There Is No Shortcut to AI Success

Everyone wants the fast AI win. The agencies that will actually get one are the ones willing to do the slow work first.

At Signal78, we talk with a lot of agencies about our platform. Two questions come up more than any others. The first: “When can we have the whole thing?” The second: “Which of our current tools can we get rid of when we sign up for Signal78?”

Both are fair questions. Neither is the right first question. They treat AI as a purchase rather than a practice — something you install, turn on and get immediate value from, instead of a platform you build your future business on.

Recently, a pilot client decided not to move forward with us. Their reasoning was candid: they don’t see enough value in the platform as it stands today, but they can clearly see where it’s going. I respect the honesty. But I keep coming back to what they walked away from — not a finished product, but a real hand in shaping what that product becomes. A chance to build it around their own agency, their own knowledge, their own way of working.

They wanted the done version. Most people do.

I understand the instinct. When you’re running an agency, every hour and every dollar has to justify itself, and “this will pay off later” is a hard sell to a CFO. But wanting the finish line doesn’t move it any closer.

The Magic Wand Problem

Part of this is how AI gets sold. Business leaders are buried in pitches — every vendor, every platform, every newly AI-powered feature promising that things will simply get better the moment you turn it on. Faster proposals. Cheaper production. More output from fewer people. The demo always works. The wand always waves.

So leaders buy. And then they buy again. What many end up with is a stack of disconnected tools that don’t talk to each other, don’t share context and don’t add up. One plus one doesn’t equal two, let alone three.

Meanwhile, the expectation of immediate payoff never changes — even for technology that is, by any honest measure, still being built while we use it.

That’s the trap: spending heavily on systems that don’t connect, while demanding instant returns from all of them. It produces exactly what you’d expect. Disappointment, followed by a quiet conclusion that “AI just isn’t working for us yet.”

It’s not that AI isn’t working. It’s that nothing underneath it has been built.

AI Is a Foundation, Not a Feature

The agencies I see making real progress have one thing in common: they treat AI as a foundational layer, not a bolt-on. And before they set that foundation, they did the strategic work of deciding what it needed to hold up.

That work isn’t glamorous. It means getting clear on what your agency actually knows and where that knowledge lives. It means deciding which workflows matter most, and which ones you’d redesign from scratch if you could. It means a leader — a named person, not a committee — owning each initiative. And it means the people at the top using the tools themselves, visibly, so everyone else knows it’s real.

It takes time. It takes money. It takes effort that won’t show up as a win in the first month. There is no version of this where you skip the foundation and still get the building.

People Want the Real Work More Than We Think

Here’s the part that keeps me optimistic.

Yesterday I presented at the Mirren AI Forum. I didn’t give a big-picture talk about the future of AI. I spent the session showing people, step by step, how to build a skill in Claude Cowork that writes client scopes, estimates and proposals. Where to save the files. How to build the layers of the skill. How to get outputs that actually sound and look like your brand.

It was detailed. It was specific. It was, by design, not magic.

And the response was one of the strongest I’ve ever had. One attendee told me it was the best presentation of its kind they’d ever seen.

That tells me something. The appetite for shortcuts is real — but so is the appetite for the truth. When you show people what the work actually looks like — not the demo, the work — they lean in. They don’t want another magic wand. They want to know how the thing gets built, so they can build it themselves.

If you want to build it yourself, the full deck is free to download — complete with step-by-step instructions.

The Race Is the Point

Every leader wants the finish line. The win, the savings, the hours handed back to the team. Those things are real, and they’re within reach. But they come at the end of the race, not instead of it.

Nobody breaks the tape at mile one. The agencies that win with AI won’t be the ones who found the shortcut. They’ll be the ones who stopped looking for it.