Analyse a Notion Workspace Without Building Another Rollup

Your team's second brain is in Notion. Project tracker, OKRs page, content calendar, hiring pipeline, customer feedback log, the wiki of how-things-work. Every meaningful decision the team makes flows through one of those databases.

The next quarterly review is in two weeks. Someone has to look at all of it and say what's working.

That someone usually screenshots six dashboards into a slide deck and calls it analysis. Notion has the data. Notion's filter-and-group views weren't built to answer "of the OKRs marked at-risk, which ones share the same owner, and which goals are blocking which?"

Why Notion is a great workspace and an awkward analytics tool

Notion is where most teams keep their operational docs. Databases inside pages, relations between them, and views you can configure for the team. Everyone who joins the company finds their footing in it within a week.

That same model is the analysis ceiling.

• Views are saved filters, not queries. You can filter your Projects database by Status = "In Progress" and Owner = a specific person. You cannot ask "which owners have the highest in-progress count relative to their team size" without exporting.
• Relations are one-way thinking. You can follow a relation from a project to its tasks. You cannot easily ask "which projects have tasks blocked for more than two weeks across the entire workspace."
• Rollups stop at one hop. A rollup can summarise a relation. It can't summarise across two relations, so any question that joins Projects → Tasks → Assignees → Teams requires either a formula chain that breaks or a new database you build by hand.
• Time-aware analysis isn't built in. "How long does an OKR sit at the same percentage before someone updates it?" The data is there in last-edited timestamps. The aggregation isn't.
• Cross-database joins are intentional, not exploratory. Notion's databases are siloed unless you've pre-built a relation. The analysis question rarely matches a relation you set up six months ago.

The traditional fix is to export every database to CSV, paste them into a spreadsheet, build the lookups, and present a slide. Or wire Notion to a warehouse with a paid sync, then to a BI tool, then hire someone to build the dashboard. Most operators have neither the patience nor the budget. So the quarterly review becomes a vibes call.

How to connect Notion to an AI analyst

Sign in to Notion. Pick the workspace, choose which pages and databases to share. Anna can read what you share — never change it. Done. About a minute.

Once connected, Anna can read pages, blocks, databases, properties, comments, users, and relations. She uses the same Notion API the rest of your tooling does — without rate limits you'd notice in a review session.

You ask the question. She walks the relations, runs the analysis, and writes the answer.

A 30-minute quarterly review in six questions

The questions an operator, founder, or head of ops actually asks before a review.

1. Pull the OKR database and audit the at-risk ones

"From the OKRs database, surface every objective marked 'At risk' or below 50% progress. Group by owner and by quarter. Show me the goals at risk with the most days since their last update."

Anna walks the database, reads the relations to owners and key results, and returns the audit.

This is the quarterly review opener every operator wants and doesn't have time to build. Anna writes it in a sentence.

2. Audit the project board

"From the Projects database, surface every project that's been 'In Progress' for more than 30 days, has no assigned owner, or has tasks with overdue due dates. Show me the worst offenders by team."

The project board has been telling you something is stuck for weeks. The interface buries it under a "Due this week" filter that excludes everything you actually need to see. Anna writes the audit.

3. Run the content calendar honestly

"From the Content Calendar database, calculate posts briefed per week and posts published per week for the last 8 weeks. Show me the gap."

Every content team thinks the problem is ideas. The data usually says the problem is the writer-to-editor handoff. Two minutes with Anna; the rest of the quarter you spend fixing the right thing.

4. Cross-database join, no relation needed

"Take the Hiring database, the Headcount Plan database, and the Active Engineers database. Surface every role on the headcount plan that has no candidate in the hiring pipeline, and every role that's filled but isn't on the plan."

This is the question your head of people gets asked in every leadership meeting and dodges with a verbal answer. The data is in three databases that don't share a relation. Anna does the join without you having to build one.

5. The "what's actually moving" rollup

"Across the entire workspace, find every database row that hasn't been edited in 60 days but still has an active status. I want a list of zombie work — things that look alive but aren't."

Every Notion workspace accumulates zombie projects. Anna writes the kill list.

6. Customer feedback themes

If you log customer feedback in a database — meeting notes, NPS responses, support escalations:

"Pull the Customer Feedback database. For every entry in the last 90 days, classify it by theme: product, pricing, onboarding, support, integration. Then count how often each theme appears. Highlight any theme that's grown by more than 25% quarter-over-quarter."

This is the analysis your product team wants every Monday morning. The database has been collecting feedback for a year. Nobody has read it cover-to-cover because doing so takes a day. Anna reads it and writes the themes.

Joining Notion to the rest of your stack

The questions that span Notion and your other tools are usually the highest-impact ones. If those tools are connected:

• Notion + HubSpot. Match accounts in your CRM against the customer database in Notion. Surface accounts with no notes from the last 60 days.
• Notion + Stripe. Compare paying customers against your "active engagements" database. Surface paying customers with no active engagement, or active engagements with no paid status.
• Notion + GitHub. Match items in the engineering roadmap database against open PRs and shipped commits. Surface roadmap items with no engineering activity in the last sprint.
• Notion + Google Calendar. Match weekly leadership meeting agendas against the OKR database. Surface objectives that haven't been discussed in three consecutive meetings.

These are the joins that turn Notion from a documentation tool into a decision-making system.

What this replaces

• The Friday before the quarterly review, when somebody screenshots Notion dashboards into a slide deck
• The "we should really build a Notion-to-Looker pipeline" idea that's been on the roadmap for two years
• The Zapier-and-Coda contraption that pulls Notion into a spreadsheet that nobody updates
• The half-hour at the start of every leadership meeting where everyone reorients themselves on the workspace
• The hire you were going to make for an analyst who would mostly read Notion databases

Questions you can also ask Anna

The review above is the foundation. Same toolset, same workspace:

• "From the Hiring database, surface every role with more than 5 candidates and no scheduled interview in the next 14 days."
• "Audit the Wiki. Which pages haven't been edited in the last year? Which have no inbound links from elsewhere in the workspace?"
• "For every customer logged in the CRM database, fetch their last activity date and group by 30-day buckets. Where's the engagement cliff?"
• "In the Goals database, find every key result tied to an objective that's been marked complete. Flag any KR where the metric isn't actually met."
• "Pull the Meeting Notes database. Surface every action item assigned more than 14 days ago and not marked done."

Each is one sentence. Each used to be a research project on its own.

The take-home

Notion is where your operations document themselves. The story of how your business is doing this quarter is already in there — across OKRs, projects, content, hiring, and feedback databases.

Connect the workspace. Ask the question. Anna walks the relations, joins to your other tools if you want, and writes the brief.

The quarterly review just got an honest first draft.

FAQ: analysing a Notion workspace with Anna

Do I need to build rollup chains or formula properties in Notion first?

No. Anna reads the raw databases and follows relation properties directly. Rollups are a workaround for not having a real analyst — connect the workspace and Anna walks the relations herself.

Which Notion permissions does Anna need?

Anna can read the pages and databases you share — never change them. You select what to share through Notion's standard integration permission flow; Anna never sees pages you have not granted.

How does Anna handle Notion's mixed page/block structure?

She reads structured database properties for analysis and uses page body content (rich text, comments) for context when a question needs it. The hybrid is the point — your operations live in both, and Anna treats them as one source.

Can Anna follow relation properties across multiple databases?

Yes — that is the unlock. "OKRs linked to Projects linked to Tasks linked to Sprint" is one prompt, not three rollup fields. Anna traverses the relation graph and runs the math at the leaf.

Can I join a Notion database to Stripe, HubSpot, or Shopify?

Yes — connect both sides and Anna joins on a common field (deal ID, customer email, project code). "Match the Notion Customers database to Stripe revenue and rank by NRR" is one prompt.

What about Notion AI's built-in Q&A?

Notion AI is good at retrieval and summarisation inside a single workspace. Anna is built for analytical questions — aggregation, segmentation, statistical comparison, cross-source joins. Use Notion AI to find a page; use Anna to compute the answer the page does not contain.

Will Anna catch data quality problems (empty cells, broken relations)?

Yes — she flags databases where >25% of records are missing key fields, relations pointing to deleted pages, and date fields with implausible values. You see what the workspace is hiding before the analysis runs.

Does this work on Notion's free plan or only paid?

The Notion API is available on all paid plans. Free personal workspaces have limited API access. Most operator and team workspaces on the Plus or Business tier already have what Anna needs.