How to Start a Mentoring Program at Work: A Step-by-Step Guide

What makes a workplace mentoring program work?

A formal mentoring program is a structured way for an organisation to pair people so that one, usually the less experienced, can learn and grow with the help of the other. What separates a program from a kind idea is the structure around the pairs: a purpose, clear design choices, fair matching, preparation, support while pairs are meeting, and a proper ending.

The research is encouraging but sobering. A 2008 meta-analysis by Lillian Eby, Tammy Allen and colleagues in the Journal of Vocational Behavior compared mentored and non-mentored people and found mentoring linked with a broad range of favourable outcomes, but with generally small effects, larger in workplace and academic settings than in youth mentoring. Mentoring helps; how you run it matters.

Design is where you have most control. In a 2006 study in the Journal of Applied Psychology, Allen, Eby and Elizabeth Lentz found that two features of formal programs were consistently related to better mentoring and relationship quality: participants feeling they had input into the process, and training they rated as high quality. Both are choices you make before anyone meets.

Demand is real. LinkedIn's Workplace Learning Report 2025 found that 55% of the organisations it surveyed offered mentorship programs, and only 15% of employees said their manager had helped them build a career development plan in the previous six months.

How do you choose the purpose and type of mentoring?

Write the purpose in three sentences: the problem you want to help with, using one or two facts from your own organisation; the people the program is for; and what will be different for them and the organisation in a year. Every later choice, from eligibility to measures, follows from this.

Then choose a type. Each one changes the design.

• Mentoring is a relationship in which someone shares experience to help another person grow, usually over months.
• Coaching works on specific goals, mostly through questions, often over a shorter period.
• Sponsorship is a senior person using their influence to advocate for someone's advancement. If your purpose is advancement, you may need sponsorship alongside mentoring.

What design decisions do you need to make?

Make seven decisions and write the reason next to each, because each gap becomes a reason for a pair to drift.

Count the true cost too. Mentors are not paid, but time is the largest cost: the program lead's, the mentors' and the mentees'. Include it in the budget along with briefings, a launch and a closing event, then ask a senior sponsor for specific things: to open the launch, to ask about the program monthly, and to approve the time and budget.

• Length: six months is a sensible default for a first cohort.
• Rhythm: every three weeks, 45 to 60 minutes. Research on protégé perceptions links more frequent interaction with better mentoring.
• Eligibility: who can be a mentee or mentor, and who cannot. Check that part-time, shift and remote workers are not excluded by accident.
• Opt-in: voluntary applications work better than conscription.
• Confidentiality: what stays in the pair, what the program team sees, and the limits.
• Line managers: informed and supportive, but not told the content.
• Work time and recognition: mentoring in work hours and a visible thank-you for mentors.

How do you recruit mentors and mentees?

Recruit mentors first, because they are usually the scarce side. Ask leaders for named suggestions of people who listen well, not only the most senior, and invite each one personally. Many good mentors hold back because they think they need all the answers; tell them good questions matter more and that there is a briefing and support.

Then open applications to mentees widely. Hold a short information session, recorded for other time zones and shifts. Ask site leads and managers to encourage specific people who would not usually apply. Use the same list of goal areas on the mentee and mentor forms, so you can compare a mentee's goals with a mentor's strengths when you match.

How do you match mentors and mentees fairly?

Matching by gut is inconsistent and drifts towards pairing people who look and sound alike. That kind of similarity is overrated: a meta-analysis by Eby and colleagues in Psychological Bulletin found that similarity in values and attitudes was most strongly tied to relationship quality, while surface similarity such as gender, race or age had very small effects.

A simple scoring method helps. Score each possible pairing from 0 to 5 on a few criteria, weight them for your purpose, and rank the options for each mentee. Then check conflicts of interest: never match someone with their line manager, and look carefully at pairs in the same team. A person makes every decision and writes down why. If nobody fits, do not force a weak match; offer the next cohort or recruit another mentor.

How do you prepare mentors and mentees?

Untrained mentors tend to tell, because telling is what experienced people know how to do. Untrained mentees tend to wait. A short briefing for each side changes both habits.

A 90-minute mentor briefing can cover what mentoring is in your program, listening and open questions with practice in threes, sharing experience as stories rather than instructions, confidentiality and its limits, the first meeting, and what to do if it is not working. A 60-minute mentee briefing can cover owning the goals and agenda, writing goals with a baseline and target, preparing for meetings, and raising concerns.

Hold a joint launch with the sponsor and make sure nobody leaves without the first meeting booked. In that first meeting, pairs should build trust and agree how they will work, not solve problems: goals in the mentee's words, the meeting rhythm, confidentiality read aloud including its limits, and the next three meetings in the diary.

How do you keep mentoring pairs meeting?

Pairs rarely end in an argument. They end in cancelled meetings and embarrassment, and nobody tells the program lead. Make drift visible: track when each pair last met and when the next meeting is due, run a four-question check-in at month one, the mid-point and month five, and log concerns.

Combine the warning signs into a simple rating. One sign, such as an overdue meeting or a low check-in score, is Amber: send a friendly nudge. Two or more is Red: call each person separately and offer three options, a reset of goals and agreement, a new rhythm, or a no-fault close. Say from day one that either person can end a pairing with no reasons and no blame, and rematch quickly when it happens.

Around the half-way point, ask each pair to review their goals against the baseline, rate the relationship separately and compare, and decide to continue, reset or close. A short clinic for mentors at the same time gives them support too.

Where does AI fit in a mentoring program?

AI assistants can help a program lead draft announcements and proposals, stress-test a design, summarise anonymised survey comments into themes, and suggest matching options from anonymised summaries as a second opinion. Participants can use them to turn a vague wish into a goal or to rehearse a conversation.

A person should decide every match, rematch and no-fault close, and every conclusion about impact. Never paste what someone said in confidence, or their name, into a tool your organisation has not approved. Regulation of AI in employment is changing in several countries, so check current rules with your data protection lead.

How do you measure the impact of a mentoring program?

Plan the ending before the launch. Each pair should hold a closing conversation that looks back at their goals, names what changed and agrees what happens next. Then measure at several levels, borrowing from Kirkpatrick's four levels and Brinkerhoff's Success Case Method: participation (pairs completed, meetings held), reaction (check-in scores), learning and behaviour (goal progress from baseline, with examples), and results such as retention or internal moves.

Be careful with results. Comparing participants with a similar group is useful, but groups are often small, people chose to take part, and other things change at the same time. Report comparisons as signals with those limits written beside them, and lead with what you can stand behind. Interviews with the most and least successful pairs often show more than any number: what helped, and what got in the way.

Mentoring can support wider goals. In a 2016 Harvard Business Review article, Frank Dobbin and Alexandra Kalev reported that mentoring programs were followed by gains in management representation of 9% to 24% on average for several under-represented groups. Treat that as a reason to design for inclusion, not as a promise.

As you grow, EMCC Global's International Standards for Mentoring and Coaching Programmes offer a useful checklist of six areas: clarity of purpose, stakeholder training and briefing, selection and matching, measurement and review, ethical standards, and administration and support.

Before the program, mentors and program leads can check their own starting point with an assessment such as AssessAll's Mentoring & Sponsorship assessment. At the close, mentees can turn their next steps into a measured plan with a structured individual development plan such as Jobulary's. If you want every step in one place, Bodhih Training's Mentoring Programme in a Box kit has a workbook, forms, guides and a calendar for each stage in this guide.