How to create a custom MEDDICC scorecard in Demodesk AI Coach

What and why

A MEDDICC scorecard in Demodesk AI Coach scores every discovery call against the seven pillars — Metrics, Economic Buyer, Decision Criteria, Decision Process, Identify Pain, Champion, Competition — without a manager listening to a single call. Build it once, set the automation rules, and AI Coach scores every matching call from then on.

This guide covers creating the scorecard, customizing questions and scoring criteria, and routing it to the right call types.

Who this is for

Sales managers, RevOps, and enablement leads running MEDDICC (or any qualification methodology) who want consistent AI sales coaching across the team instead of ad-hoc 1:1s.

Prerequisites

• Demodesk Coaching & AI plan with admin access to the Agents tab
• At least one recorded call in your account so you can test the scorecard against real data
• A clear definition of how your team applies MEDDICC — especially what “good” looks like for Metrics and Economic Buyer
• AI Coach automation enabled in your workspace

Steps

1. Navigate to Agents > Scorecards

Open Demodesk and click Agents in the top navigation. Select the Scorecards sub-tab. This is the configuration hub for every scorecard AI Coach uses to evaluate calls.

You'll see any existing scorecards — default templates plus anything your team has built.

2. Click “New skill” and choose “Create manually”

Click New skill in the top right. Demodesk gives you two paths:

• Create from template — start from a pre-built MEDDICC, BANT, SPIN, or Challenger scorecard
• Create manually — start from a blank scorecard

Choose Create manually for full control over questions and weighting. For a faster starting point, pick the MEDDICC template and skip to step 4 to edit it.

3. Add your questions

For a MEDDICC scorecard, add one question per pillar. Two options:

• Use the question library — Demodesk ships pre-written questions for MEDDICC, SPIN, BANT, and other common frameworks. Click Add from library, filter by MEDDICC, and add what you need.
• Write custom questions — Click Add question and write your own. Example for the Champion pillar: “Did the rep identify a champion who has internal influence and a personal reason to drive the deal forward?”

Most teams start with library questions and replace two or three with custom ones that reflect their specific ICP.

4. Add scoring criteria for each question

For every question, define what counts as a pass, a partial pass, and a fail. Demodesk uses these criteria to score the call.

Example for Metrics:

• Pass — Rep quantified business impact in concrete numbers (revenue, cost, hours, headcount)
• Partial — Rep discussed impact but didn't get specific numbers
• Fail — No discussion of business impact

Be specific. Vague criteria like “asked good questions” produce vague scores. AI Coach is fine-tuned on 10M+ real sales conversations, but it scores against your definitions, not a generic rubric.

5. Add custom context at the end

Scroll to the bottom of the scorecard and add a Custom context block. This is where you tell the AI how your team applies MEDDICC.

Include:

• Your ICP (e.g., “We sell to VP Sales at 50–2,000 FTE European B2B SaaS companies”)
• What “Economic Buyer” means in your deals (e.g., “Usually CRO or VP Sales — not the champion”)
• Common champion patterns (“Our champions are typically RevOps or sales managers, not individual reps”)
• Methodology nuances (“We weight Decision Process more heavily than Decision Criteria for enterprise deals”)

This context is the single biggest lever for scoring accuracy. Skip it and you get generic MEDDICC scoring instead of MEDDICC as your team runs it.

6. Activate the scorecard and save

Toggle the scorecard to Active and click Save. The scorecard is now available — but it won't score anything yet. AI Coach needs to know which calls to apply it to. That's the next step.

7. Set up automation rules under AI Coach

Go to Agents > Automations and open the AI Coach automation. Create a rule that maps your MEDDICC scorecard to the right call types.

Common setup for a discovery-focused MEDDICC scorecard:

• Trigger — Call ends and transcript is ready
• Condition — Meeting type is “Discovery” or call duration is over 20 minutes
• Action — Apply MEDDICC scorecard

You can route different scorecards to different call types — MEDDICC for discovery, a separate scorecard for demos, a closing scorecard for late-stage calls. Most teams start with one scorecard on first discovery meetings, then expand.

Save the rule. From the next call onwards, AI Coach scores every matching call automatically.

Tips

• Start with 5–7 questions, not 15. A focused scorecard produces sharper coaching. Add more later.
• Test against 5 past calls before rolling out. Apply the scorecard to recordings where you already know the outcome. If the scores don't match your read, tighten the criteria.
• Share the first round of scores with reps before using them in 1:1s. Reps adopt AI coaching faster when they see the rubric and can push back on it.
• Use the custom context block aggressively. It's the single biggest lever for scoring accuracy.
• Review scorecards quarterly. Methodologies drift. So does your ICP. Update when either changes.

Related skills and agents

• AI Coach product page — full overview of coaching capabilities
• AI Sales Assistant — pairs with AI Coach to handle recording, transcription, and follow-ups
• Marketplace — 158+ skills including MEDDICC, SPIN, and Challenger playbooks
• Solutions for Sales Managers — how teams scale coaching from 1:10 to 1:50