How matching works

Skills are only part of it.

Two people can match on every skill and still stall after the first call. What keeps a pairing going is where each of them is heading, why they signed up, and how they like to work. Our matching reads for all of it.

See how it works

The inputs

What a match is built from

Intake asks every participant, mentors and mentees alike, for more than a list of skills. Four things carry the match.

Step 3 of 7

What areas do you want to grow in?

Select up to 6 areas where you'd like support as a mentee.

0 of 6 selected

ACommunication

BStakeholders' Management

CFeedback (giving/receiving)

DConflict Resolution

ETime Management

FDecision-making

The process

From full cohort to final match

Every mentor is scored against every mentee, across everything intake captured about them both. The model weighs it together and produces a match percentage for each possible pairing. Nothing rests on who happens to know whom.
For each mentee, only the strongest candidates go through to a closer read. Where a mentee’s strongest candidates are already taken, the process works down to the next set, and keeps going until everyone in the cohort has a pairing that holds up.
For shortlisted pairings, the written answers are read closely. Two people can name the same skill and mean entirely different problems, and two motivations can sound alike yet pull in different directions. This stage exists to catch both.
Before anything is sent, a person looks at it. Pairings the model flags are held back and judged individually, whatever their score. A high percentage is a recommendation, never a decision.

Written answers, side by side

Both named delegationSame problem

Mentee“I still take the hard tickets myself because handing them off feels slower.”

Mentor“Took me a year to stop being the team’s best engineer and start being its lead.”

Both named stakeholder managementDifferent problem

Mentee“Getting other team leads to back my priorities, now that I’m the one asking.”

Mentor“Ten years of managing enterprise clients and their procurement teams.”

The highest score in the cohort was held back. Here’s why

The alternatives

Other ways to match a cohort

Most programmes match in one of three ways. Here is what each of them misses.

Other ways to match a cohort
	Spreadsheet	Self-matching	General AI tool	PeopleTopics
FairnessDoes everyone get a strong match?	×The last pairs get what’s left	×Popular mentors get over-asked, quieter people wait	×No view of who’s already taken	✓Every mentor scored against every mentee, until every pairing holds up
BiasWhat decides the pair?	×Whoever builds the sheet, and who they know	×Titles, headlines and first impressions	~Whatever made it into the prompt	✓The same four factors for everyone. Nothing rests on who knows whom
DepthDoes it read beyond the skill list?	~Only if someone reads every answer	×Profiles get skimmed	~Only what you paste in	✓Written answers read for every shortlisted pair, like Marta’s below
JudgementWho checks before it’s sent?	✓Whoever builds the sheet	×Nobody reviews the pair	~Only if you check it yourself	✓A person signs off every pair, whatever its score
Set-up timeHow much HR time does it take?	×Days of reading and sorting	~Little for HR, a lot of chasing	~Prompting, checking, re-prompting	✓About an hour of HR time
LearningWhat do you know afterwards?	~Whatever you track yourself	×Nothing beyond who picked whom	×Nothing once the chat is closed	✓Feedback after every session, in one cohort report

A worked example

Same skill on paper. A different problem underneath.

A skills-only match would have paired Marta with Anneke. Reading what they both wrote changed the pair.

Marta Reinhard

Newly promoted engineering lead

Marta, on stakeholder management“Getting other team leads to back my priorities, now that I’m the one asking.”

A score-only system stops here. Switch to see what reading changed.

Anneke de Vries
91% matchTop score
Covers every skill Marta named.
Marta’s skills · 3 of 3
- Stakeholder management, covered
- Difficult conversations, covered
- Delegation, covered
Tomás Vieira
87% match
Covers two of Marta’s three skills.
Marta’s skills · 2 of 3
- Stakeholder management, covered
- Difficult conversations, not covered
- Delegation, covered

The standard

Nothing goes out unchecked.

Every match is scored across the whole cohort, read closely where it matters, and signed off by a person before anyone receives an introduction. That is the standard the programme is built on.

Or go back to the homepage

Skills are only part of it.

What a match is built from

From full cohort to final match

Score

Shortlist

Read

Review

Other ways to match a cohort

Same skill on paper. A different problem underneath.

Marta Reinhard

Anneke de Vries

Tomás Vieira

Nothing goes out unchecked.