Illustrative scenario. Figures are modelled on typical Omnifys deployments and are not drawn from a named client engagement.
The challenge
A homeware e-commerce site with 400,000 monthly sessions showed the same bestsellers to everyone. Recommendations ignored session behaviour, cross-sell was manual and inconsistent, and AOV had been flat for six quarters.
What Omnifys deployed
- Session-based recommendations reacting to live browsing behaviour
- Complementary bundle suggestions at cart stage
- Consistent personalisation across site, email and chat
- Revenue attribution per recommendation surface
Agents used: Product Recommendation Agent, Product & Content Recommendation Agent, Product Propensity Agent.
Results
- 19% — increase in average order value
- 12% — lift in items per basket
- Consistent — personalisation across three surfaces
- Measured — revenue attribution per surface
Why it worked
Attribution was as valuable as the lift. For the first time merchandising could prove which surface earned the revenue.
Explore this for your business
Every Omnifys agent routes tasks across 15+ leading LLMs, integrates through n8n, Zapier and 300+ connectors, and ships with monitoring, audit trails and human-approval steps you control. Typical deployment runs two to six weeks from discovery to production. Book a free 30-minute consultation to map the highest-impact automation for your team.