The Recommender agent is the Genezio agent type that simulates a user who asks answer engines for recommendations in a multi-step conversation. Use it for scenarios where a persona looks for a solution that matches their needs and constraints.

​
How the Recommender agent works

For a Recommender agent topic:
  • The scenario defines the situation and the goal of the user.
  • Genezio makes a sequence of prompts from that scenario.
  • The AI conversation runs across many turns.
Example scenario:
From this scenario, Genezio makes conversational prompts such as these:
User query: I run a small startup and we’ve been tracking leads in a spreadsheet, but it’s not working anymore. What CRM tools would you recommend for a team of three?
User query: We need something that integrates with Gmail and stays under $50 per user per month. Which of those would fit?

​
Typical use cases for recommendation tests

Recommender agent conversations are useful when you want to examine these items:
  • The recommendation frequency of your brand
  • The position of your brand, when compared with competitors
  • How the recommendations change after the user adds constraints in follow-up prompts

​
Effect of the Recommender on the KPIs

Recommender agent conversations are a core input to these KPIs:
  • AI Recommendations: of the Recommender conversations in which your brand is visible, the percentage in which the answer engine recommends your brand.
  • Brand Visibility: Genezio includes these conversations because the answer engine selects the brands. The words of the prompt do not force the brands.

​
Multi-step behavior and realistic buyer journeys

The multi-step flow makes this agent useful for realistic buyer journeys:
  • A first, wide recommendation
  • A filter by budget, features, or team size
  • A shortlist of alternatives

​
Recommender data in the public API

The public API lists the scenarios and the conversations of a brand: List scenarios and List conversations.