Getting started quickly

Use existing knowledge bases and bots as recommendation sources to get started in just a few clicks.

If you're starting from scratch, try the tutorial to create a proof-of-concept.

Thinking about which agents should get recommendations from which sources

You can control which sets of knowledge and automation you want to expose through each skill serviced by human agents. Optimize recommendation results by only connecting relevant sets of knowledge and bots to particular skills. For example, an agent servicing billing questions should receive billing-related recommendations.

Providing agents with multiple recommendations

The most common way that agents will receive multiple recommendations for an utterance is if there are multiple recommendation sources connected to the same skill.

Rolling out recommendations to agents

When introducing your agents to Conversation Assist, you can control rollout and set up training groups by enabling recommendations for particular skills.

Automation first

The best performing automations for Conversation Assist are those that offload repetitive and time-consuming tasks from agents.

Answers enriched via Generative AI

Are you offering enriched answers created via Generative AI as recommendations to your agents? See our best practices.