AI - Add-on or Redesign
- Ed Fowkes
- May 8
- 6 min read
Today I’d like to talk about something in an area where Felix and I get to help our community a fair bit. The current climate, the runway used by AI to take off, aligns my interest and work in structure and systems with Felix’s skills in automation and process coding. The blend of our two interests has proved super useful to founders, for which we are both grateful to them and happy that we have a valuable service to bring. We’ve now spoken to a good number of entrepreneurs, helpinhg them resolve their bewilderment of where to turn when it comes to AI.
Business owners approach me, half curious, half concerned, typically saying something along the lines of: “We know we should be using AI… we just don’t know where to start.”
This is a pattern I’ve been seeing repeat with increased frequency over the last 12 months.
Of course we think it’s a tooling problem.
But it isn’t. Things rarely are. You can't always get what you want. This is another clarity problem.
An add-on or a redesign?
Much of the current conversation around AI in business still sits at the surface level. Which tools to use. Which subscriptions to buy. Which prompts get the best outputs. The concept is that adopting this tool or that will transform your business. Well… a restaurant doesn’t improve overnight because they buy a better oven. What I’m trying to say there is that the real opportunity here sits a little deeper.
Mark Cuban recently pinpointed the areas he considers are most exposed to AI disruption. He cited administrative work, customer support, content creation, coding and data analysis.
Most read that as a warning. As founders, we need read it as markers placed on a map. Handles on which to apply leverage. These are roles under threat, sure, but they are also roles that are ready to be re-engineered. That is a process of evolution and it brings forth possibility. So, keep ya head up!
A short story from the field
Not long ago, I sat with a founder running a service business. It was just north of £2m in revenue. Good margins. Strong reputation. A very capable team. But something wasn’t quite working.
They had recently invested in marketing. The leads were up and enquiries were flowing. Activity seemed to be everywhere. But in reality, revenue had hardly changed.
The instinct was kind of predictable: We need more leads.
Well I encouraged that we step back.
We mapped the flow.
Traffic → Conversion → Delivery → Profit.
Within about 20 minutes, the issue revealed itself. It wasn’t at the top of the funnel, but in the middle.
Sales conversations were inconsistent. There were too many options being presented, too much explaining and a distinct lack of structure in how decisions were guided. Nothing to do with AI. Everything to do with clarity.
We simplified the offer, reducing the cognitive load on the buyer. We tightened the process.
Conversions doubled within a month. With the same leads and the same team.
Then we stepped in with AI.
We introduced it not to fix the problem but to scale the solution. Automating follow-ups. Structuring responses. Analysing call patterns. Supporting consistency.
AI didn’t save the business. Clarity did.
AI amplified it. The challenge then became capacity! Oh for Spring when we test and measure…
Where AI creates efficiency
If we strip away the noise, it becomes clear that there are 4 core areas where AI currently tends to create efficiency:
1. Time compression
Tasks that once took hours can now take minutes. Admin, research, summaries, scheduling, reporting.
2. Decision support
Patterns become visible. Data becomes usable. And perhaps most impostantly, the insight arrives earlier.
3. Consistency of execution
Processes that require elements of variation become repeatable. Outputs are structured. Variability reduces.
4. Scale without linear cost
Content, communication and support can now expand without a proportional headcount.
Each of these sounds obvious. But their impact is rarely considered systemically.
When you compress time, improve decisions and standardise execution, something else happens.
The bottleneck that used to hold you back moves to a different area.
Thinking and focus
The shift most founders are underestimating is that the constraints for them is less about effort. It has become about focus and thinking.
As AI removes friction from execution, the limiting factor becomes:
What are we fighting for?
What matters most right now?
What does good actually look like?
This kind of judgement is actually the thing that successful entrepreneurs have always been good at. Those of us with an abundant mindset have always managed to leverage others into action, and achieved strangely unbelievable things in the process.
But the Spring test and measure stage is all important. And apologies I am thinking about Spring a good deal right now. It is the Spring season for the property businesses I’m helping. Soo why test and measure?
At low speed, you can get away with poor technique. You can correct mid-turn. Recover from mistakes.
At high speed, everything is amplified. Small errors become large ones. Precision matters.
AI increases the speed of business, putting us into Summer very quickly.
Clarity is now non-negotiable.
Don’t overcomplicate it
Most founders don’t actually need a grand AI strategy. I know that’s what the gurus are selling, but really?
What we do need is a better way of thinking about our business.
Here’s a simple approach I’m seeing work:
Start with friction, not tools
Diagnose where work slows down, where tasks repeat, where things get dropped.
That is your entry point.
Define the outcome before the action
Before using AI, be explicit. What does success look like? A good proposal. A clear email. A structured report.
Without this, AI will produce more noise.
Remove before you automate
Efficiency applied to the wrong activity is still waste.
If a task shouldn’t exist, don’t speed it up. Eliminate it.
Focus on one system at a time
Sales follow-up. Customer onboarding. Internal reporting.
Pick one. Improve it. Then move on.
Use AI to support thinking, not replace it
Ask it to analyse patterns. Challenge assumptions. Offer alternatives.
Not just to produce outputs. Thinking remains your most valuable skill.
Review and refine
Every system you improve should be revisited. What changed? What improved? What broke?
This is where most of the value sits.
None of this is complex.
But it does require discipline.
The hidden risk
There is a trap emerging, subtle and quietly in the background.
As AI makes it easier to produce, businesses risk producing more of the wrong things.
More content that says little.More communication that lacks intent.More activity that feels like progress.
We’ve seen this before.
Being busy masquerading as effectiveness.
And we see it in socials, where there’s a lot more content, but most is drivel.
AI can, is and will accelerate that. In spades.
It can eliminate it if we’re careful.
The difference lies in how deliberately it is applied.
The opportunity in 2026
The advantage today is not access to AI. Everyone has that.
The advantage is clarity of application.
A founder who understands:
Where their business actually creates value
Where time is being lost
Where decisions are being delayed
Where customers experience friction
…will use AI to remove constraints and improve flow.
The rest will simply layer tools on top of confusion.
And confusion, when scaled, can become very expensive.
When I look at businesses who adopt AI well, there is a common thread.
They are not chasing capability, but refining direction.
AI is a multiplier. But it only multiplies what’s already there. Which leads to a useful question to sit with:
If your business became twice as fast tomorrow…
What would double in speed?
would it become twice as effective, or twice as chaotic?
Musical Eggs
“You Can't Always Get What You Want” - You Can't Always Get What You Want by Rolling Stones
“keep ya head up” - Keep Ya Head Up by 2Pac. “What matters most” - What Matters Most by Ben Folds “And you may ask yourself “Well, how did I get here?”” - Once In A Lifetime by Talking Heads
P.S.
AI is going to speed things up. If you apply it to your system as is, what will happen?
Comments