They bought Attio because it was the most secure CRM they tested. Then the stakeholder list stayed in a Google Sheet.
The Centre for Long-Term Resilience is an independent UK think tank working on extreme risks, including AI and biosecurity. Its AI Unit had picked Attio as the most secure CRM it tested, but the unit's stakeholders still lived in a Google Sheet. We built a Workstreams model around how the unit works, migrated the sheet with no email addresses to match on, and made Attio the one place the team tracks who it knows and when it last spoke to them.
Fixed-price build
Engagement
7-person AI Unit
Team
~200
Stakeholders migrated
7
Lists built
4
Automations
The problem.
CLTR came to us through the Attio Expert Directory. The whole organisation, about 20 people, already had Attio seats. The AI Unit had chosen it because it was the most secure CRM they tested, which matters when your stakeholders are in government.
But the unit did not work in it. Its master dissemination list, about 200 stakeholders, lived in a Google Sheet. The two stakeholder lists already in Attio were out of date and unused. The only thing reaching Attio was meeting notes, posted from a Slack channel onto a card that connected to nothing else.
The sheet held no email addresses, on purpose, for security. So every stakeholder had to be matched by name and organisation against the people Attio had already created from Gmail and Calendar sync. That was the hardest part of the job.
The workspace was shared with other units. The biosecurity team had its own custom objects, and nothing we built could touch them.
Objectives.
- Make Attio the system the AI Unit actually works in, and retire the Google Sheet.
- Model the unit's four policy workstreams so stakeholders can be grouped, listed and exported for comms plans.
- Surface relationships that have gone quiet, per point person, without anyone keeping a reminder list by hand.
- Migrate the sheet without email addresses, deduping against existing synced people.
- Build only additively. No changes to existing records, attributes, or the other units' objects.
The data model.
One custom object on top of the standard People and Companies. Imogen trimmed the first draft on the build review: fewer fields on Workstreams, and cadence set per person, not per workstream.
People
Standard object. Stakeholders, segmented by type (stakeholder, employee, subscriber). New attributes for relationship owner, category, government subcategory, engagement tier and bio.
Companies
Standard object. Stakeholder organisations, with a new sector field (government, academia, civil society, think tank, industry, media).
Workstreams
Custom object for the four policy areas: National Security, Loss of Control, International, and General. Holds the lead, launch date, topics, status, and the CLTR team members involved.
What we built.
Seven lists
A master dissemination list, a priority stakeholders list, one list per workstream, and a Substack subscribers list tagged by import source. Each workstream list exports to CSV, which is how the team now builds comms plans for a new output.
Six views on People
One by relationship owner, so each person sees only the stakeholders they own. One per workstream. And one for anyone not contacted in more than 90 days.
Four automations
Setting a person's type to Stakeholder adds them to the priority list. Assigning a workstream adds them to that workstream's list, and removing it takes them off. A weekly job creates a follow-up task for the relationship owner on anyone not contacted in more than 90 days. Call recordings become a summary note on the right person.
Workstream vs list
The team asked for the difference in plain words, so it went into the runbook. A workstream is a business entity that can hold many records and fields. A list is an isolated space for running one repeatable process, like a dissemination round.
The migration.
No emails to match on
The sheet deliberately carried names, roles and organisations, and no email addresses. We matched on name and organisation against people Attio had already created from Gmail and Calendar, and pulled their emails from there.
Priority stakeholders first
The 40 priority rows went in first: 34 updated existing people, 6 were created new. We flagged 16 sets of duplicates, 25 records in all, for a reviewed merge rather than an automatic one.
Only what the team asked for
Relationship strength, influence ratings, issue alignment and free-text context stayed in the sheet by Imogen's choice. Names, roles, relationship owners and emails came across. The sheet was retired once the migration was done.
Results.
- The Google Sheet is retired. Attio is the single place the AI Unit tracks who it knows, what was said, and when it last spoke to them.
- A 90-day prompt on every priority stakeholder, which the unit adopted as its rolling review of who belongs on the priority list.
- Stakeholder lists per workstream that export straight into a comms plan, instead of being rebuilt from scratch for each output.
- A live onboarding session for the unit, recorded for the members who could not attend, plus a Notion runbook.
- A purely additive build: no existing record, attribute or other unit's object was changed.
- At handover we deleted every project file on our side, closed the shared Slack channel and removed every connection to their workspace.
Ready when you are.
Book a call and we will tell you honestly whether this is worth doing, or start with the free 48-hour audit and decide afterwards.