How to get real work done with Blu, your AI copilot

In this lesson

1. 1Go to Attributes, click Create attribute and pick Research under AI Attribute Types.
2. 2Write a real prompt in Instructions (optionally draft it from Guidance with Generate), then pick the Output value type.
3. 3Search for a sample user or account in the test panel and press Test; read the Analysis and Result.
4. 4Tighten the instructions if the answer was vague, then click Create attribute.
5. 5Optionally add a Research Agent step to a journey so it runs as people arrive.

Setup checklist

The 13 steps to put AI agents to work, from your first Blu chat to leads that are researched, scored and grouped. Tick them off as you go; your ticks stay in this browser.

0 of 13 ticked

1. 1.Open a journey, click Ask Blu in the console header (or press Ctrl+B) and notice the panel already knows the page you are on.
2. 2.Close the panel and finish a small task by hand to confirm the core product works without Blu.
3. 3.Send Blu a deliberately vague request and answer the clarifying question it asks back.
4. 4.Go to Attributes, click Create attribute and pick Research under AI Attribute Types.
5. 5.Write a real prompt in Instructions (optionally draft it from Guidance with Generate), then pick the Output value type.
6. 6.Search for a sample user or account in the test panel and press Test; read the Analysis and Result.
7. 7.Tighten the instructions if the answer was vague, then click Create attribute.
8. 8.Optionally add a Research Agent step to a journey so it runs as people arrive.
9. 9.Create an RFM attribute (the Lifecycle agent): pick the conversion event, set the Time window and the Recency, Frequency and Monetary tabs.
10. 10.Use the resulting attribute in a segment filter and as a report breakdown.
11. 11.Open Settings, Blu, and configure the Outreach methodology (T.I.P.S., P.V.P. or P.A.S.) and the Replies tab: Reply Grounding Data URLs and Draft Behavior.
12. 12.Create a Qualify attribute: on Fit score pick two or three fit attributes with weights, on Activity score pick two or three events, then create it.
13. 13.Open the attribute's Buckets tab to see the fit by activity matrix, filter users to the High fit / High activity bucket and open one lead before calling.

Your tech stack and tools

The core stack (everything you actually need)

• Intempt console (This is where Blu, your agents and their settings live)
• Blu Chat (This is your AI copilot. Ask it anything about your project)
• Research attribute (This is the agent that researches every new lead for you)
• RFM attribute (Lifecycle agent) (This is the agent that groups customers by how recently, often and much they buy)
• Qualify attribute (Qualification agent) (This is the agent that scores every lead by fit and activity)
• Tracked events with a conversion event (This is what the scoring agents learn from, so track real events first)

Highly recommended optional upgrades that level up your work

• Company knowledge (Settings, Blu) (This is how Blu learns to sound like you)
• Journeys (This is how an agent runs on every new person, automatically)
• Gmail (This is how Blu drafts replies right inside your real email threads)

Blu is included on every plan. Free gives you 15 AI credits a week plus 75 to start. Paid plans include 250 to 1,500 a seat a month.