MCP Recipes
These are starting patterns for using ClarityLoop MCP with AI agents, reports, dashboards, and internal people workflows.
If you are using ChatGPT, Claude, Codex, Claude Code, or another MCP-connected client, copy a prompt starter and adapt the placeholders. If you are building an app or dashboard, use the helpful context line to decide which ClarityLoop data to retrieve.
MCP calls run as the connected ClarityLoop user. A manager, workspace owner, and individual contributor may see different results for the same recipe because their ClarityLoop permissions are different.
How to Use These Recipes
Work in this order:
- Start with the person asking the question.
- Identify the person, team, meeting, survey, or segment they care about.
- Pull only the context needed for that task.
- Keep evidence separate from interpretation.
- Ask for confirmation before creating or updating anything in ClarityLoop.
For dashboards and scheduled reports, save the date range, filters, connected user, workspace, tool names, and extraction time. That makes the output easier to explain later.
Manager 1:1 Prep
Help a manager walk into a 1:1 with a focused agenda, useful questions, and enough recent context to avoid starting from memory alone.
Helpful context: direct reports, upcoming or previous meetings, goals, growth signals, received feedback, work context, and notes. With confirmation, the agent can also add agenda items, action items, or meeting notes.
Prompt starter:
Prepare my next 1:1 with [person].
Use recent goals, growth signals, feedback, work context, notes, and the previous meeting if available.
Return:
- 3-5 agenda topics
- evidence behind each topic
- questions I can ask
- follow-up actions from previous meetings
- anything sensitive I should handle carefully
Do not add agenda items or notes until I confirm.
What good looks like:
- agenda topics grounded in actual signals
- questions that help the manager coach, not jump to conclusions
- previous actions carried forward
- proposed agenda items ready to add after approval
Review Prep from ClarityLoop Signals
Build a grounded brief for a review conversation, career discussion, or promotion discussion. This uses ClarityLoop signals such as feedback, goals, growth signals, work context, notes, career framework expectations, and company values.
MCP does not read review-cycle submissions or calibration records.
Prompt starter:
Create a review prep brief for [person] for [date range].
Use feedback, goals, growth signals, work context, notes, career framework expectations, and company values.
Return:
- recurring strengths
- recurring growth opportunities
- goal progress and evidence
- examples linked to work context or feedback
- relevant competencies or level expectations
- open questions for the manager
What good looks like:
- themes supported by examples
- clear separation between evidence and interpretation
- language that maps to the person's role or level where useful
- open questions where more manager judgment is needed
Avoid turning one sentiment score, activity count, or isolated feedback item into a performance conclusion.
Leadership Health Snapshot
Give leaders a readable view of how a team, department, or segment is doing across ClarityLoop signals.
Helpful context: people filters, people search, Leadership Insights, and insight criteria queries. For follow-up on specific people, the agent can also use goals, feedback, and growth signals where access allows.
Prompt starter:
Create a leadership health snapshot for [team or segment] over [date range].
Use Leadership Insights for sentiment, goal progress, platform activity, focus areas, and value distribution.
Call out:
- what looks healthy
- what may need attention
- recurring focus areas
- people or cohorts that may need follow-up
- what evidence supports each point
Keep this as a coaching and operating readout, not an employment decision.
What good looks like:
- trends and clusters, not a ranked list of people
- focus areas leaders can act on
- people returned by criteria only when follow-up is needed
- notes where data is missing or too sparse to interpret confidently
Find People Who Need Follow-Up
Use Leadership Insights to find people who match a defined set of signals, then turn the result into a thoughtful follow-up workflow.
Helpful context: people filters, people search, insight criteria queries, and Leadership Insights.
Prompt starter:
Find people in [team or segment] who may need follow-up.
Use criteria such as:
- low sentiment
- low goal progress
- low 1:1 follow-through
- few completed actions
- high growth-opportunity feedback
Return the matched people, the criteria they matched, and a suggested follow-up path.
Do not present this as a risk score or final judgment.
What good looks like:
- a clear cohort and the criteria behind it
- the signal behind each match
- a next step such as manager check-in, goal check-in, feedback follow-up, or 1:1 agenda item
Survey Follow-Up
Turn survey results into a practical readout leaders can understand and act on.
Helpful context: survey list, survey insights, people filters, leader filters, and people-science guidance from knowledge search.
Prompt starter:
Analyze the latest results for [survey].
Return:
- participation and response-rate context
- strongest themes
- areas that improved or declined compared with previous sends
- questions where sentiment or distribution needs attention
- 3 practical follow-up actions for leaders
- anything that should remain aggregated because of anonymity
What good looks like:
- participation context before interpretation
- themes tied back to survey questions
- follow-up actions leaders can actually take
- anonymity preserved in small groups
Growth Plan Builder
Help a person or manager turn feedback, goals, and role expectations into a focused development plan.
Helpful context: growth signals, received feedback, feedback details, goals, career framework, level expectations, company values, and people-science guidance. With confirmation, the agent can create goals, add goal items, or link goals to supporting records.
Prompt starter:
Build a growth plan for [person].
Use growth signals, feedback, current goals, role expectations, level expectations, and company values.
Return:
- 2-3 growth themes
- evidence for each theme
- suggested goals or actions
- skills or behaviours to practise
- how progress could be reviewed in 30-60 days
Do not create goals until I approve them.
What good looks like:
- a small number of meaningful themes
- goals connected to real feedback and role expectations
- actions specific enough to revisit
- confidence notes where evidence is thin
Feedback Collection Tracker
Help someone see which feedback requests are moving, which ones are still waiting, and what might need a nudge.
Helpful context: incoming feedback requests, sent feedback requests, and received feedback. With confirmation, the agent can send a new feedback request.
Prompt starter:
Summarize feedback request activity for [person, date range, or feedback cycle].
Show:
- requests I sent
- requests assigned to me
- response counts and outstanding responses
- received feedback connected to those requests
- suggested reminders or new requests
Do not send any new requests until I confirm.
What good looks like:
- request status and response progress
- gaps in feedback coverage
- suggested nudges without automatically sending them
Incoming feedback requests are the connected user's inbox, not a workspace-wide request queue.
Values and Culture Readout
Show how company values are appearing in feedback and people signals.
Helpful context: company values, Leadership Insights, value distribution, received feedback, feedback details, and people search.
Prompt starter:
Create a values readout for [team or segment] over [date range].
Use company values, value distribution, and value-tagged feedback.
Return:
- values showing up most often
- values showing up least often
- examples where visible
- possible coaching or recognition opportunities
- questions leaders should ask before acting
What good looks like:
- value patterns by team or segment
- examples where the connected user has access
- recognition opportunities and coaching themes
- care around low-volume or anecdotal evidence
Personal Reflection Assistant
Help an individual prepare for a self-review, career conversation, or development check-in in their own voice.
Helpful context: the connected user's notes, work context, received feedback, given feedback, growth signals, goals, career framework, and level expectations.
Prompt starter:
Help me prepare for my self-review.
Use my notes, recent work context, feedback, growth signals, goals, and role expectations.
Return:
- achievements with evidence
- strengths I should mention
- growth areas I can own constructively
- examples I may want to discuss
- draft language I can edit in my own voice
What good looks like:
- evidence the person can recognize and trust
- balanced strengths and growth areas
- draft language that feels usable, not over-polished
Building Good AI Workflows
- Start with the user's actual question.
- Prefer date ranges and explicit filters over broad searches.
- Show names and plain-language labels to users; keep IDs in the background for follow-up calls.
- Separate evidence, interpretation, and recommendation.
- Ask before writing feedback, goals, meetings, notes, agenda items, or action items.
- Link back to source records where possible.
- Use aggregated views for broad leadership reporting, and individual views for coaching or follow-up where access allows.
For field details, filters, return shapes, and access notes, see the MCP Tools Reference.