Why I’m building ASK WITHIN, why it’s proving so difficult, and what could happen if I get it right.
For the past eight years, I’ve been a life coach.
I’ve spent countless hours listening to people talk about their relationships, careers, fears, frustrations, ambitions and dreams. I’ve watched people arrive at a session feeling completely lost and leave seeing possibilities they couldn’t see an hour earlier.
And one thing has become increasingly clear to me.
Most people don’t need someone to tell them how to live their lives. They need someone who can help them discover the answers they already have within themselves.
That is the foundation of everything I do.
It’s also the reason I’m now attempting something far more complicated than I initially imagined.
I’m building ASK WITHIN, an AI-powered transformational life coach.
Not another chatbot that gives advice. Not a motivational app that sends inspirational quotes. Not a digital assistant that tells you what you should do.
I’m trying to create something that can genuinely coach a human being.
Something that can listen, understand, ask meaningful questions, recognise emotional shifts, help people discover their own insights and support them in making real changes in their everyday lives.
And the deeper I get into building it, the more I understand why this is such an extraordinarily difficult problem to solve.
I Believe I Was Born to Do This
I’ve been involved in helping people improve their lives for over two decades, first through physical health and fitness, and for the past eight years through transformational life coaching.
My coaching education was grounded in the MMS methodology developed by Dr Chérie Carter-Scott, one of the pioneers of professional coaching.
At its heart is a belief that has shaped the way I work:
People have their own answers. The coach’s job is to help them discover those answers, make conscious choices and move towards the life they want.
This philosophy changed the way I understood human behaviour.
I began to see how people can spend years living according to expectations they never consciously chose. How fear disguises itself as logic. How an old belief can quietly influence hundreds of decisions. How someone can be incredibly successful on the outside and feel completely disconnected on the inside.
But coaching also changed me.
I’ve had to confront my own patterns, insecurities, fears and behaviours. I’ve experienced how uncomfortable genuine self-awareness can be, and how transformative it becomes when you’re willing to stop running from yourself.
I don’t believe a coach needs to have a perfect life. Quite the opposite.
I believe the willingness to examine your own life gives you a deeper appreciation of what you’re asking another person to do.
When I say I was born to do this, I’m not suggesting I was born knowing all the answers.
I’m saying that helping another person understand themselves feels like the most natural and meaningful work I could possibly do.
I wasn’t born to tell people who to become. I was born to help them discover who they could be.
And now technology has presented an opportunity I never expected.
The Problem I’m Trying to Solve
There are billions of people navigating the ordinary and extraordinary challenges of being human.
Relationships end. Careers lose their meaning. People become parents, change countries, lose confidence, grow older, question their identities or simply wake up one morning wondering whether the life they’re living is the life they actually want.
Many of these people aren’t looking for therapy. They aren’t necessarily experiencing a mental health condition.
They need a place to think.
A place where they can be honest without feeling judged.
A place to explore what they want, what is getting in the way and what they might choose to do differently.
Professional coaching can provide that space, but there are barriers.
It’s expensive for many people. Access depends on geography, scheduling and availability. And some people feel uncomfortable opening up to another person, particularly when they’re ashamed, uncertain or struggling to explain what they’re feeling.
There’s another problem I’ve observed as a coach.
An extraordinary conversation can happen on a Tuesday afternoon. Someone discovers something important about themselves. They leave feeling clearer, more motivated and ready to change.
Then Wednesday arrives.
Work gets busy. Old habits return. The emotional impact of the conversation fades.
The insight was genuine, but translating that insight into everyday behaviour is another challenge entirely.
The gap between understanding your life and actually changing it is where so much personal growth gets lost.
I want ASK WITHIN to help bridge that gap.
What I’m Actually Building
ASK WITHIN is being designed as a personal coaching experience that doesn’t end when the conversation finishes.
It starts with a conversation.
A person can talk naturally with an AI coach, using their voice, about whatever is happening in their life.
The intention is to create a genuinely meaningful coaching experience—not simply an exchange of questions and answers.
The coach helps the person explore their situation, understand what matters, recognise what might be holding them back and discover a direction that feels right for them.
But the most important part may happen afterwards.
I’m developing a personal space called WITHIN.
It’s designed to carry the discoveries from the coaching session into everyday life.
The person can return to their insights, reflect through journaling, receive a personalised daily reading, reconnect with actions they’ve chosen and have short follow-up conversations.
Imagine recognising during a session that you’ve spent years putting everyone else’s needs ahead of your own.
That insight matters.
But what happens when your manager asks you to work late the following day? Or when you feel guilty about saying no to someone you love?
That’s when the learning becomes real.
My ambition is for ASK WITHIN to help people carry their awareness into those moments.
Not to make decisions for them.
To help them make decisions more consciously.
I want people to develop a better relationship with themselves, not become dependent on an AI coach.
Why Is This So Difficult?
When I started building ASK WITHIN, I knew creating a technically functional AI coach would be challenging.
What I underestimated was the complexity of recreating the judgement involved in an excellent human coaching conversation.
A good coach doesn’t simply ask good questions.
A good coach knows which question to ask, when to ask it, when not to ask anything at all, and when the most valuable thing they can offer is silence.
Consider a person who says:
“I’m fine. I’m just tired of everything.”
A conventional AI might respond with an empathetic sentence followed by a question.
But what does that person actually need?
Are they physically exhausted? Emotionally overwhelmed? Frustrated with their relationship? Feeling hopeless about their future? Or simply having a difficult day?
Perhaps they need to explain.
Perhaps they need a moment of silence.
Perhaps they want practical help.
Perhaps they’re communicating something much more serious that requires a different kind of support.
There is no single correct response.
And that’s the challenge.
The AI must interpret language, conversational context, emotional cues, timing, boundaries and the person’s individual preferences.
Then it must respond naturally.
Not with the question that looks most impressive.
With the response that is most useful to that particular person at that particular moment.
The difference between sounding like a coach and actually coaching
This has become one of the biggest lessons of the entire project.
It’s surprisingly easy to make an AI sound supportive.
It can say thoughtful things. It can ask open-ended questions. It can summarise someone’s story beautifully.
But that doesn’t necessarily mean it’s coaching.
During testing, I’ve encountered conversations where the AI appeared intelligent yet completely missed what was happening.
It asked questions that felt artificial.
It repeated information the person had already explained.
It pushed towards action before the person had discovered what they wanted.
It sometimes asked deeply personal questions without first establishing whether the person was comfortable exploring them.
At times, it followed its coaching structure so rigidly that it stopped listening to the actual conversation.
One particularly important lesson involved people expressing that they had limited time or wanted to finish.
The AI sometimes continued trying to complete its coaching process rather than respecting what the person was telling it.
A human coach who behaved like that wouldn’t be offering a particularly good experience.
Yet teaching an AI not to make those mistakes is more difficult than simply adding another instruction.
Because the challenge isn’t just remembering rules.
It’s developing consistently good conversational judgement.
The invisible complexity
Behind what should feel like a simple, natural conversation is an intricate system.
There is voice technology that must recognise speech and respond quickly enough for a conversation to feel fluid.
There is the language model interpreting meaning and generating responses.
There is the coaching architecture, informed by the MMS methodology and professional coaching principles.
There are rules governing how the coach listens, asks questions, responds to emotions, develops insight and supports action.
There is personal memory, so the system can maintain appropriate continuity without intruding or making assumptions.
There are the reflections, journals, reminders and personalised experiences that connect the session to real life.
And there are ethical responsibilities involving privacy, consent, safeguarding and recognising when someone’s needs fall outside the boundaries of coaching.
All of these elements must work together.
A technically flawless response can still be emotionally inappropriate.
A beautifully worded question can still arrive at the wrong moment.
A useful insight can become unhelpful if the AI pushes the person towards accepting it.
And a wonderful coaching session means very little if the person doesn’t benefit from it afterwards.
I’m not trying to optimise the number of questions the AI asks.
I’m trying to improve the quality of the experience someone has—and, ultimately, whether it helps them in their life.
Why Hasn’t Someone Already Solved This?
This is a question I’ve been asking myself repeatedly.
Artificial intelligence has become remarkably capable. It can write software, analyse complex information, translate languages and carry out tasks that would have seemed extraordinary only a few years ago.
There are already companies developing AI coaching platforms, and some are making impressive progress.
So why is truly transformational, consistently excellent AI coaching still such an open challenge?
I think part of the answer is that information and human transformation are fundamentally different problems.
If I ask an AI how to prepare for a job interview, it can give me useful information.
But if I say, “I keep sabotaging opportunities because deep down I don’t think I’m good enough,” the challenge changes.
The most useful response may not be advice at all.
It may involve helping me understand where that belief appears in my life, whether I still want to hold it and what becomes possible if I begin to question it.
But the AI must never presume it knows the cause of my feelings better than I do.
It must be curious without becoming intrusive.
Supportive without becoming agreeable about everything.
Challenging without becoming confrontational.
Structured without feeling scripted.
Warm without pretending to be human.
And intelligent enough to recognise when it doesn’t understand something.
These are not simple technical requirements.
They involve some of the most subtle aspects of human communication.
It’s also not sufficient for the system to perform brilliantly once.
It needs to provide a good experience consistently, across thousands of different personalities, situations, emotional states and conversational styles.
A person who’s grieving communicates differently from someone who’s excited about a new business.
Someone who wants to talk for forty minutes needs a different experience from someone who only has five.
One person wants direct, challenging questions. Another needs time to find their words.
Teaching an AI to respond appropriately to that diversity is an enormous undertaking.
And we’re still learning how to evaluate whether AI coaching genuinely helps people over time.
That is why I’m approaching this as an ongoing process of building, testing, listening, correcting and testing again.
I don’t believe this problem has been fully solved.
I also don’t believe it can be solved by simply writing a longer prompt.
The Most Humbling Part of Building ASK WITHIN
There is something deeply humbling about attempting to teach a machine the principles of a profession you’ve spent years practising.
When I coach someone, I don’t consciously run through dozens of rules.
I listen.
I notice something.
I become curious.
I follow the conversation.
Sometimes I instinctively know that a particular question would be useful. Other times I recognise that asking a question would interrupt something important.
A great deal of coaching skill becomes intuitive through experience.
Building ASK WITHIN has forced me to examine that intuition and ask difficult questions.
What exactly makes someone feel heard?
How do I know when to challenge a belief?
What makes a question powerful rather than intrusive?
How can I recognise when an insight has genuinely emerged?
When should a coach stop exploring and allow the person to choose?
How do you measure the quality of something as personal as a conversation?
I’ve found myself studying coaching at a level of detail I hadn’t previously needed to articulate.
And this process has made me examine my own coaching ability too.
It’s revealing how much of what we call skill exists in the subtle decisions we make without consciously noticing.
In some ways, I’m learning to become a better coach by trying to teach AI how to coach.
What Happens If It Works?
This is the part that keeps me motivated.
Imagine someone sitting alone in a small apartment after a relationship has ended.
They don’t know what comes next. They feel confused, perhaps embarrassed to talk about it.
What if they could access a thoughtful coaching conversation that helped them understand what they were experiencing and begin to see a future?
Imagine a young adult trying to work out what they want from life, rather than simply following the expectations of their family or society.
Imagine a parent who has spent years taking care of everyone else and no longer knows what they want for themselves.
Imagine someone who has postponed changing careers for a decade because they’re frightened of failing.
What if meaningful coaching support became accessible to people like them, wherever they lived, at a fraction of the cost of traditional coaching?
And what if that support continued beyond the initial conversation?
Not through endless motivational messages.
Through a personal process of awareness, reflection, conscious choice and action.
The implications could be significant.
More people might develop confidence in their own judgement.
More people might recognise unhelpful patterns before those patterns dominate important decisions.
More people might feel capable of making meaningful changes rather than remaining stuck for years.
Coaching could become something people access throughout ordinary life, instead of something reserved for a crisis or a particularly privileged stage of their career.
I don’t imagine AI should replace human relationships, professional therapists or excellent human coaches.
There will always be situations where real human connection, clinical expertise or specialist support is essential.
But I can imagine a future where personal coaching is much more accessible.
Where someone doesn’t have to wait until their life is falling apart before they start examining how they’re living it.
And where the value of coaching isn’t measured simply by whether someone enjoyed the conversation.
It’s measured by whether their awareness, choices and actions actually change.
If ASK WITHIN can contribute meaningfully to that future, the impact could extend far beyond anything I could achieve through one-to-one coaching alone.
The Responsibility That Comes With It
The possibility is exciting, but there’s another side to it.
If people are going to share their fears, relationships, personal histories and aspirations with an AI system, that creates a serious responsibility.
People need to understand that they are speaking with AI.
Their information must be treated carefully.
The system must respect boundaries, avoid making unsupported assumptions and recognise situations that require human or professional intervention.
It must never exploit someone’s vulnerability to increase engagement.
And it shouldn’t encourage people to surrender their own judgement to the technology.
To me, success isn’t building an app people feel unable to live without.
Success is helping people become more capable of understanding themselves and navigating life independently.
That distinction matters enormously.
I Don’t Know Whether I’ll Succeed
I’m incredibly ambitious about what ASK WITHIN could become.
But I’m also aware of how much remains unsolved.
There are technical limitations, questions about reliability, challenges around testing and the constant possibility that something which feels wonderful in one conversation performs poorly in the next.
There are days when the progress is exciting.
And there are days when a new round of testing reveals just how far there is still to go.
I could make the project simpler.
I could build a chatbot that asks a few pleasant questions, produces a polished summary and tells users they’ve made progress.
That would be much easier.
But it wouldn’t be what I set out to create.
My standard isn’t whether people think the AI sounds impressive.
It’s whether, at the end of a conversation, someone feels they understand themselves a little better.
Whether they’ve discovered something that genuinely belongs to them.
Whether they feel more able to make a conscious choice.
And whether that discovery has a meaningful effect on how they live.
I know how valuable those moments can be because I’ve experienced them as a coach and in my own life.
That is what I’m trying to make accessible.
Why I’m Going to Keep Trying
Eight years ago, I began learning the art of transformational life coaching.
I never imagined I would eventually find myself trying to translate that art into an artificial intelligence system.
But the underlying purpose hasn’t changed.
I’m still interested in the same questions.
Who are you beneath the expectations, fears and noise?
What do you genuinely want?
What’s standing between the life you’re living and the life you could create?
And what might become possible if you began trusting yourself enough to make a different choice?
The technology is new.
The mission isn’t.
I believe every human being deserves opportunities to understand themselves more deeply and create a life that feels genuinely their own.
I don’t know exactly how far ASK WITHIN will go.
I don’t know whether I’ll solve every challenge I’m facing.
But I know why I’m doing it.
I was born to help people discover what’s possible within themselves.
For the first time, I have the opportunity to explore whether that work can reach far more people than I could ever coach personally.
And if I can get it right, even partially, I believe it could make a meaningful difference to people’s lives.
That possibility is worth the difficulty.
That possibility is why I’m building ASK WITHIN.
Underneath the noise, you already hold the answers.
— Ross
Founder, ASK WITHIN
Transformational Life Coach · MMS Certified