A practical, business-first approach to finding where AI can actually improve results.
AI can make individual tasks faster. But making one task faster doesn't necessarily make the business more efficient—or mean you're getting a meaningful return on your AI investment. That's the distinction that gets lost in many conversations about AI.
Businesses are being encouraged to find ways to use AI everywhere they can. Leadership teams start evaluating tools, employees experiment with ChatGPT and other AI products, and vendors demonstrate everything their latest technology can do. But there's a more important question to answer first: what does the business need to do better?
Maybe the goal is to complete existing work faster, increase capacity without adding staff at the same rate, improve customer service, reduce costs, eliminate errors, or make the organization easier to operate as it grows. Once you know the outcome you're trying to achieve—and how you'll measure success—you can examine how the work happens today and determine what actually needs to change.
Then you can decide where AI belongs.
The Problem With Adding AI to an Existing Process
Many business processes weren't intentionally designed from beginning to end—they evolved. Someone created a spreadsheet. Another employee developed a workaround. A new system was added. One department started doing things differently. Manual steps remained because that's how the work had always been done.
Eventually, what the organization calls a "process" may actually be a collection of systems, spreadsheets, emails, manual handoffs, individual knowledge, and workarounds pieced together over time.
Bolting AI onto that workflow might make individual tasks faster. But faster isn't always better. Imagine AI reduces a twenty-minute task to five minutes. That's useful, but if the work still waits for an approval, gets manually entered into another system, requires missing information, goes through unnecessary steps, or has to be corrected later, you've optimized one task without necessarily improving the business.
The better question isn't "Did AI make this task faster?" It's "Did we improve the overall process and achieve the business result we wanted?"
How Do You Improve AI ROI?
Improving AI ROI starts with measuring AI against a business outcome—not simply measuring how much the technology is being used or how much time it saves on an individual task.
A successful AI initiative might:
- Increase capacity or revenue.
- Reduce operating costs.
- Complete work faster.
- Reduce errors and rework.
- Improve employee productivity.
- Improve the customer experience.
- Allow the business to grow without increasing staffing at the same rate.
- Get more value from existing technology investments.
These are business outcomes. AI is one potential tool for achieving them. The right solution could involve AI, conventional automation, an existing business system, a process change, or a combination of them.
The question shouldn't simply be "How much time did AI save?" It should be "What measurable business result did the investment produce?"
Not sure where AI can actually make a difference in your business?
Start with the result you're trying to achieve, not another technology purchase.
The Vanderpan 7-Step Approach to AI That Delivers Results
A successful AI initiative should connect technology decisions directly to measurable business outcomes. Rather than starting with an AI product and looking for places to use it, I use a seven-step, business-first approach.
1. Define the Business Objective
What business result are we trying to achieve, and how will success be measured?
Start with the outcome. Are you trying to improve efficiency, increase capacity, reduce costs, improve quality, generate more revenue, improve the customer experience, or support growth?
Then determine how you'll measure whether you accomplished it. Defining success at the beginning gives you something much more meaningful than AI adoption to measure later.
2. Understand the Current Process
How does the work actually happen today?
Look beyond the documented procedure. Understand the systems, employees, handoffs, spreadsheets, emails, manual steps, and workarounds involved. Where does work wait? Where is information entered more than once? Where do errors happen? Which steps depend on someone knowing what to do next? Which decisions actually require human judgment?
The goal is to understand the real workflow—not the workflow everyone assumes exists.
3. Improve the Process
What should be eliminated, simplified, standardized, or redesigned?
Some problems don't need AI—they need a better process. Before automating something, determine whether unnecessary steps, duplicated work, bottlenecks, missing information, or inconsistent procedures can be eliminated.
One of the most valuable questions you can ask is: does this step need to be automated—or does it need to exist at all?
4. Identify the Right Technology
Where can AI, automation, or existing technology create meaningful improvement?
Now we're ready to talk about technology. AI may be ideal for some parts of the process. Traditional automation may be better for others. An existing business system may already have capabilities the organization isn't using. And sometimes AI won't be the best answer at all.
The objective isn't to maximize the amount of AI being used. It's to use the right technology in the right places.
5. Deploy and Drive Adoption
How do we turn the plan into the way people actually work?
Selecting technology isn't implementation. Employees need to understand the new process, how the technology should be used, what they're responsible for, and where human judgment remains necessary.
Training, adoption, security, data handling, responsibilities, and integration with existing systems all matter. Technology only creates value when it becomes part of an effective working process.
6. Measure the Results
Did we achieve the business outcome we defined at the beginning?
Return to the measures established in Step 1. Did capacity increase? Did work get completed faster? Did costs decrease? Did quality improve? Did customers receive better service? Did the organization achieve the expected ROI?
AI adoption isn't the success metric. Business improvement is.
7. Refine and Improve
What did we learn once the new process was put into practice?
Real-world use will expose things that weren't obvious during planning. Use those results to adjust the process, technology, training, and AI usage, then measure again.
That's why Step 7 leads back to Step 1:
Business Objective → Process → Improve → Technology → Deploy → Measure → Refine ↻
It's a cycle of continuous improvement—not a one-time AI project.
From AI Strategy Through Execution
A good recommendation that never changes how the business operates has very little value. That's why my role doesn't have to end with an assessment or technology roadmap.
I can work with leadership, employees, internal IT teams, vendors, and other technology partners to move the plan from recommendation through deployment, adoption, measurement, and refinement. That may mean implementing AI, improving an existing system, automating part of a process, redesigning the process itself—or deciding AI isn't the right solution.
The goal isn't to use more AI. It's to determine what the business needs to do better, choose the right technology to help accomplish it, and see the plan through execution.
What does your business need to do better?
Let's start there.
Whether you're already experimenting with AI or simply trying to determine where it could produce meaningful results, we can start with the business problem and work from there.
Comments