The Decision Backlog

THE AI AUDIT

You know that feeling on Sunday night where you're technically not working, but you're also not resting? Where your brain keeps circling the same six things like a plane that can't get clearance to land?

That isn't burnout. That's a decision backlog.

I started noticing this pattern about two years into consulting. I'd sit down with an owner doing a million and a half a year, sharp as anybody, and I'd ask what was blocking growth. They'd give me a strategy answer. New market. Better funnel. Hiring problem. Then we'd actually walk through their week, and it was never any of that. It was that they hadn't decided whether to raise prices. Hadn't decided whether to fire the account manager everybody knew wasn't working out. Hadn't decided whether the second location was a real plan or a daydream.

None of those decisions were hard, exactly. They were just unmade. And every single one was drawing interest.

Here's the thing nobody tells you about running a business. The cost of a bad decision is usually recoverable. You pick the wrong vendor, you eat six months and switch. The cost of an unmade decision compounds forever, because it doesn't just sit there. It blocks every downstream decision that depends on it, and it takes up working memory you needed for something else.

So today we're going to audit the backlog. Not your tasks. Not your calendar. The decisions.

Grab something to write with. Real paper works better here, and I'll tell you why in a minute.

Set a timer for twelve minutes and write down every decision you're currently avoiding. Not tasks. Decisions. The distinction matters. "Update the website" is a task. "Decide whether we're repositioning or just refreshing the copy" is a decision. Tasks can be delegated. Unmade decisions cannot, and that's exactly why they pile up on your desk specifically.

Most people get eight or nine in the first three minutes and then think they're done. They're not. The first nine are the ones you're consciously aware of. Keep going. The ones that come out around minute eight are usually the expensive ones, because you've buried them well enough that you stopped noticing the weight.

Prompt yourself with these if you stall out. What have I said "let me think about that" to more than twice? What have I moved to next week's list more than three times? What conversation am I dreading? What am I hoping resolves itself? What did I promise somebody I'd get back to them on?

That last one is a killer. Go through your sent mail from the last sixty days and search for "circle back," "get back to you," and "let me look into it." I've watched owners find eleven open loops in under five minutes doing exactly that.

Paper matters because writing by hand is slower than thinking, and that gap forces you to actually commit to the words instead of typing a vague placeholder you'll never decode later.

You should end up somewhere between fifteen and forty items. If you have fewer than twelve, you didn't dig. If you have more than fifty, congratulations, you've found the actual reason you're exhausted.

Now we sort, and this is where AI earns its keep, because sorting is exactly the kind of tedious clarity work that humans are bad at when they're emotionally involved. And you are emotionally involved. That's why the pile exists.

Type your list into whatever assistant you use. ChatGPT, Claude, Grok, whichever one you already pay for. Don't overthink the tool here. Then run this:

The output is going to surprise you. In my experience about a third of any decision backlog is NOT REAL. It's anxiety that got promoted to strategy because it kept showing up. Deleting those alone frees up an unreasonable amount of mental space.

Another quarter is usually NOT MINE. Those are the ones where you've accidentally made yourself the bottleneck on something a competent person could handle if you'd just written down the rule. "Should we comp this client's shipping" is not a decision you should be making forty times a year. It's a policy you should write once.

The REVERSIBLE pile is the good news pile. Those you just decide. Right now, today, in the next twenty minutes. Reversible decisions do not deserve deliberation, and treating them like they do is one of the most expensive habits in small business. If you can undo it in thirty days for a grand, flip a coin if you have to. The information you get from doing it beats anything you'd get from thinking about it longer.

That leaves EXPENSIVE. Usually three to six items. Those get real work, and we'll handle them in a second.

Before the expensive pile, deal with NOT MINE properly, because this is where the compounding actually happens.

For every item in that bucket, you're going to write a rule instead of an answer. A rule sounds like this: "Discounts up to 10% are approved by the account manager without asking. Anything above 10% comes to me with the margin math attached." That's it. Now that decision never reaches your desk again.

Here's the prompt I use to draft these fast:

Run that for every NOT MINE item, stack the outputs into one document, and you've just built the first draft of an operations manual as a side effect of clearing your head. That document is worth more than most of the courses people buy.

If your team lives in a shared workspace already, drop it there today. Not next week. The rules only work if people can find them.

THE AI WORKFLOW BLUEPRINT • $47

The exact system I use to turn recurring decisions into written rules, and recurring tasks into automations that run without supervision. Twelve workflows, the prompts that build them, and the decision thresholds that keep you out of the weeds. Built for owners who are tired of being the bottleneck in their own company.

Now the three to six that actually matter.

The mistake here is treating an expensive decision like it needs more thinking. It doesn't. It needs better inputs. You've probably been chewing on these for weeks precisely because you're missing information and haven't admitted it.

So instead of asking AI what you should do, ask it what you'd need to know before you could decide. This one prompt has saved me more time than any other on this list:

That last question is the one that breaks the logjam. Most people have never priced the delay. When you see that dragging your pricing decision another month costs you eleven grand in margin, the decision gets a lot easier to make.

Then run the pre-mortem, which is the other half:

You're not looking for a verdict. You're looking for the tripwires. If you know the month-one warning signs for both paths, you can pick either one and just watch the gauges. That's what makes an expensive decision feel reversible even when it isn't fully.

Clearing the pile once feels great. It also doesn't last, because you didn't fix the intake.

Decision backlogs form for three reasons and every one has a fix.

The first is that you don't have the data in front of you when the question arrives, so you defer. Fix that by knowing where your time and money actually go, and knowing it without a research project. I use Rize for time because it tracks automatically and I stopped pretending I'd remember to hit a timer. When somebody asks whether we should take on a project type, I can see what that project type actually cost us last quarter instead of guessing.

The second is that the context lives in conversations you had and then forgot. Half your unmade decisions were already discussed on a call in April. Recording and searching your meetings solves more of this than people expect. Fathom sits on the call, writes the summary, and lets you search across everything you've ever said, which means "what did we agree on with that vendor" becomes a ten second lookup instead of a thirty minute archaeology dig.

The third is that you have no scheduled slot for deciding, so decisions get made in the cracks between other things, badly, or not at all. Put a recurring ninety minutes on Monday morning called Decisions. Not "planning." Not "strategy." Decisions. Bring the backlog. Leave with an empty one.

The first time you run this, block two hours. It'll feel like a lot of time for something that produces no visible output, and then Tuesday will be strangely quiet in your head and you'll understand.

Second week takes forty minutes. By week four you're at fifteen, because the intake is fixed and you're only clearing what accumulated.

One honest warning. You're going to find at least one decision you've been avoiding for over a year, and it's going to be about a person. It usually is. AI will not make that one easier and it shouldn't. What it can do is stop the other thirty items from crowding it out so you can finally give it the attention it needed twelve months ago.

That's the actual win here. Not efficiency. Clarity about what deserves your worry and what was just loitering.

Run the audit today. Twelve minutes of writing, twenty minutes of sorting, and a list of rules that keeps half of it from coming back.

THE AI BUSINESS ACCELERATOR • $97

Eight weeks of building the systems behind everything in this issue. Decision rules that run without you, automations that handle the recurring work, and a stack that fits your business instead of the other way around. You bring the business, we build the machine around it.

Jordan

The AI Newsroom | Jordan Hale | ainewsroomdaily.com