You ever notice how some software has a progress bar so slow it basically gives you permission to leave? Not a quick spinner. A real one. The kind where a loan officer clicks "generate documents," watches the bar crawl, and decides this is a coffee situation.
That progress bar is the answer to this whole question. The first process you should automate is the one quietly eating the most time from your most expensive people, on work that needs none of their actual skill. Not the flashiest tool someone pitched you. Not the easiest quick win. It's the expensive, repetitive, judgment-free task hiding in plain sight.
Here's how we found one, and how you can find yours.
Start where the expensive time is being wasted
We once helped a mortgage company that had a small problem before we automated anything: no SOPs. None. So we wrote out every process, for every employee, across every department: origination, processing, underwriting, closing, and shipping.
When we analyzed all of it, one pattern stood out. The most valuable people in the building, the loan officers, were spending hours generating disclosures. That meant clicking next, next, next through the loan origination system, picking the right documents based on the loan type, and then waiting. The progress bar to generate a packet ran anywhere from three to fifteen minutes depending on its size. Long enough that the coffee break was real.
And it wasn't a once-per-loan problem. Most loans get re-disclosed at least once before signing, and then the closing disclosures go out at the end. So this was happening constantly, to the exact people you most want on the phone with borrowers.
Why we automated disclosures first
Disclosures became the first thing we automated, and we started with re-disclosure. When something on a loan changed, like the borrower's income or credit, the system used to fire an alert telling the loan officer to go re-disclose. We pointed that same trigger at an automation instead. Now the change fires, the loan re-discloses itself, the documents go to the client, and the loan officer just gets a confirmation that it's done.
Once that was solid, we expanded it to the initial disclosures, and eventually to the closing disclosures. Today every disclosure at that mortgage bank generates automatically.
This was the right first move for one reason: it removed an entire task from the most expensive people in the company, not a sliver of one. That's the difference between automation that pays for itself and automation that just looks busy. If you want the fuller version of that math, we broke down what manual work is really costing your business.
The mistake almost everyone makes first
Here's where most businesses go wrong. They go looking for the quick win, the shiniest tool someone throws their way, and then they're disappointed because the return is barely there.
Picture a voice AI receptionist to answer calls while your receptionist is at lunch. Great idea on paper. But you're still paying that receptionist to be there. You didn't save any money, and you didn't free up her time. You automated her lunch.
Or automatic AI draft replies to email. That's wonderful for one executive at home, dialing in exactly what counts as important and what counts as noise. It falls apart across a whole company, because everyone disagrees about what's important. Some people love the informational newsletters; others send them straight to spam. One set of rules can't serve all of them, so you end up writing rules per person, and even then you're saving the few seconds it takes to tweak a draft. You still review every one before it sends.
Because we're not there yet. You can't fully trust AI to speak on your behalf, or even to write a blog post like this one without a human reading it first. You think I didn't review this before it went out?
How to actually pick your first process
When I sit with a business to find the right first process, I'm scoring each one on a few questions:
- How many hands touch this process?
- Whose hands, and how much are they paid per hour?
- Does it actually require their skill set, or is it clicking next through screens, copying and pasting data, dragging and dropping files, saving and renaming things?
If a step or two genuinely needs a person's judgment, that doesn't kill the idea. You automate the rest and add a man-in-the-middle step: the automation pauses, asks the employee for the one thing only they can decide, then continues. It's exactly how the system that wrote this post worked, pausing to ask me each question before moving on.
Then the math is easy. Hours saved times the hourly rate you're paying people to do that work. That number is your return on investment, and it tells you which process wins. That's the heart of real business process automation: not replacing your people, but taking the mechanical work off the ones you pay the most.
But our process is too complex to automate
The single most common thing business owners get wrong is assuming their process is too complex, too full of edge cases, to automate.
The mortgage bank was certain disclosures couldn't be automated. There were too many variations:
- which state the property is in
- whether it was a purchase or a refinance
- whether there was a co-borrower
- whether that co-borrower was a spouse
- and a long list of others
It took time, but we locked in every one of those scenarios. Based on the data already sitting in the loan's fields, the automation decides which documents belong in the packet and which don't.
Edge cases aren't a wall. They're branches. I've built automations with five steps and automations with well over a hundred, precisely because there were that many routes to handle. Complexity isn't the reason you can't automate. It's usually the reason you should.
Where to start this week
If you're a business owner, of any size, here's the takeaway. Don't grab the easiest automation or the shiniest tool someone pitched you in a demo.
Instead:
- Sit down with your team.
- Map out all of your processes, if you haven't already.
- If you already have SOPs, pull them all up.
- Review them with one question in mind: where can I get the most savings?
The answer is almost always the same shape. An expensive person, doing something mechanical, over and over. Find that, and you've found your version of the disclosure problem. Then the real work begins: getting your team to actually adopt AI workflows is where most automation projects stall, and it deserves as much attention as finding the process.
If you want help finding yours, that's what we do. Book a free discovery call with Azuretech. We'll map your processes with you and tell you honestly where automation pays off and where it doesn't. No commitments, just clarity.
Common questions
What if part of the process needs a human decision?
You automate everything around it. A good automation can pause at the exact step that needs judgment, ask the person for that one input, and then keep going. You rarely have to choose between fully automated and fully manual.
How do I calculate the ROI of automating a process?
Multiply the hours the task takes each week by the hourly rate of the people doing it. That's the money currently going into work that software could handle. Weigh it against the cost of building the automation, and the decision usually makes itself.
Is my business too small to automate anything?
No. The test isn't your size, it's whether you've got an expensive person doing repetitive, mechanical work. A two-person company can have that problem just as easily as a two-hundred-person one.