Lesson 2 of 5

Lesson 3.1.2

AlexAI: chats, automations and memories

6 min read2 min videoAdvanced

What this costs

Whatever the underlying tools cost. AlexAI orchestrates them, it does not replace their pricing.

Before you start

A project to work on and a saved company context.

AlexAI is the conversational way into the same engine. You describe what you want and it picks the builders and tools rather than you configuring columns by hand. It is fastest for exploration and for one off work, and it needs the same cost discipline as everything else.

The Alex Agents screen in Kuration, with tabs for chats, automations and memories, and a list of conversations that each created a project
1
2
3
  1. 1Three tabs: Chats, Automations, Memories. Automations are the ones that run without you.
  2. 2How much has actually gone through it: saved conversations, messages, and how many produced a project.
  3. 3Created Project. A conversation that ended in a real table is the only kind worth counting.
Chats, automations and memories, in one place. The column that matters is Created Project: these are conversations that produced a real list, not answers about how you might build one.

When to reach for it, and when to build columns

Use AlexAI whenBuild columns when
You are exploring and do not know the shape of the answerYou know exactly what you want and will want it again
It is a one off question about a listIt is a repeatable workflow
You want a starting point to then refine by handYou need conditions, scheduling and reuse
You are demonstrating something to someoneYou are running production lists

The productive pattern is both: explore in conversation, then rebuild the version that worked as a proper column stack with conditions, and schedule that.

AlexAI being used inside Kuration to build a list conversationally
Conversation is the fastest route to a first answer. Columns are the route to a repeatable one.
Alex explaining a running enrichment beside a Kuration table of 300 companies
1
2
3
  1. 1The column it created, named for what it does.
  2. 2The title ladder it used, in priority order. Same model as lesson 1.2.5.
  3. 3Progress, in plain language, including how many rows are processing right now.
Alex working inside a real project. It names the tool it chose, the titles it prioritised, the fields it will return, and how far through the run it is.

Check yourself

0 of 2 answered

1.When is a conversation the right tool rather than a column stack?

2.What should you give it rather than re describing your ICP each time?

Answer all 2, then open Suppression before spend: never pay for a row twice to tick this lesson off.

This lesson is written and complete. Its screen recording is still to come, and the written walkthrough is designed to stand on its own without one.