Why use AI context instead of matching the meeting title?

Reps name meetings inconsistently. AI context evaluates what was said on the call, so 'Quick sync with Müller' gets scored as a discovery call if that's what it was — no dependency on the rep typing the right title.

What happens if two rules match the same meeting?

The rule higher in the list wins. AI Coach evaluates top-down and stops at the first match. Order your rules from most specific to most generic.

Can I assign more than one scorecard to a single meeting?

One rule applies one scorecard. If you need two perspectives on the same call — discovery quality and methodology adherence, for example — build two separate scorecards covering both angles and pick the most relevant one per rule.

Why are all my meetings scoring 2 out of 5?

Your scorecard criteria are probably too strict. Loosen the language, test on five recent calls, and iterate.

Do email-specific rules work the same as call rules?

Same hierarchical order, same logic. They match on email content and apply scorecards built for written follow-ups instead of calls.

Can reps see which scorecard was applied and why?

Yes. Each scored meeting shows the scorecard used, the score, and the rationale per criterion. Reps see their own coaching results.

Does this work for non-English calls?

Demodesk AI Coach supports 98 languages. Write your AI context in your team's working language and it will match meetings in that language.

What if my team has different motions per segment?

Create separate rules per segment. Reference segment signals in your AI context — company size, procurement complexity, multi-stakeholder dynamics — to route the right scorecard to the right meeting.

May 21, 2026·6 min read

How to set up AI context-based scorecard automation rules in Demodesk AI Coach

Configure AI Coach automation rules to assign the right scorecard to every meeting using AI context, not just meeting titles. Step-by-step guide.

Veronika WaxFounder & CEO

What and why

This guide shows you how to configure automation rules in Demodesk AI Coach so the right scorecard is applied to every meeting automatically — based on what was actually discussed, not the meeting title. Instead of rigid title matching, AI context evaluates the conversation and picks the matching scorecard. That's how teams scale AI sales coaching from a few power users to the whole team without manual tagging.

Who this is for

Sales managers, RevOps, and team leads running multiple meeting types — discovery, demo, closing, customer success — who want every call scored against the right framework automatically.

Prerequisites

Demodesk Coaching & AI plan (any tier)
Admin or manager role with access to Agents → Automations
At least two scorecards already created (e.g., one for Discovery, one for Demo). Build them first under Agents → Scorecards if you haven't yet
A clear picture of which meeting types you run and how they differ in content, not just title

Steps

1. Navigate to AI Coach automations

Open the Demodesk app and go to Agents in the top nav, then select the Automations sub-tab. Find the AI Coach section. This is where you control which scorecard runs against which meeting and in what order.

2. Turn on automatic evaluation

Enable the automatic evaluation toggle for AI Coach. Every recorded meeting will then be evaluated against the rules you define below. No manual scorecard assignment, no missed calls.

3. Understand rule order

Rules run top-down. AI Coach checks each rule in order and applies the first match. A generic “Discovery” rule sitting above a specific “Enterprise Discovery” rule will catch the call first — the specific scorecard never runs.

Put your most specific rules at the top. Your most generic rule goes last as the catch-all.

4. Create your first rule

Click Add rule. Each rule needs two things:

A condition— what triggers this rule
An action— which scorecard to apply

Start with the meeting type you run most often. Name the rule clearly — “Discovery Call” or “Demo — SMB” — so it's recognizable later.

5. Use AI context instead of meeting titles

Inside the rule condition, choose AI context instead of title matching.

Meeting titles are unreliable. Reps name calls “Quick chat with Anna” or “Follow-up — Müller GmbH” — title-based rules miss both. AI context looks at what was discussed and matches the rule to the content.

Write your AI context in plain language:

Discovery rule:“The meeting is a first conversation with a new prospect. The rep is qualifying the company, asking about their current process, pain points, team size, and timeline. No product demo was given.”
Demo rule:“The rep is presenting the Demodesk product. The screen is shared, specific features are walked through, and the prospect asks product questions.”
Closing rule:“Pricing, contract terms, procurement, legal, or signing timelines are discussed. The deal is in a late stage.”

The more concrete the description, the more accurately AI Coach assigns the right scorecard.

6. Create email-specific rules

If you score follow-up emails as part of coaching, add a separate rule for those. Set the condition to match email content and apply your follow-up email scorecard. Written follow-ups get the same coaching rigor as live calls — without one rule interfering with the other.

7. Assign the right scorecard to each rule

Pick the scorecard from the dropdown. One scorecard per rule. If you want two perspectives on the same call, create two separate rules with distinct conditions rather than stacking scorecards on one rule.

8. Test with real recordings

Check three or four recent meetings of different types and confirm the right scorecard was applied. A discovery call scored against your demo scorecard means the AI context for one of those rules needs tightening — or the order is wrong.

9. Calibrate if scores run too low or too high

A common pattern after setup: every meeting scores 2 out of 5. That's almost always the scorecard being too strict, not the AI being wrong. Loosen the criteria, test on five recent calls, and iterate. The goal is coaching signal, not a penalty system.

Tips

Start with three rules, not ten. Discovery, Demo, and one catch-all. Add more only when you see real misclassifications.
Re-order weekly for the first month. As edge cases appear, move specific rules higher.
Pair AI context with CRM stage when you can.A “Closing” rule referencing both conversation content and deal stage is harder to misfire.
Fix the rule, not the rep. When a meeting is scored against the wrong framework, the rule is the problem. The rule is the system.
Review your AI context language every quarter.New products, new segments, new methodology — the descriptions should keep up.

Related skills and agents

AI Coach — runs scorecards and delivers instant post-call feedback. Product page.
AI Crew — for custom coaching workflows beyond scorecards. AI Crew.
Marketplace skills— pre-built coaching frameworks including MEDDIC, BANT, and Challenger. Browse at marketplace.demodesk.ai/agents.