In February I launched the cross-mentoring pilot starting from my own experience and the conviction that it is possible to offer talent development opportunities without the high cost attached. 

I recruited three participant companies, paired 18 employees from different companies into nine mentor and mentee pairs, gave them three sessions and a light structure, and let it run until June.

I had built mentoring programmes from inside companies before, so I knew what goes wrong there: the same few senior names asked to mentor everyone, a pool too shallow to match anyone well, and a programme that dies once the person running it gets a busier quarter. The pilot was built to test whether taking the pairs outside the company solves that.

It does. Here is the evidence.

What the pilot was

End of February to June. Eighteen employees, nine pairs, each pair drawn from two different organisations. The mentee brought the topic, the mentor brought the experience, and each pair had three sessions to use. Topics were professional soft skills and leadership fundamentals.

I surveyed participants at the halfway point and again at the close.

The results

  • 8.4 out of 10 average likelihood to recommend the programme to a colleague. Nobody scored it below 8.

  • Every respondent said they would take part again.

  • Every mentor said the experience was rewarding.

  • Access to perspectives people could not have found inside their own organisation scored 4.7 out of 5, the highest-rated element of the programme.

  • At the halfway mark, 75% of pairs were on track to complete their sessions, with match quality at 4 out of 5, rising to 4.4 among pairs reporting a strong fit.

What people valued most

I asked an open question about the most valuable part of the experience, and the same theme came back in nearly every answer, from mentors and mentees alike. It was the view from outside their own company.

A mentor named the fresh perspective that comes from working with someone at a different organisation. A mentee described gaining insight into how another company handles the problems she deals with in her own. Another mentor valued going deep into problems that people do not normally bring to a networking conversation. One mentee put the most valuable part as a large dose of confidence, from talking to a senior leader in her field and having her instincts and ways of working validated.

Nobody described the mechanics of the programme. They described the access.

What surprised me, and what it changed

Some pairs that looked strong on paper did not click.

Before the pilot I assumed the thing a mentee wanted most was expertise: what a mentor had achieved and which company they had done it at. That is what I matched on. Watching well-matched-on-paper pairs stall told me skills were only part of the equation, and that the missing part looked different for every mentee.

So I spent months on what actually makes a pair work. Some mentees value a mentor with a similar background, education or industry experience. For others, personal characteristics carry more weight. Published research points at values, achievement and ambition in particular, as among the strongest drivers of affinity between two people.

Those findings are built into the matching algorithm the October cohort runs on. Skills still matter. They are no longer the only data point.

Mentors got as much as mentees

This matters, because every company in a cohort supplies mentors as well as mentees, and the first question any HR lead asks is what their senior people get for the time.

The mentors did not describe it as a favour. They described a look inside another company's way of working and a chance to test their own assumptions against it. One named the benefit as training their live coaching muscle: active listening, working with someone else's problem in real time, applying their own hard-won lessons out loud. Another valued watching a mentee's situation develop across months and coaching over that arc rather than in a single conversation.

In their words, shared with permission:

Interesting matches and a valuable opportunity to share and learn from new perspectives, lots of sparks of inspiration created for sure.

What we changed for October

Three of the nine pairs did not finish their three sessions. The pattern was the same each time: a session gets moved, then moved again, then the pair stops without anyone deciding to stop.

Inside a company, that decay gets interrupted. You pass the person in the kitchen. Their manager asks about it. Cross-company mentoring removes those nudges, so the programme has to supply them.

The pilot ran on spreadsheets. I matched the nine pairs by hand in Excel, expertise in one column and topics in another, sorted until the pairs looked right. Sessions were arranged between the two people by email. I found out how a pair was doing when I sent a survey, which was twice in four months.

Nine pairs is small enough to run that way. It is also small enough that everything going wrong sat outside what a spreadsheet could do anything about. So between June and October we built the platform Cohort 1 runs on. Five changes, each of them coming from something the pilot showed.

Matching works from more than skills. Onboarding now captures numerous signals from both mentors and mentees that a skills list does not hold, and the AI matching analyses the whole picture rather than just the qualifications. 

Booking and messaging happen in the app. Mentors and mentees connect their calendars and book a session in a click, and message each other without opening email. The pairs that faded were the pairs whose next date never got fixed, and several mentors told us they held back from pushing for a date because they felt the initiative had to come from their mentee. Taking the email negotiation out removes most of what they were reluctant to start.

Feedback right after the call. A flash survey goes to both sides while the impressions are still fresh, so we can see a match is working while there is still time to act on it. 

An HR dashboard for each company. Feedback across a company's pairs is aggregated for their HR lead: what their people are learning, at what rate, and where pairs are running into difficulty. 

Session guides in the app. Mentors and mentees each get their own guides’ section that they can browse any time. This way, they can get support between sessions or also during one, each at their own rate.

Who Cohort 1 is for

Cohort 1 opens in October for companies of 50 to 300 people. That is the size where there is real seniority in the building and not enough of it in any single topic for internal mentoring to keep working, where a development budget exists but a full L&D function does not.

If you have looked at your internal mentor pool and found the same five busy names, that is the problem this is built for.

Want to know more? Book a 20-minute call.

Illustrations by Storyset