Liberating Structures: The Bold Experiment Mentorship Programs Are Missing

Most mentorship programs are built the same way. Match a senior person with a junior person, hand them a discussion guide, schedule monthly check-ins, and hope chemistry happens. Sometimes it does. Sometimes, it doesn't, and the program quietly dies from disengagement while HR wonders what went wrong.

Employee disengagement is getting harder to avoid. As AI agents take on more of the day-to-day work, it's never been easier for an employee to go a full day without a real conversation with another person. It's efficient, and it's quietly hollowing out the knowledge sharing and mentorship that used to happen just by working alongside people. A mentorship program now has even more competition for people's attention.

If your organization wants to stand out on development and retention in the long run, you need mentorship that actively rebuilds that human connection and can compete against the quiet convenience of an AI assistant.

That's where Liberating Structures come in.

What are Liberating Structures?

Liberating Structures (LS) are a set of simple, repeatable interaction patterns designed to structure interaction in ways that are flexible but focused. In a group setting, they get everyone participating, not just the loudest or most senior voice in the room. There are 43 of them now, with names like 1-2-4-All, Troika Consulting, Wise Crowds, and Impromptu Networking. Each one has a clear structure (how you form groups, how much time you give, how people speak) but leaves the content completely open.

Henri Lipmanowicz and Keith McCandless spent over a decade developing these methods, starting in healthcare settings in the early 2000s before expanding across sectors. Their 2014 book, The Surprising Power of Liberating Structures, brought the approach to businesses, nonprofits, government, education, and healthcare organizations worldwide.

Lipmanowicz and McCandless drew on complexity science, the study of how living systems self-organize, and applied it to how humans work together. Their core critique: since the Industrial Revolution, we've run organizations like machines. Predictable parts, centralized control, information flowing from the top down. That works fine for repetitive, well-understood problems. It falls apart when you're dealing with something messier, like how to help a new employee actually grow, or how to build trust between a mentor and mentee who've never met.

LS treats groups more like ecosystems than machines. Instead of forcing order from the top, it creates simple conditions (how people cluster, how much time they get, how ideas get shared) and lets useful patterns emerge naturally. The result is less "manage the meeting" and more "design the conditions for good things to happen."

That distinction matters a lot for mentorship.

Why this fits mentorship so well

Traditional mentorship programs are hierarchical by design. One person has more experience, one person needs it, and the relationship runs in that direction. That's not wrong, but it's limited. It puts all the pressure on two people to build a working relationship from scratch, often with little training and vague guidance.

Liberating Structures offer a different entry point:

For one-on-one mentorship, structures like Troika Consulting or What I Need From You give mentor and mentee a clear, low-pressure format for working through a real problem together. Instead of "so, how's it going?" (which tends to produce vague, unhelpful answers), the structure gives both people a defined role and a reason to be specific.

For group mentorship, things get even more interesting. Structures like 1-2-4-All or Wise Crowds let a whole cohort mentor each other simultaneously, cycling people through pairs and small groups so everyone contributes and everyone gets input from multiple perspectives, not just one assigned mentor. This spreads mentorship capacity across a group instead of concentrating it in a handful of "designated" mentors, which is a real advantage if your organization doesn't have enough senior people to go around (most don't).

There's also decent early evidence that this works. A small 2022 study out of the University of British Columbia tested a modified Liberating Structure (Five Whys) in a classroom setting and found it built a sense of community, deepened engagement, and encouraged participants to reflect and support each other. This proof of concept lines up with what LS practitioners report seeing in workplace settings: people open up more when the format is clear but not controlling.

Why this matters even more in an AI-first workplace

There's another layer to this worth naming directly: the timing.

As AI agents get woven into more of daily work, from coding to serving customers, a lot of employees are slowly trading human interaction for AI interaction, often without noticing the tradeoff. It's faster to ask an assistant than to interrupt a colleague. It's easier to get an AI-generated answer than to sit with the discomfort of asking a "dumb question" out loud. That convenience is real, but it comes with a cost.

Workplace loneliness has become a mainstream business concern in 2026, not a fringe HR topic. Gartner's 2026 Future of Work research has flagged it as something organizations now need to actively manage as a business risk, on par with other operational hazards. It's easy to see how heavy reliance on AI agents for tasks that used to involve pairing with a colleague can quietly erode the everyday moments where people used to learn from and connect with each other.

This isn't just a wellbeing issue, it's a retention issue. Surveys, like the CIPD's People Management poll (in partnership with Attest), are consistently finding that a meaningful share of employees have considered leaving, or actually left, a job because they felt isolated at work. And loneliness doesn't just push people out the door, it shows up in discretionary effort too. Employees who feel connected are consistently more likely to report going the extra mile for their organization compared to those who feel isolated.

Layer on top of this a separate, compounding problem: a large chunk of the most experienced workforce is retiring right as AI is changing how novice and expert employees traditionally learned from each other. That combination puts younger or newer employees at real risk of missing out on the kind of informal coaching and knowledge transfer that used to happen naturally, just by working alongside people.

This is exactly the gap Liberating Structures are built to fill. They are a deliberate, structured way to protect and rebuild the human connection that AI adoption can inadvertently crowd out. A well-run LS-based mentorship touchpoint forces real, focused, human interaction, the kind that's hard to fake and hard to replace with a chatbot. In a workplace where it's increasingly easy to go an entire day without a meaningful conversation with another person, that's not a nice-to-have. It's becoming a genuine differentiator for organizations trying to keep their people engaged, developing, and around for the long haul.

How to actually bring this into your L&D or mentorship program

If you're an L&D lead or program manager thinking "okay, but how do I actually do this," here's a practical path.

1. Start small and low-stakes. Don't redesign your entire mentorship program around LS on day one. Pick one structure and try it in a single session. Impromptu Networking is a great first pick because it's fast, easy to explain, and helps people build initial connections before anything more serious.

2. Match the structure to the goal. Different structures solve different problems. Want mentees to get quick, focused advice on a real challenge? Try Troika Consulting or Wise Crowds. Want a mentoring cohort to reflect honestly on what's working? Try 1-2-4-All or What, So What, Now What. The Liberating Structures menu breaks these down by purpose, so you don't have to guess.

3. Train your mentors and facilitators lightly, not heavily. Part of the appeal of LS is that it doesn't require extensive facilitation training. A short walkthrough and a printed set of instructions is usually enough for a mentor or program lead to run a structure confidently. That's a real advantage if your organization is stretched thin on formal training resources.

4. Build in real feedback loops. Don't just run a session and move on. Ask people what worked and what didn't. This tells you which structures are landing and which ones feel forced or awkward for your particular group.

5. Mix formats instead of picking one and locking in. The structures aren't meant to replace every 1:1 conversation or program touchpoint. Use them to supplement your existing mentorship cadence, maybe once a quarter for group cohorts, or as a tool mentors can pull out when a regular conversation has gone stale. Flexibility is kind of the whole point.

6. Watch for the downside. Structure helps, but too much of it can backfire. Some people find tightly timed, highly structured formats stifling, especially if they prefer more free-flowing conversation. Keep an eye on this and be ready to loosen the format (more time, fewer rules) if your group signals they need it.

The bigger picture

Mentorship, at its core, is about people feeling heard, challenged, and supported by someone with a bit more experience or a different perspective. Liberating Structures don't replace that. They just make it a lot easier to create the conditions where it actually happens, especially in workplaces where time is tight, budgets are limited, and formal mentorship training with dedicated resources isn't always realistic.

And that connection is worth protecting on purpose. AI is only going to take on more of the work employees used to do together. Organizations that let human connection erode as a byproduct of that shift will feel it in engagement, in retention, and in whether people ever really learn from each other. Organizations that treat connection as something to deliberately design for, the way LS does, will end up with a real edge, one that's genuinely hard for a competitor to copy by just buying a better AI tool.

If your current mentorship approach relies mostly on two people figuring it out on their own, it might be time to borrow a little structure. Just enough to get people talking, not so much that it gets in the way.