9 fascinating things you can do with Perdoo's MCP server

Key Takeaway: Perdoo's MCP server exposes 35 tools that let AI assistants read, draft, update, and automate your strategies, goals, and projects in Perdoo. Where most OKR tools with an MCP server give you 1 or 2 read-only tools, Perdoo lets you do many useful things such as: get on-demand AI Chief of Staff briefings, ask live strategy questions from your favorite AI, draft Strategic Pillars/OKRs/KPIs, get feedback on your drafts, build integrations that auto-update goals from other tools, generate board-ready reports, and nudge the people who need it. The result: your strategy execution processes finally become AI-native.

The Model Context Protocol (MCP) is reshaping how software connects to AI assistants. Instead of building bespoke plugins for each tool, an MCP server exposes a standardized set of capabilities that any MCP-compatible AI (ChatGPT, Claude, Gemini, etc.) can call directly. It's the closest thing the industry has to a universal AI bridge.

We built Perdoo's MCP server with a specific belief: strategy execution is one of the highest-value places to plug in AI. Not for slideware or generic Q&A, but for real, connected reasoning against your actual goals and Strategic Pillars. That belief pushed us to build what is, as far as we can tell, the most capable MCP server in the OKR space. It exposes 35 tools spanning reads, writes, and progress updates across every layer of the Perdoo app. Competitor MCP servers we've looked at typically expose 1 or 2 read-only tools. That difference matters, because it's the difference between an AI that can tell you about your OKRs and an AI that can help you execute strategy.
‍

9 fascinating things you can do with Perdoo’s MCP server

1. Your on-demand AI Chief of Staff

The most common way customers start is by turning their AI assistant into a Chief of Staff for strategy execution. A single natural-language prompt, and the AI pulls together a full status picture.

Try this prompt in Claude, ChatGPT, or Perplexity:

"Give me a Monday morning briefing on our Strategic Pillars. For each one, flag any OKRs or KPIs off track and any goals that haven't had a check-in in the last 10 days. End with the 3 highest-priority things I should follow up on this week."

The AI calls list_strategic_pillars, get_strategic_pillar for each, list_okrs filtered to the current timeframe, list_kpis, and list_updates, then synthesizes the whole thing into a readable summary. What used to take 30 minutes of clicking through dashboards now takes 30 seconds.

You can schedule this as a recurring task in Claude or ChatGPT and have the briefing arrive in your inbox or Slack every Monday. Different from a standard dashboard because the AI can interpret the data (spotting stalled Initiatives, flagging Objectives without check-ins) rather than just displaying it.

2. Live strategy Q&A

Because Perdoo works in ChatGPT, Claude, Gemini, and Perplexity, the AI you already use every day becomes a strategy-aware assistant. Ask ad-hoc questions and get answers grounded in your real data.

Example prompts customers regularly use:

• "Who leads our Engineering team and what KPIs do they own?" This pulls the team lead, their active KPIs, and current values.
• "Show me every OKR under our 'Expansion Strategy' Strategic Pillar with its current progress and lead." This cross-references pillars, OKRs, and team members.
• "Compare my personal OKR progress this quarter to the company average." This uses get_me plus filtered OKR queries.
• "Which Initiatives haven't had progress updates in the last two weeks?" This pulls stalled work across teams.

None of this requires custom prompt engineering. The MCP server handles the mapping between natural language and Perdoo's data structures. If you can ask a question in plain English, the AI can answer it against your real Strategy Map.

3. Draft Strategic Pillars, OKRs, and KPIs

Writing good OKRs is hard. Writing them from scratch in front of a blank Perdoo form is even harder. The MCP server lets you draft them in conversation with your AI, iterating until they're sharp, then push them to Perdoo as DRAFT goals for review.

Example prompt:

"We're setting Q3 OKRs for our Marketing team. Our Strategic Pillar is 'Establish thought leadership in strategy execution.' Draft 3 candidate Objectives with 2-4 measurable Key Results each. Make them ambitious but achievable, and make sure every Key Result measures an outcome, not an activity."

The AI drafts the OKRs in the conversation. You review, edit, and refine in natural language ("make KR 2 more specific," "change the target on KR 3 from 15% to 25%"). When you're happy, one more prompt does it: "push these to Perdoo as DRAFT OKRs for the Marketing team in Q3." The AI calls draft_okrs to create them, and your team lead reviews them in Perdoo before moving them to ACTIVE. If you want a refresher on what makes a great Objective vs. a great Key Result before you draft, our Ultimate Guide to OKR is a fast primer.

The same works for Strategic Pillars ("draft 4 candidate Pillars for our 2027 strategy based on this positioning document") and for KPIs ("draft the 8 KPIs we should track for the Customer Success function"). It's the fastest way we've seen to go from a blank page to a coherent set of goals.

4. Get feedback on your draft Pillars, OKRs, and KPIs

This is the one we didn't expect customers to love as much as they do. Instead of asking the AI to write your goals, ask it to critique the goals you've already drafted.

Example prompt:

"Pull up all my team's Q3 OKRs currently in DRAFT stage. For each one, tell me: (1) is the Objective genuinely ambitious or is it a safe restatement of existing work, (2) do the Key Results measure outcomes or activities, (3) does each Key Result have a specific numeric target, and (4) if you were the CEO reviewing these, what feedback would you give?"

The AI reads your DRAFT OKRs, evaluates them against best-practice principles (which are baked into how Perdoo describes goal quality), and gives you specific, constructive feedback. Customers use this in two ways:

• Solo review, before showing the OKRs to their manager or team, to catch weaknesses early.
• Coaching for less experienced OKR authors, especially first-time OKR Champions who don't yet have deep pattern recognition on what makes a great OKR.

You can ask the AI to check for the common OKR dogmas that trip teams up: are you confusing Key Results with Initiatives, is your Objective a wish list, are your KRs stretch or committed, etc. The feedback quality is genuinely useful, because the AI is reasoning against goals that are structurally clean (parent Strategic Pillar, explicit KRs, dated timeframe), not against vague strategic ambitions.

5. Build integrations that automatically update your goals

This is where the MCP server becomes truly powerful, and where our customers spend the most creative energy. Because the MCP server supports write operations (update_goal_progress in particular), you can use your AI as a bridge between all your other apps and Perdoo.

The classic example: keeping your revenue KPIs live.

Set up a weekly scheduled prompt in Claude or ChatGPT:

"Every Monday at 9am, pull our current MRR from Stripe. Find the KPI named 'MRR' in Perdoo. If the current value in Perdoo is different from Stripe's live number, log an update in Perdoo with the new value and a note saying 'auto-updated from Stripe on [date].'"

No custom code. No Zapier fee for the connector. Just a prompt, an AI assistant with access to both MCP servers, and a schedule. The AI does the mapping every week.

The same pattern works for:

• HubSpot or Salesforce → pipeline and revenue KPIs. Pull open pipeline weekly, update the "New Pipeline Generated" KPI in Perdoo.
• Google Analytics → website conversion Key Results. Pull last week's conversion rate, update the KR "Increase pricing page conversion from 2% to 4%" with the current value.
• GitHub or Linear → engineering velocity metrics. Pull deployment frequency or cycle time, update the corresponding KPI or KR.
• NPS or CSAT tools → customer satisfaction KPIs. Pull the latest score, update the KPI.
• Data warehouse tables → any custom metric. If it's queryable and you have an MCP for it, it can update Perdoo automatically.

The pattern replaces a class of integrations that OKR software has historically needed to build one at a time. With MCP, the integration is the prompt. And because every update includes the source and date in the note, you get automatic audit trail for free.

6. Generate board-ready reports on demand

Building the quarterly board deck is one of the most tedious tasks in a strategy execution program. It usually involves 3 or 4 hours of copy-pasting numbers from Perdoo into PowerPoint, formatting slides, and re-writing status commentary. With the MCP server, you can compress it to minutes.

Example prompt:

"Pull the current state of every Strategic Pillar, its Company OKRs with progress percentages, and its KPIs with current vs. target values. Then generate a 3-slide Google Sheets board deck with one slide per Strategic Pillar, followed by an executive summary slide with the 3 biggest wins and the 3 biggest risks. Use our standard board template."

The AI reads Perdoo via MCP, generates the deck via a document-generation tool (Claude has one built in; several other MCP tools exist), and hands you a formatted file. You review, edit any commentary, and ship it.

Some customers do this monthly for internal leadership reviews. Some do it quarterly for the board. Either way, the tedious part goes away.

7. Draft your weekly User Review from your actual work, not from memory

Every Friday afternoon, most people sit down to write their weekly update and struggle to remember what actually happened. The MCP server can fix that: the AI reads your activity for the week from all your various tools, pulls the check-ins you logged, how your goals progressed, then drafts your User Review with concrete evidence baked in.

Try this prompt on Friday at 4pm:

"Draft my User Review for this week. Pull everything that I worked on this week, including all the meetings I’ve had, and how my goals in Perdoo progressed. Then write my Review. Structure it as three sections: Wins, Challenges, and Blockers. Keep it conversational and honest — flag anything that's genuinely off track rather than sugarcoating it. When I approve, post it as my User Review in Perdoo."

The AI calls get_me, list_updates filtered to the current week, list_okrs and list_kpis to see what changed, then drafts a review in conversation. You edit it in natural language ("cut the second win, that's not important" or "add the customer conversation blocker from Tuesday"), and once you're happy, it writes the User Review directly to Perdoo. What used to be 20 minutes of staring at a blank text box becomes 3 minutes of review-and-approve.

The interesting side effect: the reviews get better. Because the AI is working from real data (numeric progress, actual check-in notes, real timestamps) and can access all your tools, you get more specific wins and more honest blockers. Vague weekly updates tend to be a sign that someone is writing from memory. This use case fixes that at the source.

8. Auto-generated Monthly Business Reviews for the whole company

MBRs and QBRs are one of the most valuable rituals in a well-run strategy execution program, and one of the most tedious to prepare. A team lead can easily spend 1 hour pulling numbers, writing commentary, and formatting a summary before the review meeting. The MCP server compresses that to minutes, and now with Team Review writes, the final artifact lands directly in Perdoo where everyone can find it.

On the last workday of each month, run this scheduled prompt:

"Generate a Monthly Business Review for the Marketing team. Pull every Marketing OKR active this month with its progress, every Marketing KPI with current vs. target values, and every User Review the team's members wrote this month. Then write the Team Review in this structure: (1) What we achieved this month — the 3-5 biggest wins with numbers, (2) Where we're behind — the OKRs or KPIs that are off track and why, (3) What we're doing next month — the top priorities pulled from the individual User Reviews, (4) Help we need from the rest of the company. Post it as this month's Team Review in Perdoo."

The AI reads every relevant layer (Team OKRs, KPIs, individual User Reviews, progress history), synthesizes across them, and posts the Team Review directly to Perdoo. The team lead reviews and edits before it goes live, but the heavy lifting is done.

Two things this unlocks that manual MBRs don't: Consistency across teams. Because every Team Review is generated from the same prompt structure, MBRs look and feel comparable across Marketing, Sales, Engineering, and Product. That makes them dramatically easier for executives to scan and compare.

A living record. The Team Reviews become the searchable historical archive of the company's execution story — what worked, what didn't, and why — accessible to anyone via the MCP server going forward. Six months later, someone can ask their AI "how did Marketing's pipeline generation KPI trend over the last two quarters, and what did they say about it in their Team Reviews?" and get a real answer.

9. Automate weekly nudges to the people who need them

The MCP server can also drive the maintenance work that keeps an OKR program alive. This is the "AI as OKR Champion assistant" use case.

Example scheduled prompt (every Friday morning):

"Find every Objective and KPI that hasn't had a check-in in the last 10 days. Look up each lead. For each person, draft a short, friendly Slack message reminding them to check in this week, including a direct link to their goal in Perdoo. Send the messages via the Slack MCP."

That's an OKR Champion's most tedious weekly task, automated. No more chasing people. No more wondering which teams have gone quiet. The AI reads Perdoo, checks last update dates, personalizes the outreach, and delivers it, all triggered by a single scheduled prompt.

The same pattern works for:

• New-cycle kickoff reminders ("It's the start of Q4. Message every team lead to remind them their Q4 OKRs need to be drafted by Friday.")
• Score-out reminders ("Q3 ends in 2 weeks. Message the leads of every ACTIVE OKR that hasn't been closed yet.")
• Executive escalation ("Every Monday, send me a Slack DM with any OKR that's been at 'off track' status for more than 3 consecutive check-ins.")

Each of these used to require either a Champion doing manual work or a custom automation someone had to build. Now it's a paragraph of instructions.

How to set up the MCP server

The MCP server sits at https://mcp.perdoo.com and connects your Perdoo account to any MCP-compatible AI tool. Step-by-step instructions are available here. Setup only takes a minute.

You’ll also find Perdoo in ChatGPT > Plugins or in the Claude Directory.

We're actively expanding the MCP server's capabilities. If there's a workflow you'd like to see supported, tell us. A large fraction of the tools currently exposed came from customer requests over the past few months.

Why the depth of the MCP server matters

Most OKR platforms with an MCP server treat it as a marketing checkbox. They expose one or two read-only tools, use it for a headline demo, and call it done. That's fine if you want AI-generated briefings and nothing else. It's not enough if you want AI to actually help you run your strategy execution program.

The reason Perdoo's MCP server supports 35 tools including full write operations, DRAFT-to-ACTIVE stage transitions, custom-field access, and progress history is because that's what customers doing serious work with AI need. You cannot automate progress updates without write access. You cannot draft OKRs in conversation without a DRAFT stage that a human then reviews. You cannot generate board reports without progress history. The depth of the MCP server is the depth of what you can do.

If you want to move your OKR program from "we track goals" to "we have AI running alongside our strategy execution," Perdoo is built for that. Start for free or request a demo, and if you're already a customer, connect your MCP client today and try Use Case 1 this Monday morning.