The Boring Quarter

THE SUNDAY STRATEGY

There's a pattern I've watched play out enough times now that I can usually call it in the first conversation.

A business owner gets serious about AI. They do the reading, they try the tools, they get genuinely excited. And then they go build the most interesting thing they can think of. An agent that handles inbound. A content system with six moving parts. Something with a dashboard.

Six months later it's off. Not because it failed exactly, but because it needed attention, and attention is the scarcest thing in a small business. Meanwhile the same business is still manually copying data between two systems forty times a week, still writing the same five emails from scratch, still losing an afternoon a month to a report that could have built itself.

The exciting project got the energy. The boring one got the results, in the businesses that ran it.

I want to make the case for a deliberately boring quarter, because it's what I'd recommend to most people reading this, and because it's a genuinely hard sell against the alternative of doing something that feels like progress.

The argument is about maintenance, not about ambition.

Every system you build has two costs: the cost to build it and the cost to keep it alive. Most people price the first and ignore the second, which is how you end up with a business full of half working automations that everybody has learned to route around.

Boring systems have near zero maintenance cost. A pipe that moves a customer record from your form to your CRM does not need attention. It does not drift. It does not produce a surprising output one day that nobody notices for a week. It works, silently, for years.

Interesting systems have high maintenance costs, and worse, the cost is unpredictable. Anything with judgment in it needs checking. Anything with several dependencies breaks when one of them changes. Anything clever enough to impress people is clever enough to fail in ways that take a while to diagnose.

In a business with abundant technical attention, the interesting things are worth it. In a business where the owner is the technical attention, they usually aren't. That's not a counsel of low ambition. It's arithmetic about where a scarce resource goes.

The second argument is compounding. Boring systems stack. Each one you build makes the next one easier because the data is cleaner and the connections already exist. Interesting systems tend not to stack, because each is bespoke.

Ninety days. Three phases. No new tools unless the phase requires one.

Phase one, weeks one to three: find and fix the copying.

Every place a human moves data from one system to another with no judgment applied. This is the least glamorous work in the business and it's where the hours are.

You'll find more than you expect. In most businesses I've looked at, someone spends between two and six hours a week on pure data movement, and it's usually spread thin enough that nobody has ever noticed it as a single line item.

The fix for each one is a connector, and the build time per connection is typically under an hour once you're set up. Fifteen connections over three weeks is realistic and would return the whole quarter's effort by itself.

Phase two, weeks four to eight: templatise the repeated writing.

Every message your business sends more than twice a month in roughly the same shape. Quotes, follow ups, onboarding, the explanation you give every new client, the update nobody enjoys writing.

These are not automations. They're templates with the variable parts marked, stored somewhere reachable in five seconds. The AI part is minimal and sits at the edges: drafting the first version from your past examples, and adapting a template to a specific situation when the situation is unusual.

The value here is not time saved per message, though there is some. It's consistency and it's the elimination of the small dread that precedes writing something you've written forty times. Dread is a real tax on throughput and nobody puts it in a business case.

Phase three, weeks nine to twelve: make the reports build themselves.

Every number you look at regularly should arrive without you assembling it. Weekly revenue, pipeline state, whatever your two or three real operating metrics are.

Most small businesses have someone spending a chunk of a day per month producing a report that gets glanced at. Sometimes that person is the owner, which is the most expensive version.

This phase is last deliberately, because it depends on phase one. Reports build themselves easily when the data is already flowing to one place, and painfully when it isn't. Doing this first is how people conclude that automation is harder than it looks.

THE AI WORKFLOW BLUEPRINT • $47

The Blueprint is the boring quarter written down: the twelve workflows in build order, with the dependencies marked so you don't attempt phase three work in week two. It's the map I'd hand someone starting this on Monday.

Around week five, this stops feeling like progress.

The phase one wins were visible and satisfying. Phase two is templates, which is administrative work that produces no dopamine at all. Nobody congratulates you for having good templates. There's no moment where it feels impressive.

This is where most people abandon a boring quarter and go build something interesting, and it's worth knowing in advance that the urge is coming so you can recognise it as a predictable event rather than a signal.

Two things help.

Keep a running tally of hours returned. Not estimated, measured. When you automate a data handoff, write down what it used to take per week. The number gets satisfying around week six precisely when the work stops being satisfying, and it's the only thing that reliably carries people through the middle.

And do the visible thing last, not first. Phase three produces reports, which are the one output other people in the business will actually notice and comment on. Ending on that is better sequencing psychologically as well as technically.

Having watched a number of people attempt this, there's one failure that accounts for most of the abandoned attempts, and it isn't losing motivation in week five.

It's scope. People treat the boring quarter as a licence to fix everything, and then discover in week two that phase one has ballooned into a full systems rebuild.

Here's the discipline. In phase one you are not allowed to improve a process. You are only allowed to remove the human from a data movement that already happens. If you find yourself redesigning how leads get qualified while you're building the connector that moves them, stop. Write the improvement down somewhere else and build the connector as specified.

This feels wrong, because you can see the better process from where you're standing and it's frustrating to build a pipe to a broken destination. Do it anyway. The pipe takes an hour. The redesign takes three weeks and it will eat the quarter, and at the end you'll have one beautifully redesigned process and fourteen manual handoffs still running.

Fix the plumbing first. Redesign later, from a position where the data is already flowing and you can actually see what's happening.

The same rule applies in phase two. You are templatising the message you currently send, not writing the better message you wish you sent. Capture the current one, mark the variables, move on. Improving the copy is a separate project and a much easier one once the template exists.

And in phase three: build the report that answers the question you already ask, not the dashboard you've always wanted. One number per question. If nobody has been asking for a metric, it does not go in the report, no matter how interesting it would be to know.

Scope discipline is the whole difference between a quarter that finishes and a quarter that becomes a story about how you tried to fix your operations once.

Being honest about the trade, because there is one.

You give up a quarter of experimenting with the frontier. If something genuinely significant arrives in those ninety days, you'll adopt it three months later than the people who were paying attention.

I think that's a good trade for most small businesses and I want to be clear about why, because "ignore the new stuff" is bad advice generally.

The reason is that the cost of being three months late to a genuinely useful capability is small, and shrinking. Tools get easier over time, not harder. The second mover gets better documentation, fewer bugs, and the benefit of watching other people discover the failure modes. Being early is only decisively valuable when the capability is a durable competitive advantage, and almost nothing available off the shelf is, because your competitors can buy it too.

Meanwhile the cost of never doing the boring work is permanent. It compounds in the wrong direction. The data stays messy, so every future thing you build is harder, so you build fewer things, so the gap widens.

The frontier will still be there in ninety days, and you'll be in a much better position to use it.

Ask three questions honestly.

Does anyone in your business regularly move data by hand between two systems? If yes, phase one alone justifies the quarter.

Can you produce your main operating numbers in under ten minutes without asking anyone? If no, phase three is worth more than anything else you could build.

Is there an automation you built in the last year that's currently switched off or being routed around? If yes, that's a strong signal you've been building interesting things rather than durable ones, and a boring quarter is the correction.

Two yeses and I'd run it. Three and I'd start tomorrow.

Here's the whole argument in one paragraph, for the version of you that will be busy on Tuesday.

The businesses that got real value out of AI over the past two years were mostly not the ones doing clever things. They were the ones that quietly removed the friction from work they were already doing, in an order that let each fix make the next one easier, and then kept going after it stopped being fun. That's it. That's the strategy. It doesn't photograph well and nobody writes case studies about it, and it's still the best available use of the next ninety days for most people reading this.

Pick the quarter. Do the unglamorous version. Let somebody else have the interesting one.

THE AI BUSINESS ACCELERATOR • $97

Eight weeks of running the boring quarter with someone who has done it before. We find the copying, kill it, build the templates, and get your numbers assembling themselves, in the right order, documented so it survives you losing interest.

Jordan

The AI Newsroom | Jordan Hale | ainewsroomdaily.com