How to Measure Results (Leads, Conversions)

๐Ÿ“Š ChatForge Results Dashboard โ€” April 2026 CONVERSATIONS 247 โ†‘ 34% vs last month LEADS CAPTURED 43 โ†‘ 61% vs last month CALLS BOOKED 12 โ†‘ from 3 last month RESOLUTION RATE 82% โ†‘ from 71% LEADS PER WEEK Week 1 Week 14 โ†‘ Growing

You can't improve what you don't measure. Tracking the right metrics tells you whether your AI agent is performing, where it's falling short, and exactly where to focus your optimisation effort. This guide covers the key metrics, where to find them, and how to use the data.

The 5 core metrics to track

1. Total conversations

How many chat sessions has your agent had this period? This tells you whether visitors are actually finding and using the widget.

What to watch: If conversations are low, focus on widget placement and visibility rather than answer quality. You have a discovery problem, not a performance problem.

2. Leads captured

How many email addresses has the agent collected? This is one of the most direct measures of business value.

What to watch: Lead capture rate (leads รท conversations). A healthy rate is 15โ€“30%. Below 10% means your lead capture prompt needs improving or the offer isn't compelling enough.

3. Calls / appointments booked

How many booking link clicks came from the chat widget? If you have booking confirmation tracking set up, how many of those translated into actual booked slots?

What to watch: Booking rate from conversations. Even 5% is significant โ€” 5 out of every 100 conversations converting to a booked call can be worth thousands of pounds/dollars monthly depending on your deal size.

4. Resolution rate

What percentage of conversations end with the visitor's question answered (without needing to escalate to a human)? Higher is better, but aim for quality of resolution, not just the number.

What to watch: A drop in resolution rate is usually a sign that your knowledge base has gaps โ€” visitors are asking questions your agent can't answer.

5. Top unanswered questions

What questions is your agent failing to answer? This is your product roadmap for knowledge base improvements.

What to watch: Review these weekly. Any question appearing 3+ times that the agent can't answer should be added to the knowledge base immediately.

Where to find this data

In your ChatForge dashboard, go to the Analytics tab to see:

  • Total conversations and daily/weekly trends
  • Lead capture count and email list
  • Booking CTA click rate
  • Conversation transcripts (review these for quality)
  • Most common questions asked
๐Ÿ’ก
Set up a monthly review ritual On the first Monday of each month, spend 20 minutes reviewing your analytics. Note trends, review a sample of conversation transcripts, identify top unanswered questions, and make 3โ€“5 knowledge base improvements. This habit compounds significantly over time.

Setting realistic targets

MetricMonth 1 (new agent)Month 3 (optimised)
Conversations50โ€“100150โ€“300+
Lead capture rate8โ€“15%20โ€“35%
Booking rate2โ€“5%5โ€“15%
Resolution rate60โ€“70%80โ€“90%

These are benchmarks โ€” actual results vary significantly by industry, traffic volume, and how well the knowledge base is configured. The trend matters more than the absolute numbers in early months.