AI Cost Me More Than It Returned. For Months.

Almost everything written about AI promises an immediate payoff. Save ten hours a week, get back to the work that matters.

That is not how it went for me.

When I started, I could not have told you whether this was a real tool or just the thing everyone happened to be talking about. I decided to find out. What followed was months of investment before I got much back.

That shape is worth describing honestly, because this curve does not go up and to the right from day one.

Three things I built

I did not start with anything big. The first things I built were lightweight, and they were less about output than about finding out whether I could trust this at all. They worked. That familiarity is what made me willing to commit real time to something larger.

Three projects came out of that, and they turned out to be one arc rather than three experiments.

A new website, built with AI as the developer. I am not a developer and never have been. Before this I had a WordPress site I put up fast. It plugged a hole, but updates carried enough friction that they got deferred, and I was never excited to send anyone to it. That is a cost that shows up on no invoice. Now the site is the single source of truth for everything I am doing.

A standing financial advisor for my own business. Finance is not where I am strongest. That is a survivable weakness in a mature company and a real risk in a new one. So I built something that gives me a quick expert read and the tactical aids to stay current. It does not replace an accountant. It means I am not in the dark between conversations, and my customers never feel it in billing or collections.

Automated prospect research and qualification. When I told people I was starting a C12 chapter in the Pacific Northwest, the first question was almost always the same. How are you going to find Christian CEOs in Seattle, of all places?

Fair question. There is no list to buy, and most believers keep their faith fairly quiet at work. What runs now is a patient search through public signals for people who might overlap with the leaders I am built to serve. It is a first pass, not a filter, and it never replaces the conversation. It is work I could not do by hand at any real volume.

The part nobody writes about

There is a stretch in every build like this where you have already spent more time than the old way would have cost and the new thing is not ready to carry the load. That is not a sign you did it wrong. It is the middle. But it does not feel like the middle while you are in it. It feels like evidence.

The website was the hardest of the three. AI carried about 95 percent of the technical load, work I would otherwise have handed to a developer I did not have a relationship with.

But the technical part was not what made it hard.

Once the technical excuse was gone, what was left was the part I could not delegate. Deciding what I actually wanted to say. In a world this full of content, I wanted the site to be authentic, courageous, and honest. Removing the barrier did not make that easier. It took away the last reason to postpone it.

Then it turned

The return arrived in three forms.

Time came back in blocks, not minutes. Entire categories of work stopped requiring me, which changes what a week can hold.

I stopped shopping for solutions and started building them. Nothing off the shelf did quite what I needed, and I had spent years settling for close enough. The distance between “I wish this existed” and “this exists” collapsed to days.

The targeting got sharper, not just faster. Better research produced better understanding, and better understanding produced messages that were actually about the person receiving them.

The problem I have now is a better one than the problem I started with. It is no longer whether this works. It is which hurdle to point it at next.

You already know this curve

Every one of us has been through this shape before. Hiring the first real controller. Installing an operating system in the business. Building an actual sales process instead of running on referrals and hope.

All of those cost more than they returned for a stretch. All of them are reasons the business works now.

AI is that kind of investment, not a shortcut. Approach it as a shortcut and you will quit somewhere in month two, because a shortcut that takes four months is not a shortcut, and you will correctly conclude you were sold something.

Approach it as a build and the timeline stops being a problem. It becomes the plan.

The thing it could not do

For all of it, the tools never once told me whether I was building the right thing.

They executed. They accelerated. They never pushed back on a bad assumption, and never noticed when I was solving an interesting problem instead of an important one. That kind of correction only comes from people who know your business, know your character, and are willing to say the uncomfortable thing.

I get that from the leaders I sit with every month. Christian CEOs and owners working through their own version of this same question, willing to ask me why before they ask me how. They are the reason I built the right things instead of just building fast.

If you are starting, start small enough to build confidence before you commit real time. Decide the direction once. Measure at ninety days rather than weekly. Then find people who will tell you the truth about the direction, because none of us are meant to sort that out alone.

Peace,

Eric Smith, Principal Chair, C12 Seattle North