Automate research. Then generate *specific* ICP signals & fields.

It's challenging to identify nuanced attributes that truly indicate fit without spending hours manually researching each account. AI Agents solve this by automatically researching, analyzing, and enriching your ICP accounts.

Our goal is to answer ANY question for EVERY account in your market.

If you've experimented with AI-powered tools before, you may have found that they require significant prompt engineering or produce inconsistent results. Keyplay's approach is fundamentally different. You don't need to be an AI expert. Our agents combine LLMs with testing tools, quality data inputs, and specialized prompt cascades. The outputs are automatically structured and integrated with your ICP model and CRM.

Our Agents answer questions about accounts. Each agent will produce one or more custom fields on every account. Those custom fields can be used to create scoring signals, to continuously enrich CRM records, or to build target accounts lists in Keyplay.

Navigate to the Agents tab in your Keyplay dashboard. You'll see three types of agents available today, with many more coming soon.

Select some preview accounts so you can check your agent's work as you go. Once publish your agent, you'll be able to run it on your whole TAM.

You can start with a simple question. Then add clarifications as you build. Much of the prompt engineering is built in with Keyplay agents.

Before running your agent across all accounts, Keyplay allows you to run quantitative tests on lists of accounts. You can use the same test lists that you use for your ICP Model back-testing. Note that bigger tests can take a few minutes to run.

After building your agent, set it live on all your accounts or limit to a specific tag. You can set your agent to automatically run on new accounts and set a schedule to refresh data.

Once you've published your agent, you can create custom signals to use in scoring, segmentation, and list building.

Category field types will automatically create signals based on the category names

True/False fields types will automatically create signals for "Has {{Field Name}}."

1-10 Signals can be translated into ranges.
Go to the Signals menu in settings to create and manage your custom signals.

Once you've published your agent, the fields generated will be available in your CRM field mapping under "additional fields."

Agent work on a credit system. Each successful Agent execution will use credits based on the type of agent. Agents range from 1 credit for a basic Analyst to 4 credits for a Researcher and Analyst.

You credit usage for a given period will = Agent Credits X Number of Accounts X Refreshes. Credit will be communicate on the Publish screen and credit usage is tracked in your admin settings.

AI Agents transform how you research, score, and segment your market by automating the discovery of nuanced fit signals. Instead of relying solely on basic firmographics, you can now identify accounts based on business situations, use cases, and unique value alignments that truly predict success.

Get started today by creating your first agent and testing it on a small segment of accounts. Once you see the quality of insights, you can scale to your entire CRM and integrate the results into your ICP model.

Need help? Your dedicated Customer Success Manager can assist with agent setup, or you can book a custom Agent Workshop with our team.