How Grupo Garden scaled WhatsApp commerce across 10 auto brands with Customer.io and Hyppo

Garden hadn't grown to category leadership by accident. Years of investment in showrooms, digital marketing, salesforce training and brand acquisitions had built the largest auto retail footprint in the country. But the volume that footprint generated had outpaced the systems meant to manage it.

The legacy WhatsApp bot was processing massive lead volume across all brands. What it wasn't doing was producing data anyone could act on. Leads entered the bot and disappeared. Hot buyers and casual browsers got the same flow. Conversations that went quiet stayed quiet. And every brand operated as its own island: any improvement made for Nissan had to be rebuilt from zero for KIA, then again for Chevrolet, then again for every other brand in the group.

"We were selling tens of thousands of vehicles a year and had no idea what happened between a WhatsApp lead and a signed contract. The funnel existed, we just couldn't see it. Today we follow every lead from the first message to the SAP invoice, across all ten brands, in one place."

That's the moment Hyppo, a Buenos Aires-based Data & Growth agency, was brought in to audit the channel end-to-end and design the replacement.

The diagnosis surfaced a set of interlocking problems. Following a single user from first WhatsApp message to invoiced sale in SAP was effectively impossible. Internal interviews surfaced the same gap, repeatedly. "I don't know how many digital leads we receive. I have no metrics," said Diego Benitez, Commercial Manager for Fiat, in one of those early conversations. Beyond visibility, every lead got the same treatment regardless of intent: someone planning to buy in the next 60 days entered the same flow as someone casually browsing for a friend. And when a conversation went quiet (which happened constantly), there was no follow-up, no nurture, no second chance. Worst of all from an operational standpoint, every brand operated as a silo, which meant that any improvement made for one brand had to be rebuilt from scratch for the next nine.

The brief Hyppo received was specific: design a stack that could see every lead, segment in real time, reactivate at scale, and critically, be replicable across the entire group without rebuilding it ten times.

Hyppo placed Customer.io at the center of the architecture as the CDP and engagement automation engine. It became the system of record for every lead, the source of truth for segmentation, and the trigger layer for every automated touchpoint in the funnel.

Around it, each tool was chosen for one job. GoChat runs the conversational layer with an AI agent built on seven skills (initial conversation, model selection, credit, test drive, external-brand interest, human handoff) and nine AI tasks handling everything from lead classification to follow-up message generation to data validation. Microsoft Dynamics 365 holds the credit files. SAP runs invoicing and after-sales. Looker Studio on top of BigQuery delivers unified reporting across the entire stack.

The division of labor is clean. GoChat talks to the customer. Customer.io handles everything else: it identifies the user, stores attributes, tracks events, segments in real time, fires workflows, and orchestrates the bidirectional flow with Dynamics and SAP.

"GoChat talks to the customer. Customer.io is the brain. Every decision (who to follow up, when, with what message, which salesperson takes it, when to escalate) lives in Customer.io. That separation is what let us scale without rebuilding."

Every interaction in GoChat emits an event to Customer.io. Every event can trigger a workflow. Every workflow can write back to GoChat to send a WhatsApp message, alert a salesperson, or create a task in Dynamics. That round trip is what turns a passive bot into an active sales engine.

The most consequential workflow is lead classification. Once the bot has gathered enough signal during the conversation, an AI task scores the lead as Hot (buying intent within 0-60 days), Warm (intent beyond 60 days) or Cold (general inquiry only), and writes the result back to the Customer.io profile. That single attribute drives the entire downstream automation. Hot leads enter aggressive cadences inside the 24-hour WhatsApp messaging window, with pushes every few hours until the customer responds. Warm leads enter a slower 7-day nurture. Cold leads get routed to the call center for human reactivation. Same architecture, three completely different experiences, decided automatically by data the bot itself generated.

Test drives surface the same logic in a more visible way. When a customer books a test drive in the bot, Customer.io fires two journeys in parallel. One goes to the customer (confirmation email plus WhatsApp reminders to lift show-up rate) and one to the assigned salesperson (immediate alert with full lead context, plus a second reminder 24 hours before the appointment). The salesperson walks into the conversation already knowing the model, the financing scenario, and the customer's previous interactions. Conversations that used to start cold now start halfway through.

Follow-ups work the same way. When the AI detects a future commitment in the conversation, like "I'll get back to you next Tuesday" or "call me after I get paid", it extracts that date and writes it to the Customer.io profile. Customer.io waits, and on the right day fires the right message to the right person. One workflow, ten brands.

Identity resolution works across channels. Web forms via GTM, TikTok lead forms, QR codes at events like the Cadam MotorShow, and walk-in showroom registrations all flow into the same Customer.io profile layer. A lead who scanned a QR at the auto show on Saturday and walked into a KIA showroom on Tuesday is recognized as one person, not two. When a lead is closed out, the discard reason gets written back as an attribute, enabling cohort analysis on why leads die at every stage of the funnel.

Garden and Hyppo started with Nissan in November 2025. Seven months later, every brand in the group's repertoire was live on the same Customer.io-powered architecture.

Each new brand inherits the same workflow library, the same event schema, the same dashboard template. Onboarding a new brand stopped being a development project and became a configuration exercise. Hyppo can now stand up a new workspace in days, not months.

"Onboarding our last few brands took less time than onboarding the first one took weeks. Every workflow we built for Nissan was already there for Volvo. That's not integration, that's compounding."

Starting from a legacy bot no one could see into, Garden now operates a single CDP serving 100% of the group's repertoire across web, WhatsApp, showroom and event capture.

"Before, the same KPI gave us three different numbers. Marketing reported one figure, lead management another, and the commercial team a third. Nobody was wrong, everybody was looking at a different fragment. Today there is one source of truth, and the traceability improved dramatically. A lead that comes in through a KIA web form on Monday and ends up walking into a Nissan showroom on Friday is one person in our system, and marketing, sales and after-sales are all looking at the same record."

The same Customer.io foundation is being extended beyond new-vehicle sales. The next wave includes a single entry point routing customers to the right brand, used-vehicle, subscription and rental flows, plus operational AI on top of the conversations themselves.

What started as a question about a broken WhatsApp bot turned into something much bigger: a unified commerce engine that lets a national retail leader finally see, segment, and act on every lead across every brand, every channel, and every moment of the buying journey.