Best Practices

Tips for getting the best results from AI Prospecting.

Writing ICP Prompts

Be Specific

Good: "Dental practices with 5-20 employees in Texas
that don't have a full-time receptionist"

Bad: "Healthcare companies"

Include Pain Points

"Small law firms that lose clients when calls go
unanswered during court appearances"

Mention Decision Makers

"Looking for owner-operators or office managers
who handle vendor decisions"

Target Count Selection

Start small with new ICPs to validate before scaling.

Campaign Preparation

Define ICP Filters First

Campaign filters guide the Harvester. Set:

• Industries
• Company size range
• Target states
• Seniority levels

Write Clear Description

Campaign description helps agents understand context beyond filters.

During Sessions

Don’t Interrupt Unnecessarily

Let sessions complete naturally. Stopping mid-way wastes evaluated companies.

Monitor Quality

Watch the Leads tab as leads come in. If quality seems off, you might stop early and refine ICP.

After Sessions

Enrich Immediately

Contact data gets stale. Enrich within hours of session.

Prioritize High Scores

Start outreach with 18+ scored leads. They’re 3x more likely to respond.

Review Rejections

Check why companies were rejected. Patterns reveal ICP refinement opportunities.

Common Mistakes

Optimization Tips

Batch Sizes

Default settings work for most cases. Adjust in Settings > AI Prospecting only if:

• Sessions timeout → Reduce batch size
• Need faster results → Increase concurrency

Tavily Usage

Web research improves accuracy but uses credits. Default 50 lookups per session balances cost/quality.

Session Timing

• Run sessions during business hours for faster ZoomInfo responses
• Avoid running multiple sessions simultaneously

Success Metrics

Track these to measure AI Prospecting effectiveness:

If metrics are below targets, refine your ICP or review agent settings.