I don't know what software will look like in five years. I have a guess about who we'll still trust.

A few honest thoughts on AI-native software, the future of enterprise technology, and the one thing I keep betting on anyway.

Alex Rodov · PMO Leadership, PMP, Microsoft MVP Alumni

I try not to make predictions about technology anymore.

Not because I don't have opinions, but because the pattern I keep noticing is that the best insights usually come from staying curious longer than everyone else.

So take this less as a forecast and more as something I've been turning over in my head.

A few months ago, a CIO told me his company had just walked away from a platform they'd spent four years building their processes around.

It wasn't broken.

A leaner, AI-native alternative could do the same job for a fraction of the cost, and could be set up in weeks instead of quarters.

He actually said something pretty cool:

"We were maintaining a museum."

I keep thinking about that sentence.

Because I think it's about to apply to a lot more of us than we'd like to admit.

Nobody Knows What's Coming

I've been reading through forecasts from people who genuinely study the future of software and artificial intelligence, and something jumps out.

The confident predictions don't agree with each other.

Some think we're heading toward software that builds itself on the fly, creating a different interface for every person instead of forcing everyone into the same product.

Others are betting on something stranger:

AI agents that stop behaving like tools and start showing up more like coworkers.

Agents with responsibilities.

Agents with context.

Agents that become part of how teams actually operate.

The categories themselves are starting to blur.

What we currently call "software" may not look much like software in five years.

And honestly, I find that uncertainty reassuring.

Because the people making the boldest predictions increasingly acknowledge how much remains unsettled.

AI adoption is accelerating, but many organizations are still hesitant to hand over meaningful autonomy.

The problem isn't necessarily the technology.

It's trust.

Can we trust the data?

Can we trust the model?

Can we trust the recommendation?

And perhaps most importantly:

Can we trust the person or organization standing behind it?

Maybe the interesting question isn't what the software will do. It's who we'll still believe when it tells us something.

The Real AI Challenge Isn't Capability. It's Trust.

This is where I think the conversation about AI software gets interesting.

We spend an enormous amount of time talking about what AI can do.

Generate.

Analyze.

Predict.

Automate.

Summarize.

Build.

But capability is becoming less scarce.

Trust isn't.

A system can produce an impressive answer in seconds.

That doesn't mean you should make a million-dollar decision based on it.

A platform can automate a workflow.

That doesn't mean the workflow should be automated.

An AI agent can identify a risk.

That doesn't mean it understands the consequences of acting on it.

As AI becomes more capable, the question shifts from:

"Can the technology do this?"

to:

"Should we trust it to do this?"

That is a much harder question.

The Human Skills That Become More Valuable

Buried in almost every serious discussion about the future of work is a version of the same quiet admission.

As technology becomes better at handling routine work, the human capabilities surrounding that technology become more valuable.

Judgment.

Adaptability.

Context.

Communication.

The ability to read a room.

The ability to read a client.

The ability to recognize when a number everyone trusts is actually wrong.

These aren't skills that disappear because AI becomes better.

They become more important because AI becomes better.

I notice this in my own conversations constantly.

People don't ask me what the platform does anymore within the first five minutes.

They ask:

"Who will I be talking to when something breaks?"

That question didn't used to come first.

I think it's becoming one of the most important questions in technology.

The Enterprise Software Model Is Changing

For decades, enterprise software competed on features.

More functionality.

More integrations.

More dashboards.

More configuration.

More customization.

That model worked when software was expensive to build and difficult to change.

AI is changing the economics.

If an AI-native platform can understand natural language, adapt workflows, generate reports, interpret data, and configure itself around the user's needs, then the old advantage of simply having more features starts to weaken.

The software becomes less important as a static product.

It becomes more important as an adaptive system.

And that creates a difficult question for established software companies:

What happens when the product you've spent ten years perfecting can be replaced by something that takes ten weeks to build?

That CIO's "museum" comment suddenly feels less like an isolated story.

It starts to look like a warning.

A Guess About the Next Five Years

So here's my honest, hold it loosely guess about where this goes.

I think the software itself is going to keep changing shape faster than any of us can plan around.

New interfaces.

New AI agents.

New ways of interacting with systems.

New categories that don't even have names yet.

I think many platforms we consider essential today will look very different in five years.

Some will evolve.

Some will disappear.

Some will remain in place simply because replacing them is harder than maintaining them.

And some will become museums.

But I don't think the appetite for genuine relationships changes at the same speed.

If anything, the less certain the technology feels, the more people seem to want a real person who understands their business, picks up the phone, and stays around long enough to actually know them.

That part of this industry feels stubbornly human.

And I say that as someone whose company is built on AI.

Technology Can Scale. Trust Has to Be Earned.

This may be the paradox of the AI era.

Technology is becoming dramatically easier to scale.

Trust isn't.

You can deploy software globally in hours.

You can generate content instantly.

You can automate decisions across thousands of workflows.

But you can't automate your way into a meaningful relationship with a customer.

You can't shortcut credibility.

And you can't build trust simply by adding an AI label to a product.

The companies that matter in the next generation of software may not necessarily be the ones with the most sophisticated models.

They may be the ones that combine sophisticated technology with something much older:

Knowing their customers.

Understanding their problems.

Being accountable for the outcome.

And still being there when things get complicated.

What I Would Bet On

If you asked me today what software will look like in 2031, I'd probably disappoint you.

I don't know.

Maybe interfaces will disappear.

Maybe AI agents will become the primary way we interact with enterprise systems.

Maybe software will generate itself around the task we're trying to accomplish.

Maybe today's platforms will look as outdated as that four year old system the CIO described.

Probably some combination of all of it.

But if you ask me what I think will still matter, I have a much stronger opinion.

Judgment will matter.

Relationships will matter.

Accountability will matter.

Trust will matter.

And the people who understand that will have an advantage regardless of what the technology looks like.

The Part I Don't Want to Predict

Maybe in five years I'll hand this essay to an AI and ask, gently, how wrong I was.

And it will answer, as they always do now, with impeccable manners and a faint, unplaceable condescension.

Or maybe I'll hand it to you instead, over coffee, and you'll tell me the truth.

I know which conversation I'd rather have.

So consider this less a conclusion than an invitation.

Technology will keep changing.

The interfaces will change.

The platforms will change.

The business models will change.

But the reason people choose to trust one company over another may prove surprisingly resistant to all of it.

That's the bet I'm making with how we build.

And more importantly, with how we show up for the people who trust us with their work.

I don't know what software will look like in five years.

I have a guess about who we'll still trust.

Alex Rodov