Strategizing your AI and tech for human interaction

For far too long, the traditional accounting firm model has been a little abrasive for the people it serves. Clients often feel like they’re dealing with someone hiding behind glasses and spreadsheets who doesn’t want to pick up the phone. 

This disconnect in the customer service relationship is what inspired me to start my firm, Flux Accounting. The firm is centered around one goal: Creating the best experience imaginable in accounting. 

To do this, we embrace technology and other strategies to make the experience seamless, allowing us the time to focus on bringing human interaction to the forefront. 

Here’s what’s worked for us, and what you can start doing, whether you have a week or a year.

Use AI like a scalpel, not a sledgehammer

A lot of firms get AI wrong because they either avoid it entirely or they try to deploy it everywhere just for the sake of using it. Both ways will take you in the wrong direction—and hold your firm and your clients back.

While traditional software put you in a box and you tried to expand what the box could do, AI is the opposite. AI can do almost anything, so it’s vital that you narrow the focus and put guardrails around it.

So before you roll anything out, figure out what it’s good at and where it falls short, and implement it with what would best serve your clients. You almost have to be surgical in this approach.

Where to start: Solutions to save time on document intake 

For us, the best first use case was the tax organizer. During busy season, we take in tens of thousands of documents, and they show up sloppily organized, with odd file names and formats. AI has gotten very good at identifying those documents, checking off exactly what each one is, extracting the information, and organizing it.

That, alone, saves a ton of time. We use Sorban for intake and Intuit® ProConnect™ Tax for preparation, where SmartReturn does a great job reading documents and pulling out the data.

Note that the questionnaire side of things isn’t really AI. For example, it may say, “Last year, these five things happened, so here are this year’s questions.” These are just programmed rules. Right now, AI’s real strength is document handling, and that’s exactly where I’d start.

Build quality control around the AI

The most common objection I hear is, “I have to check every single line, so it isn’t saving me any time.” That’s what happens when you treat AI as a trust-it-or-don’t decision.

Just like a human preparer needs a reviewer, AI needs one, too. It’s not binary, and this should come as a comfort because our expertise is still needed. 

We don’t implement anything until there’s a quality control layer on top of it. In practice, that means AI reads a document, extracts the income and withholding, and plugs those numbers into a work paper that reconciles them against something else. The cross-check tells you whether the numbers hold up, so you aren’t re-reading every line.

Once that’s in place, you no longer need to worry about trusting the AI. The system tells you when something looks off.

Pick a philosophy and stick with it

Right now, it feels like the wild west. Software vendors are using AI to ship products faster than ever, so you could have 50 tax organizers to choose from next month. That’s a recipe for decision fatigue, causing you stress and anxiety.

My advice is to choose a philosophy and commit to it, and avoid switching for the next shiny thing. For some firms, that philosophy is moving to cloud preparation in ProConnect. For us, it was getting very clean handoffs from a business return to the owner’s personal return because that happens thousands of times each season. Small, repeatable things going smoothly add up to a lot of saved hours.

Less is more: Simplify your tech stack

For our first eight years, we leaned heavily on APIs and connected every platform we used. Now our integration web is complicated enough that we’re heading the other direction. We want fewer vendor relationships: Run on QuickBooks®, run on ProConnect, and build whatever else we need around them.

You don’t have to build your own software, and I’d caution that it’s hard when it isn’t what you do for a living. The takeaway is that every extra tool is something to maintain, train people on, and fix when it breaks. Fewer tools that talk to each other well means fewer headaches.

Tell clients how you use AI, and don’t use AI for personal communication

Messaging is everything. Good ideas fail when leaders assume everyone will receive them positively. This is one of the biggest mistakes I’ve noticed. And when it comes to AI, clients are wary of it, and they have a right to be. 

When we talk about AI, we say the same thing every time: Our biggest concern is security. 

When we rolled out Sorban, we set up an automated email that goes out the moment a client starts submitting documents. It comes from our tax concierge, a real person, and basically says, “I’m your tax concierge. I’m a human. AI helps me organize your documents, but you’re not being left to the bots.”

That one email calms nerves, sets expectations, and keeps the relationship personal.

Vision, execution, and buy-in: Standardize your process to scale successfully

This is the part that’s easy to overlook. If you have talented people, they’ll want to work their own way. I get it. Some of them are snipers who knock out work left and right.

But as the firm grows, I can’t rely on knowing the steps someone took unless every file follows the same ones. When every client goes through the same 10 steps, anyone in management can walk in and solve a problem in 30 seconds because they know step 11 is where it got stuck. I don’t need three side conversations before I call the client back, and nobody has to be the single person who holds the whole process in their head.

This will save time and turn a complex problem into a simple process. It also means you can take a vacation, and your team can hand work off without a long explanation.

The catch is buy-in. You have to train your staff on why you do things the same way every time, and why exceptions are off the table. The messaging and vision from leadership matter just as much as the automation.

Putting your plan to action

After practicing some of these strategies, what does a short- and long-term plan look like for your firm? Here are some actionable steps you can take.

In the next 30 days:

• Practice with AI. Use ChatGPT, Claude, or whichever tool you prefer, and work on your prompting. Output quality tracks input quality, and prompting is a skill.
• Audit your intake. Write down every step from “client sends documents” to “preparer starts the return.” Mark the steps that are manual and repetitive.
• Draft your client message. Write a short note explaining how you use technology, lead with security, and introduce the human who’s still their contact.

In the next 90 days:

• Pilot an AI-assisted tax organizer on a small group of clients. Keep the scope narrow.
• Add a cross-check. For every extracted number, define what it gets reconciled against before it goes into the return.
• Write your process down. Document the steps for your most common return type, then train your team on why everyone follows them.

Over the next year

• Choose your philosophy. Decide what matters most for your firm, such as cloud prep, clean business-to-personal handoffs, or a fully integrated books-to-return workflow, and evaluate new tools only against it.
• Reduce your vendors. List every tool you pay for, note what each connects to, and cut the ones that add maintenance without adding value.
• Move toward a practice that fits you. Ask what you’d change if your practice management system worked exactly the way your firm does. Even if you never build anything, the answer will tell you what to demand from your vendors.

Putting your best foot forward 

All of us can do the everyday grind of busy season, make ends meet, and keep the same routine going each year. But, I would hope that you not only want more for your firm, but also for the clients you serve.

Your clients deserve that personal touch, along with the strategies in place to allow for it. So, take a surgical approach to technology, put quality controls in place, standardize how your team works, and keep a human front and center for your clients. You’ll end up with a practice that’s better for clients and sustainable for you.

Editor’s note: Chase Vowell is a featured speaker at Empower 2026, the tax season readiness virtual conference on November 10, 2026. Register now.