +
One corridor/session, one initial quote and one recheck. Ria’s public page advertised HELLORIA; the existing secure quote returned “This code doesn’t exist.” A new-tab continuation reached login; post-login continuity was not tested. No recipient, payment details or transfer was created. Raw payloads and root cause were not verified. Values are a dated observation, not current guaranteed prices.
Evidence retained: original quote captures, both totals in one viewport, promotional terms, the rejected code and the handoff. Originals containing private ChatGPT sidebar details stay out of the public page.
Inspect the actual screens.
06 CAPTURES / NOV 4, 2026
Real product screenshots, cropped to remove private browser and chat details. Prices are a dated observation, not a current offer.
03 / LISTENING BEYOND ONE JOURNEY
The promise continues
after the quote.
A high store rating can coexist with difficult individual journeys. I collected public reviews to find the recovery questions worth investigating.
Topics mentioned in retrieved low-rated reviews
Google Play · English/Spanish request feeds · Aug 7 / Nov 4, 2026
Keyword-coded mentions among the 264 low-rated reviews. Topics overlap and can be positive or negative. These are not customer defect rates.
WHAT THIS CHANGES
Pair offer clarity
with a recovery path.
Alongside the quote pilot, I’d test whether customers can understand where their transfer stands and what to do next, without starting over in another channel.
- Show the authoritative transfer state.
- Explain the next action and expected timing.
- Carry the same context into support.
The reviews do not establish Ria’s internal cause or incident rate. The HELLORIA test remains separate evidence.
Collection and analysis method +
Captured November 4, 2026. 399,398 returned sightings deduplicated by store and review ID to 31,005 written reviews dated January 3, 2018 through November 3, 2026. These are public endpoint results, not all historical or removed reviews. Apple: 249 storefront codes attempted; the US feed reached its 500-recent-review ceiling, with empty and HTTP 400/403 results recorded. Google: 120 country/language feeds exhausted.
The bounded chart uses unique Google Play IDs returned through English/Spanish request-language feeds. Feed language does not independently establish the writer’s language or residence. Themes were coded with deterministic keywords and manually spot-checked, not manually validated for every review. Reviewer identities and raw review bodies are excluded from this page.
04 / A PRODUCT BET
One quote.
Clear offers.
Every channel.
Before adding more places to send, make the promise consistent across the places customers already use.
Separate the two offers.
The exchange-rate benefit and fee waiver each get an explicit eligibility state.
Preserve the customer’s work.
Carry the quote, selected methods and relevant promotion context into Ria’s authenticated review.
Measure completed journeys.
First paid transfer, independent repeat and contribution. Keep support and loss guardrails.
PROPOSED OFFER CARDIllustrative redesign using the Nov 4 test snapshot. No live quote or transfer.
BUILD ON WHAT ALREADY EXISTS
The next interface
should share the same rules.
ChatGPT and WhatsApp are already part of Ria’s distribution. Personal agents make a consistent quote and approval boundary more valuable.
Agent location, app, web, WhatsApp or ChatGPT
Same quote, offer states and selected methods
Authenticated review and fresh approval if terms change
Authoritative status, recovery and a clear next action
Extend carefully. A Muse quote/discovery connector is a hypothesis; its current policy does not approve new money-movement tools. Ria integration with Instinct remains unverified.
The technical view +
Shared authoritative quote and eligibility data; platform-specific adapters such as MCP where supported; scoped account authorization for permitted actions; quote expiry and fresh approval when terms change; idempotent execution and truthful status. This is a proposal, not a claim about Ria’s internal architecture.
Dots shares connected ChatGPT plugin permissions; a Dots-specific Ria quote call remains to be tested, and money transfers require user takeover. Grok Bot supports connected apps and browser work; a Ria-specific connector is unverified. Start with permitted quotes and preserve the customer’s approval boundary.
Start with US → Peru and bank → wallet. Establish offer-comprehension and handoff baselines, then compare a controlled pilot with the existing flow. Track first paid transfer, 30-day independent repeat and retained contribution, with support-contact, fraud/loss and refund-time guardrails. These are proposed measures, not promised lift.
What I’d benchmark. Félix’s AI-driven WhatsApp remittance journey is a direct product reference; its documented payout options include Yape in Peru. Remitly’s WhatsApp sending and ChatGPT discovery are existing capabilities to compare. Stripe is a financial-infrastructure reference. For Ria, I’d test continuity from trusted retail introductions through first payment and independent repeat, rather than treating another chat channel as the outcome.
Sources & scope
BCP outcomes are my reported program results, with reconstructed journeys. The quote study is one corridor and account session, plus one recheck; it does not establish final checkout eligibility or an internal root cause. Review themes are customer self-reports. Product proposals are hypotheses.
Ria website ↗Ria WhatsApp journey ↗Muse platform policy ↗My resume ↗Full portfolio ↗