Your First 90 Days in a New CTO Role: What the CEO and Board Judge

Most new CTOs plan their first 90 days as an assessment. Learn the stack, meet the team, review the architecture, ship something early so everyone can see the company hired the right technologist. That plan is aimed at the wrong examiners. In about ninety days, the first three months, the CEO, the CFO and the board decide whether they hired a business executive or a head of engineering, and they decide it in money and risk.

This is for the chief technology officer, or the VP stepping into the seat, who reports to the CEO in front of a board or a private-equity sponsor, in a company where technology is material to how it earns. The technology assessment feeds four deliverables. It isn't one of them. If your new seat runs enterprise IT instead, read the CIO's first 90 days: that seat is judged on run cost, contracts and risk.

What the CEO Judges a New Executive On

The CEO judges a new executive on business results, stated in money and risk. Until now you were judged on what your organization built: the platform, the releases, the uptime. From this seat the question is what technology did for revenue, cost and exposure. When the CEO asks for something strategic, that is what they mean.

The technical skills that got you this new job are the ones that pull you off course under pressure. Say you're a tech leader six weeks in. The CEO wants a plan and a number. The CFO questions every dollar. When the board asked about AI, you gave them architecture. Every one of those answers was technically sound, and none of them answered the question that was asked.

I made that climb myself: individual contributor to SVP and Chief Architect at Citigroup Energy, then IT Director at S&P Global Platts and Head of Application Development at Anadarko. What I've seen since, from the coaching side, is a pattern. A bigger leadership role doesn't change who you are. It amplifies the habits you bring, and the first months are when changing them costs least. That pattern explains most of why technical leaders struggle when the room changes.

Standard onboarding plans solve the wrong problem for this seat. The 90-day frame comes from Michael D. Watkins and his book The First 90 Days, and the frame holds. What most CTO plans put inside it assumes your credibility is technical, and the first quarter is exactly when technical credibility stops being enough.

There is rarely a mentor inside the company who has sat in both chairs. That gap is what executive coaching for technology leaders is built for, and it's why the plan below is organized around the people who judge you.

Here are the 90 days the CEO actually judges, at a glance.

Day One to Day 30: Agree With the CEO What Success Means

By day 30, put what success means on one page with the CEO. Success measures, decision rights and how the two of you decide, written down and agreed. It feels bureaucratic to an engineer. It is also the first time the CEO sees technology framed as business terms they can hold you to, and that is why it works.

Start in week one with three conversations with the chief executive officer. What does the company need from technology this year, tied to the business plan? How do they want bad news? What does the board already believe about technology, right or wrong? The third answer tells you what your first board meeting has to correct.

Then run a listening tour across the executive team. Ask the CFO, the COO, the head of sales, the head of customer success and the chief people officer one question: what does technology cost you? Some will answer in money, some in lost deals, some in hours. Each answer becomes a line on the CEO page.

The grievances you collect are your first agenda with your peers and the broader business. Nobody expects you to fix them in month one. They expect you to have heard them. Every stakeholder you asked is now watching to see whether the question was real.

Two situations change the page. With a founder-CEO who wrote the first version of the platform, the decision rights line is the hard one: it's their code and their opinions. The page surfaces that in month one, while it is still cheap to settle.

In a private-equity-backed company, the operating partner and the value-creation plan set the clock. Ask to see the plan in week one, and write your success measures in its language.

Resist announcing a reorg or a rewrite in this window. The stop list waits for days 31 to 60, when you understand the estate well enough to defend it. If this is your first seat at the top table, read up on what changes at the top for any executive.

Days 31 to 60: Translate the Technology Estate Into Money, Risk and Priority

Between days 31 and 60, turn your assessment into a document the CFO reads to the end. State the technology estate in money, risk and priority, in four parts. The architecture review you've been building is an input to that document. The CFO will never read the architecture review, and shouldn't have to.

1. Spend by business capability: where the money goes and what each capability buys the company.
2. Exposure: the three things that could stop revenue or make a headline, each with a likelihood and a cost.
3. The plan in priority order: what is high priority because the business said so, and what will accelerate once the debt is paid down.
4. The stop list: what technology will stop doing, and what that frees.

Technical debt belongs in the exposure section, stated as business exposure: delivery drag and incident cost. The DevOps Research and Assessment (DORA) delivery measures give you a starting set. How often you deploy, how long a change takes to reach customers, how often a change fails, how fast you recover. Each one is a metric the CFO can follow quarter to quarter without a glossary.

The stop list is the credibility move most new CTOs skip. Expect pushback. Somebody owns every system on it, and somebody's favorite vendor is on it too. Stopping things earns more trust with the CFO than any new initiative, because it is the first sign that you will spend the company's money as if it were your own.

Stopping things earns more trust with the CFO than any new initiative.

Here is the same estate told two ways. At S&P Global Platts I led the consolidation of more than 280 market data products onto Amazon Web Services. Told as architecture, that is platforms, data flows, migration waves and a target state. It is accurate, and told that way to any CFO, it loses them by the second slide.

Told as money and risk, the same work becomes three questions a CFO already asks. What does running duplicate products on duplicate infrastructure cost us every year? What would an outage on the old estate cost, and how likely is one? How fast could we launch a new product once everything sits on one platform? Same estate, same facts. Told the second way, it becomes the CFO's agenda too.

Close the document with scaling. Evaluate whether the estate holds the next stage of the business plan: twice the customers, a new market, an acquisition. If it doesn't, say what breaks first and what it costs to fix before it breaks. That is the sentence the CEO will repeat to the board.

Leave the Incident Channel: Build the Bench and a Weekly Cadence

By day 60, decide which of your direct reports can run delivery without you, and which seats have to change. Make the people calls you've been delaying. As long as you're the first person paged at 2 a.m. and the last word in every design review, the company is paying an executive salary for a principal engineer.

The pull back is predictable. Under stress, leaders return to familiar ground. A technology leader designs a better system while the CFO across the table asks for more data, and each thinks the other is missing the point. The quieter version is the voice that says “I can do it better, faster.” Maybe you could, once. Your people are better and faster at it now, and every time you take the keyboard back you tell them otherwise.

Reshape the leadership team and its structure in steps. A big-bang organizational redesign in your first quarter mostly teaches people to wait you out. Name one owner for production stability and one for design authority, so you can leave both rooms and stay out of them. If neither name comes easily, that is your first hiring problem.

Start the hiring pipeline now for the seats you change. The hire you need is a leader. At this level you build a team of leaders, and the headcount of the engineering organization stops being the measure of your job. Engineering keeps its rituals, and agile ceremonies, sprint planning and story-point debates belong to your leaders now. Nobody will miss you at the retro.

Then audit your own week. For two weeks, mark each hour on your calendar as business or organization: the CEO, peers, customers and the board on one side, management of the technology organization on the other. The split is usually a surprise. Leadership at this level means the first column grows every month, and the second shrinks as the bench gets stronger.

Leaders who make decisions without you are the point. A bench that waits for you is a queue. If letting go is the hard part, there is a longer piece on delegating when you used to build it yourself.

It is also the work that coaching for chief technology officers spends the most time on, because the instinct that built your career is the one you are retiring, and nobody inside the company is placed to tell you so. Tandem's executive coaching led by Master Certified Coaches starts there.

Days 61 to 90: The Board Meeting, Your Technology Strategy and AI Readiness

Between days 61 and 90, build the story you'll tell the board. It carries your technology strategy in four parts that a director without a technical background can follow from start to finish. The architecture review is an input. The board story is the deliverable, and the first board meeting is where the CEO finds out whether you can tell it.

1. Where the estate stands, in money and risk.
2. The three moves for the year, and how each one serves the business strategy.
3. What you need from the board: money, patience or a decision.
4. What you will not do, and why.

Frame the technology decisions inside that story as options, each with a cost. A board can choose between options. It can't choose between technologies, and it shouldn't be asked to.

Then prepare for the AI question, because it will come early. In BCG AI Radar 2026, a Boston Consulting Group survey of 2,360 executives in 16 markets published in January 2026, half of the 640 CEOs surveyed said they believe their jobs are on the line if AI does not pay off, and 72% said they are now the main decision makers on AI. That pressure lands on the CTO the moment a director asks.

What goes wrong most often is answering a business question with architecture: models, platforms, a vendor list. The board hears that the CTO hasn't decided what the company should get from AI. The second failure is no position at all. Then the board adopts the CEO's view or a vendor's, and you spend the next year executing someone else's plan.

A board can choose between options. It can't choose between technologies, and it shouldn't be asked to.

A strong answer has three parts. One business outcome the company wants from AI. One guarded first bet, with a cost and a point at which you stop it. One line the company will not cross, on data, customers or safety. A first bet with a stop point promises less than a vendor's slide deck, and that is why the board can trust it.

If your own fluency is thin, start with the AI skills executives need. A board can tell a CTO who uses these tools from one who has only read about them. For the longer arc, including what boards now expect from the seat, there is more on how AI is reshaping the CTO role.

One more case. If the CEO hands you the whole AI agenda in this window, that is a second transition stacked on the first, with its own 90 days. Treat it that way.

Four Common Mistakes New CTOs Make in the First 90 Days

The common mistakes new CTOs make in the first 90 days share one root. Each answers the question the CTO is comfortable with instead of the question the CEO asked. Four show up again and again, and each one is cheap to fix if you catch it before day 90.

Proving You're the Best Engineer in the Room

The early technical win is the classic. You fix the flaky build, rewrite the slow service or settle a design argument nobody else could settle. Engineers notice. Nobody above you asked for that win, and the CEO is still waiting for a plan and a number. Keep the skill. Spend it through your leaders.

Announcing the Big Reorg or the Rewrite in Month One

A reorg or a platform rewrite announced before you understand the estate tells everyone you judged it without context. Morale drops first in the teams that were doing fine, and your best people start taking recruiter calls. Make the people calls by day 60, in steps, with reasons you can say out loud.

Reporting Activity When the CEO Asked for Outcomes

Weekly updates full of migrations started, vendors evaluated and headcount added read as busy. The CEO asked for outcomes against the page you agreed in month one. When the feedback comes back as “think more strategically,” ask what it means in money and risk. The answer is usually somewhere in the business plan.

Treating the CFO as an Obstacle

The CFO questions every dollar because that is the job. Treat those questions as an audit of your story and answer them in the CFO's terms. The CFO is the first peer to win. Once they repeat your numbers in the executive meeting, the board hears them from someone who isn't you. If the seat feels isolating while you build that, it is: the loneliest seat in the C-suite is often the newest one.

Check Your Plan Against the Five Areas at Day 90

At day 90, check your plan against five areas. They are the five a CEO, a CFO and a board weigh when they decide how a new technology executive is landing. Rate yourself first, then ask who would rate you differently.

Each of the five maps to a part of this plan.

• Business and money. The estate stated in spend, exposure and priority, with a stop list.
• The CEO and the board. A one-page agreement, and a board story with a position on AI.
• The leaders under you. Owners for production stability and design authority, and the people calls made.
• Where your week goes. The share of your calendar spent on the business, measured.
• Your agenda: AI and cost. One outcome, one guarded bet, one line you won't cross.

Your own read on these five is the least reliable one you have. The measurable signal is what the CEO, the CFO and your direct reports would say, and theirs is the view that counts at day 90.

Coaching works on how you think, decide and land with the CEO, the CFO and the board. If you want the AI strategy written for you, hire a consultant; for a second opinion on architecture, find a technical mentor who has held the seat. I'm Alex Kudinov, an International Coaching Federation (ICF) Master Certified Coach who climbed from individual contributor to VP-level leadership of global technology organizations.

For a private first read, the Technology Executive Scorecard takes about four minutes: fifteen statements across these same five areas, free, with your result on the page.

Frequently Asked Questions