The final coaching session goes well. The coachee can describe what changed, the coach closes the engagement, and HR receives a neat completion note. Three weeks later, the line manager still runs the same meeting in the same way. Nobody has agreed when the new behaviour will be used, who can see it or what to do when it slips.
A good closing conversation is not a handover. If the coachee only changes while the coach is present, the workplace has borrowed a behaviour rather than learnt to sustain it. The manager does not need the coach’s private notes. The manager needs an observable action, a place for it in the working week and permission to ask about it. Those are different things, and each should have its own small artefact.
Why the handover disappears
Coaching works in a protected conversation. Work happens in meetings, reviews and difficult decisions where the coach is absent. At the end of an engagement the coachee may know exactly what they intend to do, yet their manager sees only the old routine. That is a transfer problem, not necessarily a coaching-quality problem.
The research is less conclusive about long-term maintenance than vendor claims suggest. A workplace coaching meta-analysis found positive outcomes but called for longitudinal studies that track when outcomes change and how long coaching needs to last. It did not establish that a particular closing checklist preserves gains. Treat the handover below as a practical design to test, not a proven formula for permanent behaviour change.
There is also a boundary that matters. The International Coaching Federation’s ethics code requires a clear agreement about what information moves between coachee, coach and sponsor. A manager may observe a behaviour; that does not entitle them to the contents of a confidential session. The handover must be written with the coachee, not extracted from the coach. This distinction belongs in the coaching engagement before the first session, then gets tested again at the last one.
Start with a behaviour card the manager can see
About two sessions before the end, ask the coachee to select one behaviour that occurs in ordinary work. Not “shows more confidence” or “becomes strategic”. Those require a guess about a person’s inner state. Use a verb, a setting and a visible finish.
For example: “At the start of the weekly production review, asks each shift lead for one unresolved safety constraint before assigning actions.” A manager who attends that review can tell whether it happened. They can also offer a different interpretation of its quality. That is useful evidence; an adjective on a development plan is not.
Put five fields on a single card: the situation, the action, what counts as done, the next opportunity to practise and the coachee’s name. The coachee writes or approves every field. The manager may suggest a work setting, but must not turn the card into a rating of the person’s character. If several behaviours came up in coaching, choose one for the first handover. A crowded card becomes a performance review in miniature.
This is where a competency framework helps if it contains observable anchors rather than a list of virtues. The test in a competency framework nobody uses is subtraction: a small number of actions people can recall during work is more useful than a long vocabulary nobody can apply. The handover card is smaller still.
Do not ask the coach to write a diagnostic paragraph for the manager. That risks disclosing material the coachee did not consent to share, and it replaces a future action with a retrospective description. The manager needs to know what to watch for next Tuesday, not what was said last Tuesday.
Give the action a calendar, not a promise
The second artefact is a short practice calendar. It names the actual work events where the behaviour can happen, and the person who will notice. An abstract goal such as “practise active listening regularly” has no slot on a calendar. A real weekly review does.
List the next four opportunities rather than booking four extra coaching meetings. For each one, record the date, the meeting or task, the cue that starts the action and a tiny record the coachee can make afterwards. “When the shift lead names a constraint, I will ask who owns its removal before I offer a solution” is specific enough to rehearse. “I will be a better leader” is not.
If the routine is less frequent, use the opportunities that exist. Do not pretend every leadership behaviour has a weekly occurrence. A disciplinary conversation might not arise in the next month. In that case choose a related regular behaviour, such as asking for the employee’s account before offering a judgement during a one-to-one. If even that cannot be observed, the handover card needs a different setting.
The calendar should belong to the manager’s existing rhythm. Add a brief check to a one-to-one or a review already on the diary. A separate HR reminder often dies when the programme administrator moves on. The wider programme design should reserve this manager time at the outset rather than sending an optional follow-up request after the coach leaves.
Agree what the manager may ask
The third artefact is a review agreement, no longer than a paragraph. It names the manager’s question, the evidence they will use and the information they will not request. A workable opening is: “At Tuesday’s review I noticed you asked for constraints before assigning actions. Which part helped the team, and what would you change next time?” That starts from an event both people saw.
A poor opening is: “What did your coach tell you about your confidence?” It pressures the coachee to report a private conversation and leaves the manager unable to help with the next work event. Another poor opening is: “Are you fixed now?” It turns development into a verdict. Use the observed action, then ask what the coachee will try at the next opportunity.
Write down the limit of disclosure: the manager receives the card, calendar and the coachee’s account of work behaviour. Session notes, personal disclosures and the coach’s interpretation stay private unless a separate agreement permits their release. The ICF code asks for clarity about information exchanged among all parties, not a vague assurance that “coaching is confidential”. HR can check that an agreement exists without collecting the private content itself.
Set an end condition too. If the manager misses two agreed reviews, the coachee asks to reset the diary. If the behaviour cannot be observed because the role changed, they replace the action rather than logging a failure. These are design rules for the engagement, not research-derived thresholds. Their value is that a silent lapse becomes a visible decision.
Worked example: a plant operations manager
Illustrative example. A plant operations manager finishes a six-session coaching engagement. The original brief was “delegate more”. That phrase would be useless on a handover. The coachee and line manager instead choose the weekly maintenance review as the setting.
The behaviour card says: “Before taking a maintenance issue back, ask the responsible supervisor for a proposed next action and a date; record the owner in the meeting note.” The practice calendar lists the next four Thursday reviews. After each, the coachee notes whether the question was asked and whether an owner was recorded. The review agreement gives the line manager ten minutes in the following one-to-one to discuss one example, without asking what the coach said.
Suppose the first review has three eligible issues and the question is asked on one. The next has four and the question is asked on three. Those figures are invented solely to show how the observation works. They are not a client result, a benchmark or proof that coaching caused improvement. The useful question is why the missed issues differed. Perhaps an urgent breakdown made delegation inappropriate. Perhaps the coachee fell back into solving the problem. The manager can distinguish those possibilities by discussing the work, not by demanding a coaching transcript.
At the fourth review the pair decide whether the behaviour is now part of the meeting or needs another cycle. If it is only performed when the line manager watches, the handover has not finished. Move the observation to a routine note or ask another meeting participant to confirm ownership, with the coachee’s agreement. Do not quietly turn the card into surveillance.
Where this method stops
Three artefacts will not repair a manager who punishes a coachee for trying the new behaviour. If a supervisor asks for delegation in a development meeting but takes every decision back under deadline pressure, the coachee has learnt a rational lesson from work. Change the manager’s response first. An external coach cannot override it with a better closing session.
This method also needs a behaviour that can be observed without exposing private material. A coachee working on grief, identity or a deeply personal conflict may have a valid coaching goal with no suitable manager-facing card. Do not force one. Agree a narrower work behaviour if the coachee wants it, or keep the engagement private and do not claim an organisational transfer measure. A completion rate is not a substitute for consent.
Nor should the card be treated as a causal evaluation. Many things change during an engagement: team staffing, workload, the manager’s own behaviour. The smallest honest evaluation can document whether a target action appeared in work, but it cannot attribute every change to the coach without a stronger design. Keep that distinction explicit when reporting outcomes to a sponsor.
What to do on Monday
Take the next coaching engagement due to finish and book a short closing handover with the coachee and their manager. Ask the coachee to bring one action they want visible at work. Leave with a behaviour card, four genuine opportunities on a calendar and a paragraph spelling out what the manager may ask. If the coachee declines to share a work goal, respect that decision and report the handover as not agreed. That is more honest than a glossy closure slide.
Sources
- https://coachingfederation.org/credentialing/coaching-ethics/icf-code-of-ethics
- https://pmc.ncbi.nlm.nih.gov/articles/PMC10597717
- https://www.worklearning.com/wp-content/uploads/2020/01/Transfer-of-Training-Quick-Research-Review-2020-v1a.pdf