AI is the Hardest Thing CTOs Have Ever Done
Yeah, AI is fun. But it's also the biggest challenge we've ever faced.
I’m going to preface this article by saying that I really believe in the potential of AI.
What we’ve seen in the last few years with LLMs has already been revolutionary. And technical innovation is still moving at breakneck speed.
But I also believe AI is the most challenging transformation CTOs in the Enterprise have ever dealt with — more difficult than Agile or any tech that came before AI.
Consider this:
An E&Y Survey says: More than half of senior technology leaders feel like they’re failing amid AI’s rapid growth.
3 in 4 senior tech execs say driving measurable enterprise value from AI is their top 2026 priority — and that their job security depends on it (Deloitte’s Global Tech Leadership report)
In fact, I think AI will be the hardest thing CTOs will ever do in their careers…but it also represents the biggest opportunity CTOs have ever had to make an impact.
AI will not only personally stretch tech leaders beyond anything we’ve done in the past, but it will also completely redefine the role of CTOs.
What’s Bothering CTOs About AI
External Pressure
Every CTO feels the pressure — Boards are asking harder questions, ELTs are wondering what the AI plan is, and CEOs are dumping their efficiency goals on tech leaders.
In 2023 & 2024 AI was still a novelty. In 2025 the first inklings of stakeholders requiring results started to be seen. Now in 2026 AI is in the cross-hairs of the powers that be, and they want business outcomes.
But delivering productivity from AI is very different than with other transformations. AI impacts are much harder to measure. The tech is much more complex. AI costs are super vague. AI requires a TON of change management. And most fundamentally, its still unclear whether AI really leads to sustained productivity or not.
Self-Imposed Pressure
CTOs are always their own worst critics. We know whats possible and therefore we know what we could achieve vs. what we reported up to the Board/ELT.
With AI not only is it a full-time job to keep up with the technology, but we also want to deliver results that match up to the huge hype from stakeholders. Because of this, most CTOs I know have ratched up the pressure on themselves significantly.
So AI is definitely making our job MORE stressful, not less.
Too Many (AI) Problems to Solve
AI has very broad-ranging impacts in and out of engineering. This makes the set of problems we’re solving for exponentially larger.
Driving cross-functional impact is incredibly difficult. We never had to do that with Cloud or Bitcoin — those technologies weren’t so all encompassing. As CTOs we have to enable Sales, Marketing, Support…every function in the business to get faster and more efficient with AI.
This is 10x harder than if AI just applied to engineering like previous transformations.
Showing Results from AI
Lets face it — as technologists we were always pretty bad at translating engineering work into business results. Well, AI is going to make things tougher.
For example, one of your engineers saved 2 hours a day by using Claude Code. Fantastic! But will that translate into 2 more hours worth of software being released?
It depends on a lot of variables. And the truth is…nobody knows right now.
The Politics Around AI
In most organizations AI is rife with political impact. And its not just the usual “who gets to be the top AI person” type of power-grab issue.
Job loss (RIFs) because of AI is really controversial. As CTOs we’re right in the thick of that conversation. Maybe AI creates more jobs, maybe 50% of dev jobs are gone in 5 years. Again, nobody really knows the answer to what’s going to happen but CTOs are expected to build resource plans and execute on them.
Navigating politics with this kind of optics is not something CTOs have had to do before.
Security is a Nightmare
Shout-out to all my CISO colleagues because AI is going to make a lot of them very successful.
For CTOs, we are facing a minefield of security concerns with AI that are both highly-complex and extremely impactful. AI represents not just another threat vector but an entire new classification of security concerns that is still being defined by both malicious actors and experts who define the new security standards.
AI Forces CTOs Back into the Weeds
I know a bunch of CTO friends who were happy, laid back managers. Their schtick was making their teams feel good & ensuring nothing blew up.
But the days of just managing the team are pretty much over.
Unless you’re in the weeds with AI-driven software delivery you’re going to have a bad time. I’m already seeing a lot of folks changing jobs because delivering results became a thing again, not just managing teams.
The Changing CTO Role
Humans are imperfect — that’s why engineering organizations set things up the way they have for the last 30 years, for example:
Lots of meetings to remind people of things they should already know
Formal / heavy-weight processes like SDLCs
Things like code-commenting & documentation
Middle management layers to shuffle tasks around
1 on 1’s so people feel included
Humans needed all that.
But AI is all-knowning, has near instant recall, and has no feelings to hurt.
In light of this, traditional organizational structures and boundaries are going to collapse. And with it many roles will transform completely, including the CTO job.
Key Changes
CTOs will need to be business leaders more than tech leaders (code gen gets trivialized).
CTOs will have to deliver bigger results than we’ve been used to (AI is fast).
CTOs will have to juggle more products at the same time (AI can parallelize).
CTOs will have smaller organizations and less middle management (AI can handle).
CTOs will be valued for communication skills, not tech knowledge (AI knows all anyway).
CTOs will be required to personally innovate/code (AI assists everyone).
CTOs will have to manage lots (like 100s) of agents (not just people).
I do think CTOs will be around 5 years from now. But the role will change dramatically.
For example, major decisions that CTOs typically make, like whether to modernize or which skillsets to hire for, will be handed off to AI.
In fact, think about all the things that will dissappear because of AI:
No more sprint planning
No more writing most code
No more manual testing
No more writing requirements
The list is long, but for sure each will fall eventually.
What Should CTOs Do?
First, every leader must come to grips with the fact that AI is redistributing where the value lives in an organization.
We used to spend hours writing or reviewing code and that goes away with AI. With that time now available back to us, what should engineers work on?
Here are some examples of new value engineers can create:
Better problem framing
Better system design
Better measurement of metrics
Better talking to customers
Better business translation
The #1 goal of every CTO should be to help their teams understand the new kind of value they can create in the organization. Helping yourself & the team understand the impact and meaning of the AI transformation is ultimately the most beneficial thing tech leaders can do.
As a CTO we must think several steps ahead in order to articulate:
“Here’s what software engineering & QA may look like 2 years from now”
“Here’s what cloudops may look like 2 years from now”
“Here’s what security may look like 2 years from now”
You’re painting a brand new picture nobody has seen before. In the process, you should invite your team to help you sketch out the future.
Then you’ll have to operationalize.
Vision alone doesn’t change everything — people need clear direction on what to stop, what to start, and what to double down on. If you don’t explicitly reallocate time and expectations to new things, teams will default back to the “old way.”
You should aggressively implement AI where it makes sense and then shift capacity toward the new value areas (or nothing meaningful happens.)
At the same time, you have to get ruthless about measurement. Not vanity metrics like “AI usage” or “# of copilots deployed,” but actual business impact: cycle time, revenue velocity, cost reduction, margin expansion.
Your job is to show impact. If you can’t measure it, your Board will assume it’s not working.
Finally, personally lean in harder than you probably want to. AI is not about delegation. The CTOs who win are the ones in the details — testing tools, building prototypes, understanding where AI breaks the SDLC to push their teams forward.
The gap between leaders who “get it” and those who don’t is already massive, and it’s only going to widen. Personal curiosity and intensity will directly determine your outcomes.