[2 min read] Steal Our Secret AI Agents

My business partner will be upset I did this....

We are in this for the long game so I don’t mind if you steal our secret agents, we plan to deploy many different AI solutions from big to small. 

To help you grasp the mental framework, here is a breakdown of how we build our secret agents…

Data sources 

Where are the Data points coming from? These are going to be the triggers for the ai agents. For the sake of sales, here are some examples of Data Sources

  1. If you are running ads, it will be a form

  2. Inbound Calls, Text, and Emails

  3. Tags in the CRM on a batch of old leads

  4. Intent Lead Alerts

  5. Landing on a website and unmasking contact data

Data Aggregation  

These will be the pipes behind the scenes of your AI agent. How data will move across different tools and platforms that are being used. Here are some examples:

  1. Zapier is the most commonly used

  2. Make is the most up and coming

  3. Lindy.ai 

  4. I have 2 more but thats a secret I can tell…. yet

Data Center/Warehouse 

Ok, now let’s talk about the hub where your AI agent is going to harvest the information being used to do “action” this is the command post for your agent. Here are some examples of Hubs.

  1. CRMs that collect many data points like opens, clicks, and activity

  2. Google Sheets is the most basic its a “poor mans CRM”

Data Enrichment 

Now data enrichment is about using AI within your Data Hub/Center/Warehouse to feed your output framework with the most complete information. These are the small things that push the needle, I like to call this “inside baseball”. Here are some examples

  1. Research of prospects to help personalize communication

  2. Analyzing the data they submitted to extrapolate predictions or solutions

  3. Pre-qualifying prospects before even reaching out

Data output frameworks 

Now this is the part where things get good. If you don’t want your agent to sound like a robot, you have to nail this. Here are things you need to add into your framework to have an agent sound “human”.

  1. Objectives and Clear Scenarios

  2. Speech Patterns

  3. Knowledge Base

  4. Background Sounds

  5. Answers to the hardest questions they could get

  6. Objection Handling and Looping

  7. When to pause and when to show “empathy” (Program it into a predictable Q&A)

Hopefully you found this newsletter useful, feel free to connect with me on LinkedIn Desmond Dixon. Don’t be shy, we are in this together.