Most Businesses Don’t Need More AI Tools. They Need Better Workflows

Every week, another AI tool promises to make businesses faster, smarter, and more productive.

And every week, companies add another tool to the stack.

ChatGPT. Copilots. Automations. AI agents. Meeting assistants. Content generators.

Yet inside many of those businesses, the same problems remain:

Slow follow-up. Messy handoffs. Inconsistent customer communication. Repetitive administrative work. Information trapped in different places. Reports nobody fully trusts.

So maybe the problem isn’t that businesses need more AI.

Maybe they need to redesign the work.

That distinction matters.

Because using AI is not the same as improving a business.

The Tool Is Not the Strategy

I use AI tools. I test them. I build with them.

But I don’t believe the tool should be the starting point.

When companies start with technology, the conversation usually sounds like this:

Which AI platform should we use?

What should we automate?

Which model is better?

What prompts should our team learn?

Can AI create more content?

Those are reasonable questions.

They’re just not the first questions.

I believe the better place to start is:

Where is the work breaking down?

Where are customers waiting?

Where does follow-up disappear?

Where is the team repeatedly recreating the same work?

Where does quality depend on who happens to be doing the task?

Where is information difficult to find?

Where are decisions delayed because nobody has the right context ready?

Where are opportunities being missed?

That is where the AI opportunity usually becomes visible.

Not in the software catalog. In the workflow.

AI Activity Is Not AI Value

This is one of the most important distinctions businesses need to understand.

AI activity is easy to see.

Someone drafts emails faster.

Someone summarizes a meeting.

Marketing produces twice as many posts.

A team automates a repetitive task.

A manager creates a custom GPT.

It looks like progress.

But ask a harder question:

What actually improved?

Did response time decrease?

Did follow-up become more consistent?

Did customers receive a better experience?

Did the team gain meaningful capacity?

Did quality improve?

Did sales opportunities move faster?

Did decision-making improve?

If nobody can answer those questions, the company may have increased AI usage without creating much AI value.

And those are not the same thing.

The number of people using AI is not a business outcome.

A Simple Example: Client Communication

Imagine a service business answering the same customer questions every week.

Questions arrive through email, forms, calls, DMs, and internal messages.

Employees repeatedly write similar responses.

Some are excellent.

Some are rushed.

Some leave out important information.

And much of the process depends on whoever happens to respond.

The tool-first solution is obvious:

“Let’s use AI to write the replies faster.”

That may save time.

But it doesn’t necessarily fix the workflow.

A workflow-first approach asks different questions:

What questions repeat most often?

Where should approved information live?

What information is required before someone responds?

What can be standardized?

What can AI prepare?

Where is human review necessary?

What should happen after the response?

How will we know the process improved?

Now AI has a defined role.

It might classify the inquiry.

Retrieve approved information.

Prepare a first response.

Suggest the next follow-up.

Summarize context for the employee.

But AI is no longer the strategy.

Improving the customer communication workflow is the strategy.

AI is one capability inside it.

Don’t Automate the Mess

There is another mistake I see businesses making:

They find something repetitive and immediately try to automate it.

But repetition does not automatically mean a process is good.

Some workflows are slow because nobody has defined them properly.

Some contain unnecessary steps.

Some exist because “that’s how we’ve always done it.”

Some are inconsistent because the business has never established a standard.

And some probably shouldn’t exist at all.

So before asking:

“Can AI automate this?”

Ask:

“Should this process exist this way in the first place?”

Maybe it should be automated.

Maybe it should be simplified.

Maybe parts should disappear.

Maybe AI should prepare the work while a human makes the decision.

Maybe the business needs a better standard before technology touches it.

Because:

Automating a bad process doesn’t make it strategic.

It just makes the mess move faster.

More Output Is Not Always More Value

Marketing is a perfect example.

AI can help a business create:

More posts. More emails. More captions. More videos. More variations. More ideas.

That sounds productive.

But if the business still doesn’t understand its customer, has weak follow-up, lacks a clear offer, or doesn’t know what converts...

AI may simply help it produce more noise, faster.

The better question is not:

“How can AI help us create more content?”

It is:

“Where is our marketing workflow actually failing?”

Customer research?

Message development?

Content production?

Lead capture?

Follow-up?

Sales handoff?

Nurture?

Measurement?

The answer determines where AI belongs.

Perhaps the highest-value use of AI isn’t generating another Instagram post.

Maybe it’s extracting recurring objections from sales calls.

Maybe it’s turning customer conversations into content insights.

Maybe it’s improving lead follow-up.

Maybe it’s helping the team reuse institutional knowledge.

Maybe it’s connecting marketing activity to actual sales conversations.

More output is easy to measure. Better outcomes are what matter.

Human Judgment Still Designs the System

This is also why I don’t believe the goal should be to remove humans from every workflow.

AI is extremely useful for work such as:

Drafting. Summarizing. Classifying. Extracting. Comparing. Organizing. Preparing. Standardizing.

But businesses still need people to determine:

What matters.

What good looks like.

Where trust matters.

Where context changes the answer.

Which customer moments require care.

Which tradeoffs are acceptable.

Which decisions carry real consequences.

AI can prepare the work. Humans still need to define, verify, interpret, and decide.

That distinction will become more important—not less—as AI becomes more capable.

The Metric That Changes the Conversation

There is one question that immediately separates AI experimentation from serious business transformation:

How will we know this worked?

The answer doesn’t need to involve a sophisticated dashboard.

Start with the workflow.

If you’re improving customer communication:

Did response time improve?

Did consistency improve?

Did manual rewriting decrease?

If you’re improving sales follow-up:

Are more follow-ups actually happening?

Are they happening faster?

Are more opportunities moving forward?

If you’re improving reporting:

Are there fewer manual steps?

Is information prepared faster?

Are decisions being made with better context?

If you’re improving onboarding:

Is there less back-and-forth?

Are people becoming productive faster?

Without some measure of improvement, a company can confidently say:

“We’re using AI.”

But it cannot confidently say:

“AI improved the business.”

That gap matters.

A Better AI Adoption Sequence

I believe businesses need a simpler way to think about AI:

1. FIND THE FRICTION

Where is time, quality, capacity, or opportunity being lost?

2. REDESIGN THE WORKFLOW

What should be eliminated, simplified, standardized, automated, or redesigned?

3. DEFINE AI’S ROLE

Where can AI prepare, accelerate, organize, analyze, or support the work?

4. PROTECT THE HUMAN LAYER

Where do judgment, trust, context, creativity, accountability, or relationships matter?

5. MEASURE THE CHANGE

What business outcome should improve?

That changes the conversation from:

“How do we use more AI?”

to:

“How do we build a better business—and where can AI help?”

That is a much more valuable question.

The Companies That Win Won’t Have the Most AI

AI tools will become easier to access.

Models will improve.

Features competitors charge for today will become standard tomorrow.

The technology itself will continue to spread.

Which means access to AI will become less of a differentiator.

The advantage will come from knowing where to apply it.

The businesses that create the most value from AI won’t necessarily be the ones with the largest technology stack.

They’ll be the ones that understand their work.

They’ll know where customers are waiting.

Where employees are wasting time.

Where information gets lost.

Where follow-up breaks.

Where quality becomes inconsistent.

Where decisions slow down.

Where opportunities disappear.

And then they’ll redesign those workflows deliberately—with AI supporting the work where it creates measurable value.

So before your company buys another AI tool, ask two questions:

Which workflow actually needs to improve?

And how will we know when it does?

Start there.

Miguel Humbria

AI Strategy, Productivity & Business Transformation