AI and the Coaching Profession: Key Considerations for Coaches Using AI

Has the coaching profession taken a hit from the uptake in AI? Yes, it likely has. Many people are heading to AI to seek advice, to help them better plan their days and to set goals. When used conversationally, AI has the potential to guide users to unlock ambitions and then put together a personalised plan for achieving them. So, can AI replace coaches? No. No, for so many reasons.

Firstly, because transformation is only possible when real work is done to uncover true wants and desires and confront obstacles and self-imposed limitations. This demands absolute honesty, and because AI is designed to please and generally agree with the user, it cannot scrutinise or apply that human understanding that a good professional coach can. Even where it can process tone or text cues, it can't apply the instinctive, relational judgement a trained coach brings. Secondly, no matter how much people may come to rely on AI, they will never feel accountable to it, and many seek out a professional coach because they need to be held accountable. AI isn’t going to know whether you’ve met this week’s goals, and it isn’t hosting a weekly meeting to check in and talk through how it’s all going. So, whilst some coaches may have felt the sting of AI as potential clients opt to experiment with using Claude or ChatGPT, it’s never going to have the same value.

Can coaches use AI to their advantage though? Of course. There are many ways that AI can help service providers free up more time to actually help clients by lessening the admin load and streamlining workflow. Working solo, as many coaches do, AI can also be a great sounding board to bounce ideas off of, untangle thoughts and identify ways to market and develop your coaching business.

With the world of work and lifestyle changing so rapidly with the somewhat sudden common use of AI, it can feel like a sink-or-swim situation. A call to panic.

Instead, though, let’s take some time to reflect on how we might use AI responsibly, what the legal considerations are for coaches using AI, and how we might mitigate risk.

How AI Can Be Used In Coaching

Most service providers, especially those in caring professions, have long wanted to dedicate less time to administrative tasks and more time to client care. Client care is emotionally taxing. Doing this and then having to run a business in between these hours is a strain. AI can't completely solve this - it needs oversight and leadership, but it can help.

Here are some areas where coaches may consider using AI –

Admin and Operations: Tools you’re already using may be more capable now of scheduling, invoicing, content management and automating reminders, if you make use of their AI capabilities. Or, AI agents can be trained to automate, create and perform operations, including invoicing, onboarding and other processes.

AI Notetakers and Session Transcription: Otter.ai, Fireflies, Fathom and other tools can join calls to transcribe and/or summarise, meaning you’re not having to physically take notes during sessions or write them up afterwards.

Session Prep and Brainstorming tools: You may instruct AI to come up with coaching questions and session structures, but you (the coach) must retain full oversight. If you’re not applying your expertise and insights as a professional, then you risk diminishing rather than amplifying your value. You may also wish to use AI as a sounding board, helping you to unravel thoughts and ideas in the way you might if you had a partner working beside you.

Of course, another revenue stream, if you wish, could be to create AI coaching tools yourself. These can be modelled on your own training methods and intended for exclusive use by your clients. This could add value to your offering by providing interim support between sessions, or to provide you with useful information and insights prior to a client's first consultation.

What UK Law States About AI

Currently, there is no specific legislation around AI use in the UK. Instead, the UK points to other existing regulations, including UK GDPR.

Fundamentally, this means that no use of AI overrides existing legal obligations around data protection, responsible advertising, copyright, transparency, etc.

In practice though, using AI can make it harder to ensure compliance, since you are outsourcing to a machine capable of operating autonomously. And we all know AI makes mistakes. It oversteps and has been known to harvest data, share confidential information and breach copyright.

AI is also known to be prone to bias, reinforce stereotypes, give false or misleading information and even encourage harmful behaviours. As such, there are many cases being brought against AI builders. Most often though, when AI makes a mistake, the user is the responsible party and, if used within a business, then responsibility lies with the business owner. Therefore, before utilising AI in your business, it’s wise to know how it works and put some checks and processes in place to ensure AI-generated work and actions have human oversight, and that that person is capable of properly dealing with any AI-related issues.

The coaching relationship is built on trust and confidentiality, so in this profession, it’s especially important to trust any tools you’re using in your practice.

Let’s look at some of the key risks AI poses and how coaches can mitigate against these -

The Key Risks Of AI Use In Coaching (And How To Mitigate Them)

When using AI, you are delegating. Whether delegating to a person or a system, you are still ultimately responsible. If AI makes a mistake, applies bias, or intrudes on data protection, you are the responsible party, and so we must identify where risks lie.

Confidentiality

Coaching professionals often get deeply personal with clients, and this information is meant to be private. When using AI tools in therapy, even outside of sessions, this does not override your responsibility to uphold confidentiality. For example, AI notetakers recording/transcribing during sessions means a third-party tool now holds a recording of a confidential conversation, and some free or consumer-grade tools use this data for model training, which risks it being made public.

Mitigating Risk of AI Confidentiality Breaches

Obtain explicit consent for any use of recording or transcription tools. This consent should be given as part of your coaching agreement/contract. However, it’s also wise to obtain consent before switching these tools on during sessions, as clients have the right to change their mind.

Before using any such tools, though, or inputting any client information into an AI tool, you should vet the vendor's policies, particularly around data handling and retention. Crucially, check whether the tool is training on data and if there is an opt-out for this. If using AI at scale to process client data, you should run a DPIA (Data Protection Impact Assessment).Opting out of training on data models can prevent accidental data leaks. However, you should generally be cautious about inputting any sensitive data into AI chats.

Accuracy

AI is very impressive, no doubt, but just like any high-flying employee, it’s still capable of making mistakes. Unlike a human, though, it’s not likely to realise a mistake has been made. So it’s not a matter of training an AI and then fully entrusting it; it can let you down at any time and will always need its work checking.

AI can generate plausible-sounding content that may still contain inaccurate summaries, generic insights, and speak in false confidence. Therefore, leaning on AI-generated frameworks can mean you’re relying on mass generalised data conclusions rather than treating/coaching the individual.

Mitigating Risk Of Mistakes Or Generalisation in AI

Treat AI outputs as a starting point, not the final word. Check through AI-generated summaries and scrutinise insights using your own honed skills and experience. The benefit of using AI to transcribe or summarise conversations, for instance, is so you can give greater focus to the client. Therefore, you will know, when checking the AI output, if it’s accurate. Agentic AI has been trained to encourage you to give them more and more responsibility as you work together, but you have to retain your red line and decide at what point you take over and apply your expertise. For example, you might ask AI to transcribe a session, then summarise. It might then ask you if you’d like it to put together a coaching plan for your client. And here may be the point where you decide, no - I am the more trusted expert and that part should not be outsourced to AI. It’s all up to you, of course,  but as well as assessing the risk of inaccuracies and poor advice, you must also be able to judge where your professional value lies and what should be entrusted only to you.

AI Bias

AI models can reinforce stereotypes and practice profiling, which can lead to career ‘fit’ suggestions or life coaching models that are generic and don’t take full account of the client as an individual. Coaches excel at this. And, especially in one-to-one coaching, clients expect a highly personalised approach. Something AI is not yet highly capable of.

Mitigating Bias and Profiling in AI

If you’re using AI to profile or assess clients, don’t use the tool as a stand-alone verdict. You must apply professional judgement and be able to identify and challenge outputs that feel reductive or stereotyped. This may not be a legal pitfall, but relying on AI-generated plans, pathways and assessments can risk damaging your professional integrity as a coach.

Duty Of Care

The coaching profession is not regulated, but you may still come into contact with vulnerable people and so you have a duty of care. Outsourcing coaching work to AI can lead to not recognising red flags; a client disclosure that needs referral, for example. AI cannot be trusted to recognise this.

Upholding Duty Of Care When Using AI

Even when using AI tools for screening, initial assessment or in between sessions, a human coach must be reading and reviewing everything a client shares with the tool, in good time. The global association for coaches, the ICF (International Coaching Federation), positions AI as a complement to coaching, not a replacement. They rightly maintain that due diligence is required at every stage of client care, which would include interactions with AI.

Consent and Transparency

When using AI, it’s important to be transparent. Disclose AI use in all relevant contracts and detail what it is and is not used for. You may also disclose tools that you use so clients can conduct their own due diligence.

For executive/group coaching within organisations, you should also clarify who can access AI-generated notes, etc.

Key Takeaways for Coaches

Ultimately, whatever you do to mitigate risk, responsibility lies with you. Whether you're using an AI notetaker, building your own AI tool for clients, or simply asking AI to draft a session plan, you, not the AI, and not the AI provider, remain accountable for confidentiality, data protection, and the accuracy of anything that reaches a client.

In practice, that means human-led guardianship of your client, over everything you do and every tool you utilise. Every tool you utilise - and this is more personal than legal - should be used to enhance client care, not to outsource it.

Technology can give us back time to focus on the client and increase outreach too, meaning you can help more people. Just remember, when you’re using these AI tools that are forever fostering you to outsource more and more to them, that your professional integrity is your biggest asset, and something that should always be applied.

Lastly, it’s crucial to protect client information when using AI and to be upfront with your AI use in all contracts.

If you haven't reviewed these since introducing any AI tool into your practice, that's the one action from this article worth prioritising. Get in touch with me to review your contracts and consents.