Beyond the AI Hype: Why the Best AI Projects Often Aren't AI at All

Everyone wants a piece of the AI. Startups are rebranding overnight. Enterprises are cutting jobs to make room for algorithms. And somewhere in a boardroom near you, someone is demanding an AI strategy before the next quarter ends.

But what if the most valuable thing AI can do for your business is reveal that you didn't need it at all?

That is the central argument Max Golikov, Chief Business Development Officer at Sigli, makes. In a recent episode of the Goodfirms Podcast: Conversations That Matter, Max unpacked what really happens when companies pursue AI initiatives, why so many quietly evolve into something else, and why that evolution is often the best possible outcome.

Drawing on nearly 15 years in technical consultancy, Max offers a pragmatic playbook for business leaders who want to invest in real innovation rather than expensive illusions.

1. Why Do So Many AI Projects End Up Becoming Something Else Entirely?

"AI" has become an umbrella term for everything, and the further a client is from the technology world, the broader that umbrella becomes.

When business leaders say they want AI, they are really expressing a desire for innovation. A need to compete, to be faster, to stand out. AI just happens to be the most visible shorthand for that today.

"Most businesses are still very profit-driven. A lot of people think AI would be the answer. But that's a catch-22, where everybody thinks that, but nobody is actually that unique when it comes to trying to do something with AI," says Max.

The real work is uncovering what a client actually needs. And that conversation almost always leads somewhere more grounded than we need AI. Sometimes it is faster logistics. Sometimes it is better data infrastructure. Sometimes it is simply modernizing aging systems that have been quietly holding the business back for years.

2. How Does AI Act as a Trojan Horse for Deeper Business Conversations?

AI is not always the solution, but it is often a powerful conversation starter.

When a client walks in asking for AI, they are signaling openness to transformation. A skilled partner uses that openness to dig into the underlying business: the processes, the data, the legacy systems, and the misaligned incentives, surfacing the problems that actually need solving.

In this sense, AI is the Trojan Horse. It gets you in the door. But the value is in what you find once you are inside.

3. When a Client Realizes Their AI Project Is Actually a Data Cleaning Project, How Do You Manage That Shift?

When a client realizes their ambitious AI initiative is actually a data hygiene project, egos can bruise.

Max’s approach is to meet people where they are. No organization has perfect data. Some have too much unstructured, outdated, and decentralized. Others have too little. Either way, the data is the foundation, and without it, nothing else works.

"Nobody's data is perfect, but everybody's data is beautiful, and everybody needs to work on it to succeed," Max explains.

The key is helping stakeholders see data work not as a step backward, but as the essential groundwork that makes everything else possible.

4. What Does Sigli's Discovery Phase Actually Look Like and What Red Flags Tell You AI Won't Fix the Problem?

Sigli's discovery process is anchored by what Max calls the AI Value Matrix. It is a framework adapted from Gartner, refined through years of client engagements. It evaluates every project across two dimensions:

Technology:

• Quality and availability of data
• Presence of the right human expertise (including internal subject matter experts)
• Overall technical feasibility of the proposed solution

Business:

• Measurable KPIs and OKRs
• An internal sponsor who believes in the transformation and can champion it through uncertainty
• Alignment between business goals and the technical approach

Max added that nobody scores a perfect score. Everybody has something that could have been better.

Max warns, “The biggest red flag is misalignment. When business expectations are high, but technical readiness is low, the project is not ready for AI.”

5. Why Do So Many AI Initiatives Evolve Into Process or Data Projects, and Why Are Those Boring Foundations Actually More Valuable?

Max has been in consultancy for nearly 15 years. He describes it, proudly, as a boring business."

"The boring things, the processes, the data, they are less exciting. The governance of your projects, your infrastructure, your innovation, they don't scream 100 times or a 1000 times growing in like a month or two and becoming a billionaire overnight. But that's where most of the work is done," clarifies Max.

When a company fixes its data infrastructure, streamlines a clunky process, or finally integrates siloed systems, it creates real, durable value. That value compounds quietly while the flashy AI project across the street is still waiting for clean training data.

6. How Often Is the Real Hurdle Not the AI But the Fact That Existing Systems Aren't Ready to Talk to It?

Clients often arrive convinced that their obstacle to progress is an inability to adopt AI. In reality, the obstacle is usually the environment AI would need to operate in:

• Legacy systems built on vendor relationships years ago
• Cloud infrastructure belonging to a provider that no longer exists
• Data living in three formats across six departments

However, AI's newness makes it malleable. It can be shaped to fit complex, messy environments, but only if you invest the time to do that shaping properly.

7. Why Is It Actually a Good Sign When a Company Decides Not to Force AI Where It Doesn't Fit?

There are moments in any engagement when the honest assessment is: you are simply not ready.

The data is not structured. The internal champions are not aligned. The infrastructure cannot support what you want to build. Pressing forward anyway is not bravery but a waste.

"Accepting the limitations of your situation is the better choice a lot of the time. You understand what kind of homework you need to complete before you're ready," adds Max.

8. Can You Share a Real Example Where a Flashy AI Concept Evolved Into a Stable, High-ROI Solution?

Max's favorite example of a project evolving in exactly the right direction is his work with the Allkind Group — a long-standing client that creates learning tools for people with disabilities.

The original vision: A conversational AI tutor providing personalized, one-on-one instruction to students with dyslexia, hearing impairments, and other learning differences.

The problem: AI hallucinations. In a learning management context, inaccurate information is actively harmful.

The pivot: Rather than attempt to replace teachers entirely, the team narrowed focus to the most repeatable, lowest-risk layer of teacher interactions, the how-to questions:

• Where do I click?
• What does this task expect of me?
• How do I interpret this coursework?

"The smaller the context window for an AI, the fewer mistakes it makes," emphasizes Max.

The result was a measurable reduction in how-to support tickets, freeing tutors to focus on the irreplaceable work of actually teaching. Student outcomes improved. Teacher bandwidth expanded.

Three to four years in, conversational tutors are beginning to roll out. But the teachers remain at the center.

9. When the Project Ends Up Being "Something Else," How Do You Redefine the KPIs to Show the Pivot Was a Win?

When a project evolves away from its original AI mandate, the question is: how do you prove to the board it was actually a win?

Max's answer: count what matters from the very beginning.

In the Allkind example, tracking how-to ticket volume was not an obvious business metric, but it connected directly to teacher capacity, which connected to student outcomes, which connected to the organization's core mission.

10. How Can Service Providers Ensure They Are Acting as Partners Rather Than Just Vendors?

Every service provider claims to be a partner. Very few actually are. Max draws the distinction clearly:

• A vendor waits to be handed the missing pieces of your puzzle. You tell them what you need; they deliver it.
• A partner asks to see the whole picture first, advising on which pieces you actually need, and how to find them, even if the honest answer means recommending someone else.

"The actual value for our customers is much more important than delivering on what is said in the moment and making some money off of it,” says Max.

That kind of partnership cannot be declared. It has to be earned over time, through trust and the accumulated evidence of telling clients hard truths.

11. What Is the One Question Every Business Leader Should Ask Their Technical Team Before Budgeting for AI?

Max returns to a question that William De Prêtre, Head of AI at the Allkind Group, asks regularly:

"Do we really need AI for that?"

It sounds almost too simple. But Max argues it is the most powerful filter available.

"From my experience, if the answer is no, you are not going to succeed in implementing it. You're going to spend a lot of time and money and get discouraged," clarifies Max.

For business leaders currently allocating budget for AI, this is the starting point. Not "how do we implement AI?" but "do we actually need it here?"

12. If You Could Rename "The AI Era" to Something More Accurate, What Would It Be?

The AI Era reminds of the California Gold Rush. Everybody's rushing to get their nugget of gold. Startups are now AI startups, regardless of whether or not they need to be. Organizations are firing thousands of people because that's the next big thing, and they don't want to be left behind.

"The only people really seeing the ROI are the ones selling the shovels, the data centers, the video cards. This is why Nvidia is now a $3 trillion company," adds Max.

At some point, he believes, the music will stop. The organizations that spent these years cleaning their data, aligning their processes, and building their infrastructure will be the ones still standing.

Rapid Fire Round with Max Golikov

We put Max on the spot with a few quick questions. Here is what he had to say:

Buzzword most likely to die in 2026?

AI agents. They are sitting right at the peak of the Gartner hype cycle — and what goes up must come down.

AI ROI: Real or myth?

Both. It is very real for Nvidia. Still elusive for OpenAI. For everyone in between, it largely depends on whether you are selling the shovels or panning for the gold.

Prompt engineering: A career or a temporary skill?

A basic skill, like typing on a keyboard. Everyone will need to know how to do it in their own professional context, but it is not a career in itself.

Watch the full episode on YouTube to know what Max answers to questions on open source vs. proprietary, legacy systems, and whether AI will replace developers.

Conclusion

The organizations that win the next decade will not be the ones who moved fastest to attach "AI" to their strategy. They will be the ones who asked harder questions, built cleaner foundations, and resisted the rush long enough to figure out what they actually needed.

As Max Golikov makes clear, that kind of discipline is not boring. It is the most ambitious thing a business can do.

Sometimes, the most intelligent AI strategy is the one that starts somewhere else entirely.