CONTEXT
AI Teammate for Live Sales Calls
Real-Time AI by Amotions Inc. is an AI sales assistant that AI Sales Teammate that guides sales reps during live prospect calls with real-time suggestions.
Live Suggestions to Sales Reps
Emotional Analysis of Prospects
Post Call Analysis and Scoring
PROBLEM
Initial Ask
"We want a way to measure our sales reps’ performance in live calls!"
Stakeholders
As straightforward as that sounded, our deeper conversations with both stakeholders and users revealed something else beneath the surface.
Stakeholder Expectations
Assess Sales Rep Performance
during calls with prospects and clients
Scoring System
to rate sales reps during their calls
Why Scoring was Hard
No two Sales Calls are alike
A single Scoring System can’t capture the reality of all sales calls, because every call has a different goal, approach, and agenda.
Pure AI Scores were inconsistent
For the AI to properly evaluate a sales rep’s performance, it needs access to the company’s RAG (Retrieval-Augmented Generation) documents.
Stakeholders requested Detailed Scoresheet
The requirements made it clear that what was needed was a thorough, detailed assessment of the salesperson’s performance.
Building the Framework
I got to work and built a comprehensive scoring system tailored to different types of sales calls. The core idea was to break each call into distinct segments, allowing me to evaluate the salesperson’s performance in every segment.
Testing Reality
A month-long pilot test with a client quickly revealed the ground reality.
One system doesn’t fit all calls
Scores improved when reps gave context
Human input was still essential
SOLUTION
The Breakthrough
I asked our client managers to walk me through how they rate calls at their company and noticed a pattern.
Managers didn’t have time to review detailed score sheets.
Every company had its own approach, and the AI could not provide a score without first having the playbook in its knowledge base.
What Managers really needed
Their only concern was whether the sales rep followed the playbook during the meeting.
They needed a simple scoring system with results that could be reviewed at a glance.
SHIPPED DESIGN
Shipped Design
A simple comparison with the company’s sales playbook worked best.
A simple checklist was easier to scan and make quick decisions.
Easy to spot errors or ambiguity the AI may have missed.
Saved server load as scoring calculation became easier.
Pilot Results - 3 Months, 4 Companies
Before - Hard to assess
After - Easy to assess
88% managers: clearer, more useful insights.
Training time significantly reduced.
50%+ new reps reported higher confidence.
Nudging for Better Data
Nudging users to add basic content before meetings improved assessments, as even a little context helped the AI provide better evaluations, while keeping it optional ensured that busy sales reps were not burdened with extra work.
LEARNINGS
Key Learnings
Having stakeholders aligned early is crucial.
Human intervention is still necessary despite AI being advanced.
Simplicity wins over complexity for adoption.
This is just one story. See what else I’ve designed
There’s a lot more where that came from. Browse through my other projects to see how I tackle problems, experiment with ideas, and create meaningful experiences.
