Intro to Agents

Intro to AI Agents

AI Agents in Unify are programmable assistants designed to research, qualify, and score records based on natural-language prompts. These agents go far beyond static filters by automating real-time research—surfacing nuanced, high-intent buying signals that traditional CRMs often miss. Whether you’re trying to determine if a company is B2B, check if they use a specific technology, or analyze sentiment from online reviews, AI Agents give you a flexible, intelligent way to drive precision in your outbound motion.

To create your first agent, head to the Agents tab in Unify and click “New Agent.” From here, you’ll configure three components: a name, the record type (Company or Person), and one or more open-ended questions you want the agent to answer. Each question includes a prompt or instruction, such as “Check the company’s website to determine if it uses Auth0,” along with a response type—True/False, Number, Multiple Choice, or Text. This setup allows you to encode custom signal logic tailored to your outbound strategy.

Let’s walk through a few examples. If you only sell to B2B companies, you can build an agent called “Is the company B2B?” that searches the web, visits the company’s website and LinkedIn page, and returns a yes or no answer with justification. Need tech stack detection? You could ask “Does the company use Auth0?” and have the agent inspect the site’s HTML or JavaScript tags for evidence. For softer signals like support quality, an agent can research review sites and forums, analyze sentiment, and return either a True/False result or a summary text that can even be used in your outbound messaging.

After creating an agent, you can run tests by selecting sample companies. As the agent operates, you’ll see its thought process—what it searched, which sources it reviewed, and how it arrived at its answer. You can refine the prompts to improve accuracy and reuse agents across different plays once they’re working well.

When you’re ready to use an agent in production, it becomes a trigger or condition inside a Play. For instance, you can launch a sequence only if the company is B2B, send alerts to your team when sentiment turns negative, or spin up a competitive displacement play when the agent detects a rival tool in use. This level of dynamic logic turns every play into an intelligent, real-time automation.

You’ve now learned the basics of creating and deploying AI Agents in Unify. These agents unlock a new level of contextual targeting, allowing you to automate what was once manual research—at scale. For even more inspiration, check out the AI Agent playlist on Unify’s YouTube channel for advanced strategies and creative outbound workflows.