Fiske Journalen had 30,000 Products and Nobody on the Floor to Advise. We Built the Advisor.
A catalogue that size buries the right rod. Neuramonks delivered a recommendation agent that talks like an experienced angler, narrows down 30,000 products to the right one, and attaches the tackle that goes with it.
Delivered for Fiske Journalen
- 30,000+
- Products covered
- 45%
- Faster product discovery
- 25%
- Higher conversion
- 30-40%
- More buyer confidence
Delivered for the L&D team
- 30,000+
- Products covered
- 45%
- Faster product discovery
- 25%
- Higher conversion
- 30-40%
- More buyer confidence
The Client's Problem
Thirty Thousand Products, and Nobody on the Floor to Ask.
- Filters and search across 30,000 items
- A conversation that narrows to the right gear.
- Shoppers guessing at compatible tackle
- Cross-sells attached to what they actually chose.
- Browsers leaving without buying
- Conversion up 25%, discovery 45% faster
What We Delivered
Six pieces of work, one advisor.
- 01 · Recommendation AI
- Intent-based product matching
- The AI product recommendation engine reads intent, not keywords.
- 02 · Domain LLM
- Fishing expertise layer
- Answers gear questions the way an experienced angler would.
- 03 · Applied AI
- Upsell and cross-sell engine
- Attaches the line, lure and leader that fit the choice.
- 04 · Data Engineering
- Catalogue attribute model
- 30,000 products structured so the agent can reason across them.
- 05 · Product Engineering
- Conversational storefront
- Shoppers ask in plain language and see live products.
- 06 · Analytics
- Engagement and revenue reporting
- Which conversations convert, and what they add per basket.
- Across all six · Shopper data
- Encrypted end to end
- Shopper sessions and order data encrypted in transit and at rest.
- Access-controlled by role
- Merchandisers and support see only their own scope.
- On-prem capable
- Deployable on-premises or air-gapped where policy requires it.
Big Catalogue, Small Conversion?
We scope which pieces your build actually needs, in 30 minutes.
The Result
A Bigger Basket, Not Just a Faster Search.
Recommending the rod is half of it. Because the agent understands what the angler is fishing for, it attaches the line, lure and leader that belong with it, which is where the conversion lift comes from.
25%
Higher conversion, with product discovery 45% faster across a 30,000-item catalogue.
How The Engagement Ran
Four Phases, Catalogue to Storefront.
- Phase 1
- Discovery
- Shopper journeys and catalogue gaps mapped with the team.
- Phase 2
- Catalogue model
- 30,000 products structured into reasoning-ready attributes.
- Phase 3
- Advisor agent
- Intent matching and fishing-domain answers tuned.
- Phase 4
- Upsell engine and rollout
- Cross-sell logic added, then released storefront-wide.
- LLM
- Recommendation engine
- Catalogue modelling
- E-commerce integration
Why Neuramonks
Why Fiske Journalen Choose Neuramonks.
- Outcome-driven delivery
- Conversion and discovery targets set before development started
- Domain-trained, not generic
- The agent answers like an angler, not a search box.
- Built for large catalogues
- Reasoning across 30,000 products, not a curated subset.
- Deployable on your terms
- On-premises or air-gapped where policy requires it.
Common Questions
What Teams ask about this Build.
How does an AI product advisor work?
It reads the shopper's intent, species, water, and conditions rather than keywords, then matches products whose attributes fit that situation.
How does it increase basket size?
Once the main item is chosen, the agent attaches the line, lure and tackle that genuinely belong with it, so cross-sells are relevant rather than generic.
Does it work on a catalogue this large?
Yes. It was built for 30,000 products, with attributes structured so the agent reasons across the full range rather than a curated subset.
What did the engagement include?
An AI product recommendation engine for intent-based matching, plus the fishing-domain layer, the upsell engine, the catalogue attribute model, the conversational storefront and reporting.
How is shopper data secured?
Sessions and order data are encrypted in transit and at rest, with role-based access and on-premises deployment available.
