How to Get the Best Results from the AI SDR
Overview
The AI SDR is designed to handle incoming prospect replies and automatically guide conversations toward booking a discovery call. This article covers common scenarios and best practices to help your AI SDR perform reliably and as intended.
Use the Knowledge Base for Context, Not Behavior Control
The Knowledge Base supplements the AI SDR's understanding of your business—it doesn't change how the agent behaves or makes decisions.
What to Include in Your Knowledge Base
Company background and mission
Product or service details
Common FAQs prospects ask
Industry-specific context or terminology
Pricing information or unique selling points
The AI SDR uses this information to personalize responses, but its core goal remains the same: move qualified prospects toward booking a call.
The Knowledge Base cannot override escalation triggers or change AI SDR workflows. For example, adding “send a calendar booking link instead of suggesting times” won’t change the AI’s behavior—it will still send available times for the prospect to choose.
Keep It Concise
You have up to 2,000 characters for your Knowledge Base. Prioritize the most relevant details prospects need during early-stage conversations. Avoid long paragraphs or instructions intended for the AI itself.
If you notice unexpected behavior, review whether your Knowledge Base content describes your business versus attempting to direct the agent's actions.
Keep All Prospect Responses Inside the AI SDR
For best results, all responses to prospects should flow through the AI SDR platform.
When conversations are handled partly inside the platform and partly through external tools or direct email replies, it disrupts conversation continuity and can lead to unexpected outcomes like duplicate messages or lost context.
Best Practice
Once a prospect replies, allow the AI SDR to manage the conversation end-to-end unless the conversation has been escalated for personal attention. If you need to intervene manually, use the platform's inbox to maintain full thread history.
For consistent conversation handling, all prospect replies should be managed inside the AI SDR. Responding from external tools or email clients can disrupt conversation continuity and lead to unexpected outcomes.
Creating Campaigns: Platform-Native vs Imported
If you use Instantly, how you create campaigns matters for AI SDR performance.
Platform-Native Campaigns
Campaigns created directly in the Cold Email Agent contain full metadata (like company fields and campaign intent) that the AI SDR uses to guide conversations. This ensures the most consistent and reliable behavior.
Imported Campaigns from Instantly
Campaigns imported from Instantly may lack certain metadata, which can affect how the agent responds. Common issues include:
Campaign type is not available, which means settings that depend on campaign intent (such as booking a specific call type) cannot be applied
CRM pipeline association is not included, which may limit the AI SDR’s ability to create or update deals automatically
Only new positive replies are ingested, so prior conversations from before the import are not visible to the AI SDR
For Legacy plans that use Instantly, use campaign imports only for existing campaigns that are already running. Going forward, create all new campaigns directly in the Cold Email Agent for the most consistent results.
Learn more: How to Create a New Cold Email Campaign | Connect Existing Instantly Accounts
How the AI SDR Handles Future Timing Requests
The AI SDR is designed to handle immediate communication with prospects (within 24 hours). When a prospect requests follow-up outside that window—like asking to reconnect next week or next quarter—the AI SDR will:
Recognize the delay request
Stop automated engagement
Escalate the conversation for personal attention
This ensures prospects don't receive delayed or out-of-context automated messages and gives your team the opportunity to take over at the right time.
The AI SDR uses your connected Google Calendar to check availability and propose meeting slots. Make sure your calendar is up-to-date to avoid booking conflicts.
Learn more: Setting Up AI SDR with Google Calendar | Appointment Booking Logic with Google Calendar
Understanding Escalations
Escalations are a feature, not a failure. The AI SDR escalates conversations when it detects scenarios that require human judgment, such as:
Security concerns or prompt injection attempts
Timing requests beyond 24 hours
Channel switch requests (e.g., "call me instead")
Complex questions the Knowledge Base doesn't cover
Signs of prospect frustration
Out-of-office or auto-response messages
When escalation happens, your team can step in with full context rather than trying to adapt the agent's behavior through configuration.
If you're seeing frequent escalations on simple replies, review your Knowledge Base to ensure it covers common questions. You can also dismiss false-positive escalations and re-enable the AI SDR for that conversation.
Learn more: Understanding AI SDR Escalations | Why the AI SDR Stops Responding
Additional Troubleshooting Tips
The AI SDR Isn't Responding
Common reasons include:
AI SDR is disabled for that conversation (check inbox filters)
You or a team member replied manually
Prospect opted out (DNC list)
Meeting already booked
Calendar disconnected
See: Why the AI SDR Stops Responding
Summary of Best Practices
Use the Knowledge Base for context (max 2,000 characters), not behavior control
Let the AI SDR manage conversations end-to-end whenever possible
Create new campaigns inside the platform for best results, not via imports
Expect escalation when timing or intent falls outside the standard booking flow
Keep your Google Calendar connected and up-to-date for accurate availability
Review escalations to improve Knowledge Base coverage over time
Following these guidelines will help the AI SDR perform predictably and reduce friction for both your team and your prospects.
For more details, see AI SDR Overview.