A CFO sat across from me and spent twenty minutes explaining why his company could not cut costs without impacting service delivery.

What he was really saying was: “We have always done it this way, and I have no idea where the money actually goes.”

This is the accounting equivalent of assuming the map is still accurate after the territory changed. And for building owners and developers running capital programs right now, the territory has changed dramatically.

The problem with building on last year’s numbers

Most organizations budget incrementally. Take last year’s numbers, add a percentage for growth or inflation, negotiate at the margins, lock it in. It is fast, it is familiar, and it embeds every assumption from last year into this year’s plan, including the assumptions that are no longer valid.

Consider the implications for a development pro forma: when was the benchmark you are using for cost-per-square-foot actually established? What market conditions, what labor rates, what supply chain dynamics did it reflect? How many of those conditions still apply?

The contingency baked into the model may not be protecting you from project-specific risk. It may be compensating for the fact that your baseline assumptions were outdated when you built the model.

Zero-based budgeting is the corrective. Not a management consulting trend, a straightforward discipline: start from zero and justify every dollar you plan to spend. Not “we allocated 12% to G&A last year so we will allocate 12% this year,” but “what do we actually need to spend on G&A to deliver our objectives, starting from nothing?”

This is a difficult exercise because it makes things visible. It surfaces expense lines that exist because they always have, not because they create value. It exposes departments whose cost structure reflects a decade-old operational model that AI could now handle at a fraction of the cost. It creates accountability for headcount decisions that have never been rigorously examined.

That exposure is productive. The market is indifferent to your organizational politics.

The AI wedge

What changes the calculus in 2026 specifically is that AI does not create incremental efficiency. It creates step-function changes.

Traditional thinking: reduce administrative costs by 2% annually through better processes. Reasonable, achievable, and largely irrelevant at the scale of what is actually possible.

Grounded thinking: if AI handles 70% of your accounting close process, what does your accounting department actually need to look like? Not “2% smaller.” Potentially 50% smaller, and then another 30% smaller in two years as the next capability arrives.

Organizations still building pro formas by adjusting last year’s numbers are planning to be uncompetitive. A competitor who has rebuilt their cost structure around AI-native workflows will be able to price deals that others cannot. They will carry less overhead, move faster, and close business that slower-moving firms lose. The efficiency gap will show up in outcomes long before it shows up in anyone’s strategic plan.

The firms with the most exposure are those whose cost structures were calibrated for a world where scaling required linear increases in headcount. That assumption no longer holds.

What this means for your project partners

The same logic applies to the firms you hire to manage your projects.

An owner’s rep or program manager with ten people and AI tools can now do work that previously required twenty. If they are not passing some of that efficiency back to you (in the form of better coverage, more thorough analysis, faster responses) the question worth asking is where that efficiency is going.

Ask directly: how has your use of AI changed your capacity per team member? What can you do now for the same fee that you could not do two years ago? Firms with honest, specific answers to those questions have genuinely integrated AI into their workflows. Firms with vague answers have added AI to their marketing materials.

The right question

The question is not “how do we do what we are doing more efficiently?” That framing keeps you optimizing a model that may not be the right model.

The question is: if we were starting this project program from scratch today, with all current tools available, what would it look like? How would we staff it? What would we spend? What would we measure?

That is zero-based thinking. It is harder. It is more threatening to established structures. And it is the only intellectually honest way to plan in an environment where the fundamental economics of how work gets done are changing faster than annual budget cycles can accommodate.

The organizations that navigate the next five years effectively will not be the ones that cut 10% across the board. They will be the ones that rebuilt from first principles and asked what they actually need.

Start from zero. Justify everything. Build forward.