Challenge

Altitude Sports has been selling outdoor and sporting gear to Canadians for over 40 years. What started as a single retail operation grew into two ecommerce properties: Altitude Sports, the flagship, and The Last Hunt, an outlet channel, spanning more than 350 brands, 150+ marketplace brands, and roughly 2.8 million SKUs. A dev team of 20-25 keeps it all running.

For most of that history, the commerce platform handled everything: catalog, cart, checkout, content, promotions. That sounds efficient until the seams start showing. Highly custom HTML templates meant the development team was required to support most changes on the front end. Homegrown workarounds were accumulating. Site performance was suffering. And the cost of maintaining it all was becoming prohibitive.

The incentive environment made the problem acute. At any given moment, Altitude operates 10+ campaign types simultaneously: membership discounts, gift cards, store credits, coupons, automatic cart promotions, shipping offers, always on, always interacting with each other in ways that got harder to predict as the catalog grew. 

What should have been marketing decisions were engineering projects. The team was sitting on customer data with no fast path to act on it.

Strategy

Altitude didn't go shopping for a bigger suite. They decomposed the problem.

Working with Orium as their systems integrator, the team took a phased approach, starting with proofs of concept to validate the architecture, then launching The Last Hunt on a fully composable platform before migrating Altitude Sports. 

The monolith gave way to best-of-breed, API-first tools. commercetools for commerce, Algolia for search, Contentful for content, Vercel for hosting, Fluent Commerce for order management, Bloomreach for CRM and cross-channel messaging, and Voucherify for the full incentive layer: promotions, loyalty, gift cards, and credits.

Each system does one thing well. Each exposes clean APIs. And each becomes a capability that other systems, including AI agents, can call. That's what made the next step possible.

Impact

Altitude Sports became the first live deployment of Vincent, Voucherify's AI incentive optimization agent. 

Vincent operates as a reasoning layer across the incentive stack. It doesn't replace Voucherify's promotion engine, it calls it. Two use cases are live today. First, customer support: when a customer needs a compensation voucher, Vincent generates it, checking if the customer exists, selecting the right campaign, creating the code, setting expiration, and assigning it. What used to be a multi-step manual process happens in a single prompt. Second, campaign creation: marketing describes what they want in natural language, and Vincent builds the campaign directly, surfacing competitive benchmarks on its own to inform the setup.

The CRM team of four and marketing team of five manage the full incentive operation, building campaigns, testing whether offers genuinely shift behavior, and protecting margin, without developer involvement.

Composable architecture in action

Altitude's architecture is what makes the agent work. 

That's the composable payoff, applied forward. Each best-of-breed system in the stack exposes a clean API surface. Bloomreach holds CRM and engagement data. commercetools holds catalog, cart, and margin context. Voucherify holds the full incentive layer. Because each system is API-first, an agent doesn't need a special integration to reach them. It needs an API call.

The more structured and accessible the data, the more an agent can reason about it. The pattern is replicable. Any brand running a composable, API-first architecture with structured promotion logic has the same foundation. The business defines targeting conditions, success criteria, and guardrails. The agent handles execution within those boundaries.