Scorecard MCP
Scorecard MCP lets agents manage evaluation projects, testsets, testcases, systems, metrics, runs, records, scores, annotations, and documentation searches.
Verdict
Common use cases
- Audit support tickets flagged for coaching
- Compare agent scores across product lines
- Identify recurring customer pain points
- Draft feedback emails from low-scoring calls
- Summarize weekly QA trends for leadership
Integration
- Vendor
- Scorecard MCP
- Category
- other
- Auth
- OAUTH2
- Composio slug
scorecard_mcp
Tools
Setup
Setup guide
- 11. Open your Switchy workspace and navigate to Settings > Integrations. 2. Find Scorecard MCP in the directory and click Connect. 3. You'll be redirected to Scorecard's OAuth consent screen—log in with an account that has admin permissions. 4. Grant read access to conversations, scorecards, and team analytics when prompted. 5. Switchy confirms the connection and displays your Scorecard workspace name. 6. Open any Project and type '@Scorecard' followed by a request like 'show me yesterday's lowest-scoring tickets' to test the link. 7. If the MCP returns data, you're ready—if it errors, revisit Settings > Integrations to re-authorize or check that your Scorecard plan includes API access.
What teammates see: by default, memories from Scorecard MCP are scoped to the Project (PROJECT visibility) - you can mark any memory PRIVATE or share it ORG-wide.
Works well with
Top models
Compatibility data appears once enough Projects have used this MCP together with a given model.
How Switchy teams use it
Starter prompts
Weekly QA Summary
@Scorecard retrieve all scored conversations from the last seven days and summarize the three most common quality issues, then list agents who need coaching.
Coaching Opportunity Finder
@Scorecard find tickets scored below 70 in the past three days, group them by agent, and draft a brief coaching note for each person.
Product Feedback Rollup
@Scorecard review this month's transcripts tagged 'feature request' or 'bug report' and create a ranked list of the top five issues customers mentioned.
Agent Performance Snapshot
@Scorecard show me each agent's average score this month alongside the team median, then flag anyone more than ten points below the benchmark.
Rubric Compliance Check
@Scorecard pull the last twenty tickets and check whether agents followed our greeting protocol and offered a follow-up—list any that missed either step.
Example outputs
Illustrative - representative of the model's voice and quality, not literal recordings.
@Scorecard analyze our engineering velocity scorecards from the past 6 months and identify which metrics correlate most strongly with on-time delivery
Analysis of 26 weeks of engineering velocity data shows three strong correlations with on-time delivery: Code review turnaround time (r=-0.72, faster reviews → better delivery), Sprint planning accuracy (r=0.68, better estimates → better delivery), and Bug escape rate (r=-0.61, fewer escapes → better delivery). Interestingly, raw commit volume showed weak correlation (r=0.23), suggesting quantity of code changes matters less than process discipline.
This example pairs Scorecard data with AI reasoning to surface insights. Requires read access to historical scorecard data across multiple periods. The correlation analysis is illustrative—actual statistical rigor depends on your data quality and sample size. Best used for hypothesis generation rather than definitive conclusions.
Use-case deep-dives
When Scorecard MCP fits a 5-person sales pod tracking pipeline health
A five-person sales team runs weekly pipeline reviews where they score deals on likelihood, budget fit, and timeline. Scorecard MCP works here if the team already uses a scorecard platform and wants AI to pull scores into Switchy for cross-referencing with CRM notes or Slack threads. The OAuth2 flow means each rep authenticates once, then the AI can fetch their assigned deals without manual exports. The trade-off: if your scoring logic lives in spreadsheets or Notion tables instead of a dedicated scorecard tool, this MCP adds a layer you don't need—just use the Sheets or Notion MCP directly. If you're paying for Scorecard and want AI to surface scores alongside other context, this is the right connector.
How Scorecard MCP supports quarterly roadmap planning at 8-person product orgs
An eight-person product team scores feature requests on impact, effort, and strategic fit every quarter. Scorecard MCP lets the AI pull those scores into Switchy so PMs can ask 'show me high-impact, low-effort features from last scoring session' without leaving the workspace. OAuth2 means each PM's view respects their access level in the scorecard platform. The boundary: if your team scores fewer than 30 items per quarter, a shared Airtable or Notion board is faster to set up and the AI can read it just as well. Scorecard MCP pays off when you have 50-plus items, multiple scorers, and need version history or audit trails that a spreadsheet won't give you.
When Scorecard MCP makes sense for a 3-person CS team tracking account risk
A three-person customer success team monitors health scores across 80 accounts, updated weekly by their scorecard platform. Scorecard MCP connects that data to Switchy so the AI can answer 'which accounts dropped below 70 this week' or cross-reference scores with support ticket volume from Zendesk. OAuth2 keeps each CSM's view scoped to their book of business. The catch: if your health scores come from a BI dashboard or get manually updated in a CRM field, this MCP is overkill—use the CRM's native MCP instead. Scorecard MCP is the right call when your scoring logic is complex enough to warrant a standalone platform and you want AI to treat scores as first-class data alongside everything else your team references.
Frequently asked
What does Scorecard MCP do in Switchy?
Scorecard MCP connects your Scorecard account to Switchy's AI workspace, letting your team query and reference Scorecard data in conversations without switching apps. Since no specific tools are documented yet, expect read-access to scorecards, metrics, and reporting data that your OAuth token permits.
Do I need admin permissions to connect Scorecard MCP?
You'll need whatever OAuth scopes Scorecard requires for the data you want to access. Typically read-only integrations work with member-level accounts, but if you need to pull org-wide scorecards or sensitive metrics, your Scorecard admin may need to approve the connection or grant elevated permissions.
Can Scorecard MCP write data back to my scorecards?
That depends on which tools the MCP exposes and which OAuth scopes you grant. Most read-only integrations won't update scorecards. If you need write access—like logging new scores or updating metrics—check the MCP's tool list once it's populated, or use Scorecard's native API for those workflows.
Why use Scorecard MCP instead of just opening Scorecard in a browser?
The MCP brings Scorecard data into your team's AI conversations, so you can ask questions, pull metrics, and reference scorecards inline without context-switching. If you only check Scorecard occasionally, the browser works fine. If your team discusses performance data daily, the MCP saves time.
Does connecting Scorecard MCP count against my Switchy plan limits?
MCP connections themselves don't count as seats or users. However, API calls the MCP makes to Scorecard may count against your Scorecard plan's rate limits or usage quotas. Check your Scorecard billing if you're on a metered tier.