Most AI conversations in the MSP world are obsessed with speed.
Faster tickets. Faster resolutions. Faster reporting. Faster everything.
But when I spoke with Darryl Cresswell, the conversation landed somewhere far more useful.
Not everything needs to be automated.
And automating the wrong thing is how MSPs waste months thinking they’re building AI capability.
This wasn’t a conversation about shiny tools.
It was a conversation about discipline.
I’ve known Darryl’s journey for a while, and I respect it deeply. He started his company at 13 and has reinvented it through every major shift in our industry – from hardware and break/fix to cloud and now AI.
That kind of journey doesn’t give you theories.
It gives you instincts.
And those instincts matter right now.
One of the first things Darryl said separated him from the noise:
“Not everything needs to be automated and not everything should be automated.”
That line explains why so many MSPs feel busy but not better.
They automate workflows that look impressive.
They test copilots.
They experiment with agents.
And months later, nothing meaningful has changed.
No capacity freed up.
No margin gained.
No customer experience transformed.
Just activity.
And activity is not capability.
Here’s an infographic that conveys the message:
What Darryl reinforced is simple: AI only creates leverage when it’s applied to the biggest constraint in the business. Not the coolest use case. The real bottleneck.
That’s why one of their most impactful AI wins wasn’t flashy at all. It was auditing vendor invoices and reconciling billing. Painful, repetitive work and exactly the kind of constraint AI is meant to remove.
Another point Darryl was clear on: don’t learn AI on your customers.
If you practice AI in client environments, you introduce risk, inconsistency, and loss of trust. The smarter move is to start in-house. Make the mistakes internally. Build the playbook. Then take something proven to market.
That’s often how new services are born.
Not from selling AI.
From operationalizing it first.
Darryl also reminded me of something we tend to forget: this industry has never been evergreen, and it never will be. If you’re waiting for someone else to teach you, you’ll always be reacting.
The MSPs who win in AI are self-learners. Not because hustle is cool, but because being first to capability is how you become first to market.
And finally, there’s the power of peer groups. Being around other operators who are building and sharing what works compresses time in a way tools never will.
In AI, speed doesn’t come from technology.
It comes from exposure.
My takeaway from this conversation is simple:
AI capability isn’t built by automating more.
It’s built by automating what matters.
Because the goal isn’t to “use AI.”
It’s to remove constraints, free up capacity, and deliver outcomes you can repeat.
That’s how MSPs become thought partners in the AI era – not by being first to talk about AI, but by being the first to make it work predictably.
What’s next
In Part 2, I’ll go deeper into:
- How MSPs move from AI curiosity to customer confidence
- How education and structure become high-value services
This is the work.
This is the responsibility.
And this is exactly why AI By Design exists.
More to come next week. Stay tuned.
