Most AI strategy conversations have a framing problem. They assume AI investment is an add-on - in other words, you can keep the entire 2025 plan intact, sprinkle in some “AI initiatives,” and plan for transformation.
That’s not how this works. AI doesn’t plug into old plans. True transformation is a major operations shift.
If you want AI transformation in 2026, you’re going to need to invest in AI at the expense of other things. Let’s break it down.
How to think about AI investments
You need to answer these two questions before proceeding:
- How does my company actually create value?
- What are we not going to invest in anymore so we can fund the future?
You need to go into budgeting with the mindset that AI investment is reallocation. If nothing gets cut, nothing meaningful changes.
You also need to know where reallocation of resources is going to have the biggest impact.
Most companies can describe what they do at a high level but they can’t clearly explain the mechanics of how money is made, where leverage lives, or which signals actually drive decisions. You need to model your company before you can transform it.
The critical step becomes abstracting the goal from the current process. Leaders say things like: “We do market research by buying these reports, using this agency, and running these monthly meetings.”
But that’s the method. The goal is: “We want to reliably understand the market so we can make the right product decisions.”
When you separate goal from process, everything opens up:
- You can rethink which inputs you use (e.g., what data matters)
- You can rethink how core work gets done
- You can decide what AI should do vs. what humans must do
- And you stop wasting time “AI-ing up” a process you should have redesigned
This approach is fractal. You can apply it at the enterprise level and at the smallest level of work.
3 AI strategy investments to make in 2026
Once you’re solid on those two mindset shifts, you need to work on this one: Think about investments as more than money. To do this right, you’ll also need to reallocate leadership time, employee bandwidth, and strategic focus.
Most leaders don’t want AI to take away - they only want it to add. But that’s just not going to work. Here’s what leaders need to focus on, even at the expense of other things.
1. Build an actual AI strategy for the next half decade
And no, “we rolled out ChatGPT licenses” is not a strategy.
AI will touch every process in your company. You as a leader need to invest time and strategy into thinking about 3 layers:
- Level 1: Personal productivity. This is not just LLM licenses – so if that’s all you’ve done, you can’t check this off. How do you also raise output quality, get your team comfortable with a new technology, and assemble champions who will bring about the real value driving use cases?
- Level 2: Workflow automation. This involves using low-code platforms or AI features that take over manual work. There’s a host of questions you have to answer here to do this strategically: Which workflows can be redesigned vs. just partially automated? Which ones are worth investing in? Which processes can you stop doing at all?
- Level 3: Core process automation. This involves building custom software to automate a process with a lot of leverage in your business. So which processes are driving revenue for your business that, if automated, would be transformational?
You can’t phone this in and you can’t treat AI as your IC’s pet project. Sit down and scope out the next 5 years of your business and how AI will transform it.
2. Focus on AI-powering your workforce
You need to reskill or replace your people to build an AI-enabled team. Again, this is not the same as rolling out LLM licenses, and it’s not about reducing headcount. You need everyone using AI and using it well.
You should expect to free up 20% of your people’s time to allow them to build competence, learn to apply AI to real work, and build solutions. That’s the tradeoff of the investment.
Half-measures create confident amateurs – people who become just knowledgeable enough to make bad decisions. But done right, the payoff is a highly performant team.
They may have less bandwidth during this period of reskilling, and you may lose a little bit of output. But you need to be willing to say, “We’re going to take a hit this quarter, and this year we expect flat growth as we invest in AI.”
3. Start rethinking your systems of record
This is not a mandate to rip out Salesforce. Keep your CRMs and other systems running.
But ask yourself: Are these the AI systems of the future, or are they dying beasts we’re hitching our wagons to because they suit our current processes? If you do step 1 right, this one starts to answer itself.
And don’t get tricked into the “AI SaaS” BS either. That’s the new “all natural” label for the tech world. Once you look under the hood, very few features are actually AI-powered.
That’s because SaaS and AI are a bit antithetical. AI says “you can do anything”, SaaS says “you can do these 5 things one way”.
Take your five-year business vision and start investing in systems that support it. In some cases, this will mean moving away from “systems of record” toward systems that support hyper-personalization.
My top advice if you want to win 2026
Besides the obvious takeaways from the above – accept the tradeoffs of doing this well, level set on how AI can actually transform your company, and don’t phone it in – this is what will separate leaders in 2026.
1. Ask yourself: “What can I do now with AI that I couldn’t do before?”
There are two categories of AI opportunity:
- Incremental: AI makes existing workflows faster/cheaper.
- Transformative: AI enables workflows you literally couldn’t run before.
Most companies only do #1 because it feels safer. But the big unlocks come from #2. And you don’t get there by automating what a human team used to do. You get there by asking: what outcome do we want, and what process would we design if we started from scratch?
2. Don’t be a quarterly CEO
If you wait until AI strategy is obvious, you’ll be behind. You need to invest before there’s certainty and tolerate the short-term discomfort.
That means telling your board, stakeholders, and team:
- “We’re going to take a hit”
- “This might flatten growth this year”
- “We’re doing it anyway”
If you can’t lead through that – if you need every bet to pay off in a quarter – you’re not leading an AI transformation, you’re managing decline. Be a 5-year CEO and be willing to bet on the long game.
3. Think hard about build vs. buy
Building can sound expensive until you consider:
- Years of license creep
- The cost of customization consultants
- The tradeoffs of inefficient workflows and operational workarounds
Do the math on the effectiveness gap vs. tool cost and make the long-term play.
