Keep your hands on the steering wheel

There's a tension a lot of founders carry around AI, and it has nothing to do with prompts or tools - it's trust.

What will my clients think if they notice I'm using AI too?

Will they quietly downgrade me in their head, assume I'm cutting corners, or start doing that mental math of "if he's using this as well…what am I actually paying for?"

That little spiral is real, and I don't think people talk about it enough because it sounds insecure on the surface, even though it's a very normal response when the ground under value perception is shifting.

Last week I wrote about the shift from expert to curator, and how the value is moving from "having the answers" to being able to choose what matters and guide what happens next.

Today I want to make it practical - how to use AI inside client delivery in a way that supports the work, without replacing the part people came for in the first place.

Tech nerd reporting for duty

Most people know me as the client flow guy now - momentum, renewals, win-win scenarios, the client-first side of business.

But the tech nerd part was there from the beginning. Back in 2014-15 I was a complete nobody in online space, so I got hired for one reason: I could build funnels and automations fast, and I undercharged because I was grateful anyone trusted me.

That skill led into bigger and bigger launch projects over the years, where I'd first build the entire funnel and automation stack myself, and later have a full-time team running those done-for-you launches.

Some of my favourite client launch projects:

Which is why it might sound strange when I say this:

AI agents are the first tech wave in a long time where I didn't jump in the moment everyone started shouting about it.

I've learned the hard way what happens when you build something impressive before you build something usable (but that's a long story for another time).

So I stayed away from the over-engineered setups and waited until I found a tool that fits how I want to build now.

Something I can use weekly without having to turn it into a 5-person team tech project, and something my clients can build with too (if they ever want to model the same approach).

The trap I don't want you to repeat

Corporate world isn't immune to trend hype either, and I had to double check this because it sounded too extreme to be true.

The MIT NANDA report on GenAI in business claims that 95% of organizations are seeing zero measurable return from GenAI efforts so far, even with tens of billions being poured into it (check the full report here).

The problem is that most teams try to drop AI into a messy way of working and hope it magically fixes it.

If the workflow is unclear, AI just makes the confusion faster - and then someone has to step in to double-check it, rewrite it, and literally become its babysitter.

My answer starts with constraint

Think of AI as a space engine, and your framework as the steering wheel.

The space engine is powerful.

It can generate hundred options in seconds, summarize, structure, rewrite, brainstorm, and sound confident even when it shouldn't.

But without a steering wheel it doesn't take you where you actually want to go - it just drifts in space.

That's why prompts are not the starting point for me. The starting point is deciding what we're protecting, and then giving the tool a clear role inside that.

And in client delivery, the best constraint you can have is your method - your lived experience distilled into a journey that guides client progress.

Remember, clients don't need more information anymore, they need help staying in motion with fewer options and easier support.

In short, they need guiding signals, not noise.

This is what I meant in last week's curator letter. The value is no longer "I can give you the best strategies."

The value is "I can help you choose what matters next for you, and stay with it long enough to win."

(more on what I mean by Curator in letter 007)

What my constraint looks like in practice

In my work, it always starts in the same place - Client Flow framework

People first need to understand what's really happening with their clients, not what's happening in a clean curriculum on paper.

What's happening with client momentum:

• where the blocks are
• what the micro-wins are that keep people engaged
• and where they get stuck even though the content is "great"

I built the MicroWins frame out properly in letter 012.

Then we simplify the journey based on what's real.

Sometimes that means installing a renewal or upgrade path at a specific milestone.

Other times it means removing things from the curriculum, because "adding more value" is making it harder for clients to win.

Only after that do I care about tools, because then the tool has a job. It supports the client journey instead of becoming another shiny distraction.

What I'm building right now

This is the part I've been excited to share with you, because it's where this whole "AI without losing trust" conversation becomes real.

I'm starting a nerdy practical series (FlowOS Lab) where each issue covers one tool, workflow, or agent I'm building or testing behind the scenes, and the philosophy behind it.

Not "here's a 30 minute tutorial on how to build an AI tool" - more like "here's what I'm trying to protect, and here's the simplest way I've found to build support around it."

There are two kinds of things I'm building: core framework tools, and individual tools.

Framework tools are the ones everything else plugs into. They're the reason the rest doesn't turn into random toys.

For example…

One of the reasons I stopped doing 1on1 audits in the past is that they required a ridiculous amount of prep.

I'd spend hours creating custom docs, gathering context, chasing inputs, then putting it together into a big strategy document - before we even got to the part that matters, which is reviewing together.

This approach changes that.

Tools help the client do the first layer of work before we ever get on a call, so our time together is not spent extracting context - it's spent making decisions (that's the curator role in action).

So what are the two framework tools?

1/ Client Flow Architect

This is my way of mapping the real client journey so you're not guessing where clients drift or stall.

Not the "curriculum journey" you planned in Notion, but the journey your clients are actually having - where momentum drops, where questions repeat, where people go quiet, where micro-wins show up, and where your best clients naturally ask for the next step.

The output is clarity: what needs fixing, what needs simplifying, and what needs a human touch point because that's where trust is either built or lost.

2/ Upgrade Path Architect

Once that map is clear, this tool builds on it.

It's my way of designing renewals and next-step support so it feels clean for the client and sustainable for you, not like last-minute renewal roulette.

It takes what you've discovered in the Client Flow Architect - the momentum blocks, the micro-wins, the moments where clients either light up or quietly stall - and uses that to shape upgrade offers that actually make sense.

Mini exercise for you

Before you think about AI tools, come back to your core framework for a second.

If your framework is the thing that gets clients results, then one "foundation tool" should exist to make that framework easier to deliver - without you having to hold the whole thing in your head every week.

So here's the question:

If you could build one foundation tool that supports your framework, what would it make consistently easier for you to do?

A few directions to spark the right kind of thinking:

• Make your framework easier to apply to real client situations (not just "teach it")
• Help you spot where people get stuck inside the framework before they drift
• Help you guide the next step inside the framework without rewriting the same guidance from scratch every time

Questions to get you started in the right direction:

• "What part of my delivery currently relies on me noticing things too late?"
• "Where do my best clients get momentum - and where do others quietly stall?"
• "What do I need to make consistent so trust doesn't depend on my mood, memory, or bandwidth?"
• "What I know I should be doing when it comes to client delivery, but I keep delaying?"

If you caught yourself nodding at any of those questions, good.

With that foundation in place, the rest gets fun.

These are the smaller tools I'm building behind the scenes that help the day-to-day of delivery and client work without making it feel robotic.

3/ GoldDigger is my current favourite.

Every coaching call, group call, workshop, Q&A session you run gets turned into tagged, reusable plays. Patterns, objections, breakthroughs, the moments you normally forget by Tuesday.

Your experience compounds instead of disappearing into Zoom recordings and messy notes, and the goal is to build structured memory that makes you faster and more consistent without losing the human touch.

GoldDigger is my antidote to one of the biggest hidden costs in mentorship: experience that disappears.

4/ WinCatcher, just like the dreamcatcher but for wins, is a lightweight system I'm using to capture testimonials in real time, in a way that still feels like a conversation.

Most clients don't write great testimonials because they answer vaguely.

So, WinCatcher asks a couple of simple follow-up questions to help them land the specifics, without you having to chase them or write the testimonial for them.

And once that's done, it runs a neat little two-step that helps you potentially capture referrals at the same time - again, without making it weird.

(Version 2 will expand this beyond testimonials into micro-wins and language shifts, but right now the job is simple: make proof easier to collect, consistently)

5/ FlowInspector is a diagnostic tool for people who are curious about working with me.

It runs a simple Client Flow check, a lighter version of the Client Flow Architect, so you can quickly see where your client journey is strong and where momentum tends to stall.

Then it gives you a few specific next steps based on real patterns I've collected from live client work through GoldDigger, so you get clarity and a direction fast - before we ever get on a call.

If you want a structured version of this, that's exactly what The Gameplan is now: a guided self-paced FlowOS diagnostic, a 90-minute 1:1 retention call, and a written 60-90 day action plan built around the Three Blocks (Momentum, Founder, Upgrade).

(Yes, the Tom Cruise "help me help you" meme is playing in my head as I write this.)

6/ Momentum Check-in Architect

This helps you design a simple weekly or bi-weekly check-in system that actually becomes a ritual, not another form your clients ignore.

The goal is to turn check-ins into a small habit your clients repeat, while giving you a clear signal system (green-yellow-red) so you can see who's winning, who's drifting, and who needs a human touch point before they disappear.

7/ Renewal Flow Builder

This one helps you build a natural renewal timeline so you're not doing the awkward "so… are you renewing?" message at the end.

You map a simple 30-60 day runway with a clear plan for when to start the conversation, what to focus on, and how to make the next step feel like the obvious continuation for the right clients.

I have a lot more in production, but this is more than enough to get your mind going too.

Which one looks the most interesting to you? If any of this hits that "oh damn, I should build this properly" feeling, The Gameplan is where we map it together.

In the end, keep it human

None of this is about building the smartest or the most complex setup or proving you're "ahead" because you've cloned yourself.

That's why I'm using AI like an engine, but I'm keeping my hands on the steering wheel and building everything around one question:

Does this help me see what's happening early, and respond with better judgment, without the relationship feeling automated?

If you're experimenting with AI in delivery right now and want a structured place to map what you're trying to protect most, that's exactly what The Gameplan is built for.

See you on the Gameplan call,

Filip 🌊

Filip Sardi

Retention Strategist · Founder of Client Flow & FlowOS™

I built Client Flow and FlowOS Lab because I've felt what it's like to give your all and still have clients fade away. Twelve years in the online arena - crafting offers, running launches from €50k to million-dollar campaigns, driving sales. It never made sense that everyone would put so much time, money, and energy into their launches just to lose most of those clients before the next one.

I'm building the system I wish had existed - for the mentor who senses the drop-off but can't fix it with another Zoom call, for the coach who knows most people aren't finishing and secretly wonders if it's their fault, for the founder who shows up fully and still feels like they're holding it all up.