A simple yet illuminating assessment!
Reviewing Discovery Calls with AI Summaries
A quick assessment of noun-verb pairs or the first few words in each summary section can be very enlightening:
- The more summary sections that describe the prospect’s situation, the better the call.
- The more summary sections that describe the vendor’s capabilities, the worse the discovery call.
Good calls have phrases like:
Prospect is seeking…
Current issues include…
Current challenges include…
Data is stored…
Prospect aims for…
Prospect aims to…
Prospect shares…
Prospect details…
Prospect detailed…
Prospect desires…
Prospect describes…
Prospect envisions…
Prospect mentions…
Prospect clarifies…
Prospect confirms…
Prospect needs…
Prospect expresses…
Prospect explains…
Prospect notes…
Prospect prefers…
The discussion…
Goals include…
Prospect highlights…
Prospect prioritizes…
Prospect currently…
Prospect uses…
Prospect is…
Prospect has…
Prospect is looking for…
Prospect inquired…
These phrases may also be indicators of good calls:
Vendor recaps…
Vendor confirms…
Poor calls include phrases such as:
Product is presented as…
Product offers…
Product is…
Product can…
Vendor introduces…
Vendor positions…
Vendor explains…
Vendor explained…
Product originated as…
Company originated…
Company is…
Vendor highlighted…
Consider a metric: The “Prospect:Vendor Statement Ratio” defined as the number of desires identified by the prospect vs the number of pitches by the vendor. Clearly, a good discovery call will have more prospect statements than vendor declarations.
Calculating the percent of prospect statements vs vendor statements can yield a rough but potentially illuminating metric.
For example, a call that shows 9 sectional summaries where 2 are prospect statements and 7 are vendor statements yields 22%:78% – that’s a pretty bad discovery call!
On the other hand, a call with 6 prospect statements and 3 vendor statements is quite a bit better: 67%:33%.
Of course, analyzing the full transcript may yield even more intriguing results, but I love that you can just scan the AI summaries and come to a rapid conclusion about the effectiveness of the call!
I’m sure this can be refined further… Thoughts?