How Dreaver Reached 82% Automation on Highly Technical Products
Dreaver is a leading European walking pad brand, selling across the Nordics and beyond. Their products are compact machines with real specs: motors, automatic incline, load ratings, noise levels, app pairing and model-specific accessories. An AI Agent from Minimal now handles the majority of that technical support in Gorgias, Shopify, Channeldock and Klaviyo. On the storefront especially, it recommends the right model while shoppers are still deciding.
The Challenge
Dreaver does not sell a single walking pad. The range is technical: the Motion Pro for walking and running up to 12 km/h, the Elevate Pro with automatic incline to 9%, the Assist One with optional side rails for extra stability, plus app control, remotes, bottle holders and recovery products. They sell across Europe, including the Nordics, so support is not a single-market problem. Pre-purchase questions are spec-heavy. Post-purchase questions are too: motor noise, Bluetooth pairing, display readings, belt tracking, warranty checks.
That is hard to automate with a generic chatbot. A wrong answer on incline, load capacity or whether a model fits under a sofa is not a small miss, it is a wrong product. Dreaver needed an agent that could work from live product data in Shopify, fulfillment in Channeldock, customer profiles in Klaviyo, and the macros Anouk already used in Gorgias, without growing the support team.
“Really good service. If something comes up, they look for a solution right away, even on the weekend.”
Recent Trustpilot review on dreaver.com
The Solution: Real Actions, Not Just Replies
Dreaver went live in August 2025. The AI Agent connected to Gorgias, Shopify, Channeldock and Klaviyo in onboarding, then learned from the five highest-volume topics and macros Anouk already ran. Automated replies go out directly. Suggestions land as internal notes, ready to send.
The same agent is now live as storefront chat on dreaver.com. That channel is especially strong for product recommendation: matching a shopper to the Motion Pro, Elevate Pro or Assist One from live specs, then escalating to a ticket only when a human is actually needed.
Storefront product recommendations
This is where the agent is at its best. On dreaver.com it asks what someone needs, under-desk walking, extra stability, or incline, and recommends the right model from the live catalog. Spec questions turn into a product match on the page, before a ticket is ever opened. Internally, Dreaver's storefront sits alongside Cloudpillo and XXL Nutrition as a benchmark for low escalations.
Technical product advice
Choosing a Dreaver is a spec decision, not a style one. The AI Agent compares models on motor, speed (1–12 km/h in 0.1 increments), automatic incline up to 9%, noise, load capacity, fold height, and whether the Assist One needs side rails. It uses live Shopify catalog data instead of a generic FAQ.
What sets the deployment apart is that the AI Agent also takes real actions across Dreaver's systems, resolving cases that normally mean jumping between the warehouse, the store and marketing:
Address changes in Channeldock & Shopify
When a customer needs a different delivery address, the AI Agent updates it in Channeldock before the warehouse picks, and keeps Shopify in sync, without anyone on the team opening either system.
WISMO: tracking, updates and postal investigations
When a customer asks where their order is, the AI Agent reads the track and trace, tells them where the parcel is, and either handles a postal investigation or escalates the edge cases Anouk still needs to see.
Order holds in Channeldock
Putting an unpicked order on hold is an action currently being built. Once live, the AI Agent will pause fulfillment in Channeldock when a customer needs a change, then confirm back in Gorgias.
Refunds and order edits in Shopify
Cancellations, refunds and order edits run in Shopify end-to-end, with a clear confirmation to the customer instead of a back-and-forth for the team.
Profile and list updates in Klaviyo
Marketing opt-outs, list changes and profile updates happen directly in Klaviyo. The AI Agent looks up the customer, applies the change, and confirms, so Anouk does not have to jump into another tool.
Manuals and ebooks on demand
When a customer is missing a user manual or ebook, the AI Agent sends the right link in the reply, so they can set up or use the product without waiting for the team.
More than 82% of incoming tickets are now handled fully end-to-end, without anyone on the team in the loop. The remaining cases arrive as ready-to-send replies in Gorgias.
The Results
Technical questions that used to need a specialist are now answered automatically. On the storefront, that same knowledge turns into product recommendations at the moment of choice.
Ticket Automation Rate
- 82% ticket automation on a catalog where answers depend on motors, incline, noise, load ratings and model-specific parts, not a one-size-fits-all script.
- End-to-end actions in Channeldock, Shopify and Klaviyo. Address changes, cancellations, refunds, WISMO tracking, manuals and marketing opt-outs run in their systems, not as tasks left for the team. Order holds in Channeldock are currently being built.
- Storefront as a product recommender. The agent is especially strong on the site: it turns spec questions into a model recommendation before checkout. Dreaver is used internally as a benchmark for low storefront escalations, next to Cloudpillo and XXL Nutrition.
- Service that holds on Trustpilot. Dreaver keeps a 4.6 score across 3,400+ reviews from customers across Europe. They still call out speed, including replies at the weekend.
What This Means
Technical products are usually the last place teams trust AI. Dreaver shows the opposite: when the agent has live product data, it can recommend the right walking pad as precisely as a specialist, and then actually change the order, the warehouse and the marketing profile, for customers across Europe. That is most visible on the storefront, where recommendation happens before the ticket exists, at a volume no lean support team can match by hand.
Published on August 14, 2026 by Niek Hogenboom