Gary Does Strategy

Gary Does Strategy Welcome to Gary Does Strategy!

We specialise in helping businesses grow by optimising sales pipelines, enhancing customer engagement, and implementing effective marketing strategies.

09/04/2026

Here's something that surprised me when working with a contact centre last year.

They wanted to automate ticket resolution. Big operation, serious volume. Before we touched any technology, I asked one simple question: "What does resolved actually mean?"

Customer service said: the customer confirmed they were satisfied.
Operations said: the ticket was closed in the system.
Management said: the issue wasn't re-raised within 30 days.

Three different answers. Same word. Same organisation.

They had been measuring "resolution rate" for four years. Three teams, optimising for three completely different outcomes. The dashboard looked fine. The customer experience was quietly falling apart.

No amount of AI would have fixed that. When you automate an unclear process, you just get confusion at scale, faster.

We spent six weeks getting everyone aligned on a single definition before any technical work started. The improvement that followed was significant, and almost entirely nothing to do with technology.

Here's the question worth sitting with today: how many different definitions of your most important KPI exist across your organisation right now?

You might be surprised by the answer.

08/04/2026

The real ROI on AI? It almost never comes from the deployment itself.

Every AI project I've worked on has delivered its most valuable output before a single piece of automation was built. It happens in the mapping stage, during the process audit, in that moment when someone in the room says "wait, that's not actually how we do it anymore."

And suddenly you realise the whole system was designed around a process that changed 18 months ago.

I call it the discovery dividend. You can't deploy AI into a process you haven't properly mapped. And the mapping forces a level of clarity that most organisations should have had years earlier.

The companies I've seen extract the most long-term value from AI didn't just implement a tool. They used the implementation as an excuse to redesign how work actually happens.

The tool is the catalyst. The process redesign is the prize.

Honestly, most businesses have at least one process running on pure habit and historical accident. Nobody's questioned it in years because it technically still works.

Where's the most outdated process design sitting in your organisation right now? I'm genuinely curious what people are finding when they start looking.

07/04/2026

Here's something that catches almost every business owner out when they start automating.

The instinct is to go straight for your biggest headache. The process that eats time, causes stress, and has everyone frustrated.

But that's usually the wrong place to start.

The most painful processes are painful for a reason. They're complicated, full of exceptions, and often held together by one or two people who just "know how it works." Nobody's ever written it down properly. Automating that first is a recipe for building something that breaks constantly, or spending months mapping a process that exists mainly in people's heads.

The better move: start with whatever is high volume, low complexity, and already well documented. Not the most painful thing, but the most automatable thing.

Small wins matter. They show your team what automation actually looks like in practice (usually different from what they expected), and they build the internal case for tackling the harder stuff later.

Sequence matters as much as selection.

So here's a genuine question: what's the most automatable process in your business that nobody's really looked at yet? I'd be curious to hear what comes to mind.

Before you buy another AI tool, answer these 5 questions.I've started a lot of engagements over the years, and these are...
07/04/2026

Before you buy another AI tool, answer these 5 questions.

I've started a lot of engagements over the years, and these are the ones that almost always change everything:

1. Where does your data actually live, and who owns it?
2. Which decisions do you want AI to support, and which ones must stay human?
3. What does "good output" look like, and who signs off on it?
4. Do you have someone internally who can catch when the AI is wrong?
5. What happens to your process if the tool goes down tomorrow?

Most teams skip straight to the demo. They pick a platform, run a pilot, wonder why adoption stalls.

The answers to these questions don't just prepare you for AI. They reshape the entire scope of what you actually need, sometimes dramatically.

I've seen companies realise mid-conversation that they were solving the wrong problem entirely.

That's not a failure. That's the work.

Get the foundations right first. The tools are the easy part.

Before you buy another AI tool, answer these five questions.I've asked them at the start of every engagement for the pas...
07/04/2026

Before you buy another AI tool, answer these five questions.

I've asked them at the start of every engagement for the past two years. The answers almost always change everything.

1. What decision does this AI need to support, and who makes it?
2. Where does your data live, and is it clean enough to trust?
3. If the AI gets it wrong, what's the cost?
4. Who owns the output when it lands?
5. What does success look like in 90 days, not in theory?

Most teams skip straight to the technology. They pick the tool first, then work backwards trying to justify it. That's how you end up with expensive software nobody uses.

These questions aren't complicated. But sitting with them honestly, before anything is scoped or signed, saves months of rework.

The readiness isn't about whether your team can use AI.

It's about whether your business is set up so that AI actually has something solid to work with.

Get that right first. Everything else gets easier.

03/04/2026

A client once told me they hired me to "implement AI."

That's not what they actually paid me for.

What they paid for was clarity. Someone to look at a messy operation and say: here's what's really broken, here's the order to fix it, here's where technology helps and where it doesn't.

AI implementation is step four or five in a six-step process. I spend most of my time on steps one through three: diagnosis, definition, and design.

Most consultants skip straight to the exciting part. They demo the tool, map it to a use case, run a pilot, and then they're gone. The client is left with technology they half-understand, a process they didn't quite fix, and a metric nobody thought to define.

I'm slower than that. More expensive too. Because I'm solving the actual problem, not just the visible one.

There's a mindset I keep coming back to: I don't solve problems. I build systems that make problems impossible.

So tell me, what's the real problem you're paying someone to solve right now? Not the surface one. The one underneath it.

Most companies I work with have 12 dashboards and no idea what to do on Monday morning.That's not a data problem. That's...
02/04/2026

Most companies I work with have 12 dashboards and no idea what to do on Monday morning.

That's not a data problem. That's a measurement problem.

When everything gets tracked, nothing gets decided. The dashboards multiply, the meetings fill up with charts, and somewhere in all that noise, the actual decision quietly gets postponed.

Good measurement isn't about coverage. It's about consequence.

Here's the test I use with every organisation: for each metric on your dashboard, ask one question. "If this number moves, what do we do differently?" If you can't answer that in one sentence, the metric isn't measurement. It's decoration.

Strip it out.

What you want is a small set of numbers where the action is obvious before the data even arrives. The threshold is set in advance. The response is agreed. The decision almost makes itself.

Fewer numbers. Clearer ownership. Faster action.

That's the whole framework, honestly. The organisations getting the most from their data aren't tracking more. They're tracking less, but with genuine intent behind every single number they choose to watch.

01/04/2026

"Let's just try AI on this" is one of the most expensive phrases in business right now.

Not because trying is wrong. Because without a defined baseline, you will never know if it worked.

I have walked into organisations six months after an AI tool was rolled out and asked: "What were you measuring before, and what are you measuring now?" The room goes quiet.

Before any AI pilot, answer three things: What does the process produce today? What would measurably better look like? How will you know you got there?

Those questions take forty minutes. They are worth more than any subscription.

What does your current AI evaluation process actually measure?

30/03/2026

Friday reflection.

Three years ago I started ending every client meeting with one question: "What's the one thing that would make this irrelevant in six months?"

Not because I had a framework for it. Just seemed useful.

Over time that single question became a standing agenda item. Then it shaped how we structured quarterly reviews. Then it changed the way clients hired, planned, and allocated budget.

One question. Asked consistently. For three years.

It didn't feel like building anything. It felt like a habit. But habits, repeated in the right direction, quietly become infrastructure.

That's the thing about small wins. They don't announce themselves. You don't notice them compounding. You just look back one day and realise the ground has shifted, and you're standing somewhere completely different to where you started.

No hustle story here. No 4am wake-ups or grind montages.

Just one question, asked the same way, every single time.

What small thing have you been consistent with this week, even without knowing why it matters yet?

Address

50 Princes Street Ipswich
London
IP11RJ

Website

Alerts

Be the first to know and let us send you an email when Gary Does Strategy posts news and promotions. Your email address will not be used for any other purpose, and you can unsubscribe at any time.

Shortcuts

Share

Category