By the end of this tutorial, you’ll know how to build an AI Signal agent in lemlist, configure the signal you want to monitor, choose the right scope, define how signals should be processed, and deploy the agent.
Why this matters
AI Signals help you monitor target accounts for highly specific business events that matter to your outbound strategy. Instead of relying only on predefined signal types, you can create custom signal questions and let lemlist monitor the web for relevant matches.
Before you start
You should already know:
What a company list is in lemlist
That an account usually refers to a company name monitored by the agent
Which signal processing workflow you want to use after a match is found
Key concept: AI Signals let you create custom monitoring questions about your target accounts and then decide how lemlist should handle matching results.
Core lesson
Phase 1: Start a new signal agent and choose AI Signals
Go to Signal agents, then click Create signal agent. This opens the setup flow where you’ll choose the signal type, configure it, define the scope, and set the processing method.
In the Signal to detect step, expand Buyer intent and select AI Signals.
Phase 2: Configure your AI signal
In the Configure your signal step, enter a Signal name, choose the question opening from the Question to monitor dropdown, and type the rest of your question. Use Add a signal to detect if you want to monitor more than one question in the same agent.
Signal formatting rules to follow:
Signal name: keep it short and easy to scan
Question opening: choose one of the available dropdown formats
Question body: type the rest of the question in free text
Language: questions need to be written in English
Important: Questions need to be in English, but results can still appear in other languages if lemlist detects them.
Use templates to speed up setup
If you do not want to write every signal from scratch, you can use one of the suggested templates shown in the Templates section.
Click Use template next to any suggestion you want to add.
Phase 3: Choose the scope and identification limit
In the Scope step, choose how broadly you want the agent to monitor accounts. You can select All segments, Company list, or Specific segment.
Then set the Maximum number of signals identified per week in the Identification limit section. This determines how many signals the agent can identify weekly and shows the maximum credit usage for that setting.
Available scope options:
All segments: retrieve all companies that trigger the signal
Company list: monitor companies from an existing company list in lemlist
Specific segment: retrieve companies that belong to a segment using a CSV file
Phase 4: Decide how identified signals should be processed
Next, choose what should happen when lemlist finds a relevant signal.
You can review signals manually, auto-create tasks, or auto-push leads into a campaign. In this example, Auto-push to campaign is selected, then AI-powered outreach is chosen as the campaign type.
Phase 5: Review the summary and deploy the signal agent
Before launching the signal agent, review the summary carefully. You’ll see the selected signal type, the configured questions, billing details, processing setup, and scope.
When everything looks correct, click Deploy agent.
What it unlocks
You can monitor highly specific account-level signals with a dedicated AI Signal agent
You can control how many signals are identified each week
You can route detected signals into manual review, tasks, or campaigns
Practical example
Here’s a simple real-world setup for an account-based sales team:
Signal agent: AI signal for expansion accounts
Signal name: M&A
Question: Did the target account acquire or merge with another company?
Scope: Existing company list or all matching segments, depending on your strategy
Processing: Auto-push matching leads into an outreach campaign
This setup works well when you want to act quickly on high-intent or high-change company events.
Best practices
Write questions in natural English so lemlist can interpret them accurately
Keep signal names short and recognizable for faster review
Start with a focused weekly identification limit so you can control credit usage
Use templates when you want to launch quickly, then customize later
Choose the processing method that best matches your team workflow
Troubleshooting and common pitfalls
Issue: I can’t move forward from the signal configuration step
Root cause: One or more required fields are incomplete.
Fix:
Make sure you entered a Signal name
Choose a valid Question to monitor opening
Type the rest of the question in the free-text field
Issue: The signal agent is not finding relevant results
Root cause: Your question may be too vague, too narrow, or not phrased like public web content.
Fix:
Rewrite the question to sound closer to how the event would appear in articles or LinkedIn posts
Test broader wording first, then refine it over time
Avoid internal jargon that is unlikely to appear publicly
Issue: Billing is higher than expected
Root cause: Your weekly identification limit is set too high.
Fix:
Review the Identification limit before deploying
Use the summary step to confirm the expected credit usage
Lower the weekly limit if needed
Issue: I’m not sure which scope option to choose
Root cause: Different scope options fit different monitoring strategies.
Fix:
Use All segments if you want the broadest monitoring
Use Company list if you already have a list in lemlist
Use Specific segment if you want to monitor a CSV-based segment
What happens next?
Once your signal agent is live, lemlist starts monitoring for the signal you configured. As signals are identified, they are processed according to the workflow you selected, helping your team act faster on relevant account intelligence.