%20(6).png)
.png)
Zero-based budgeting has been through several cycles of enthusiasm and abandonment since Peter Pyhrr introduced it in the 1970s. Every time cost pressure rises sharply — a recession, a rate cycle, a margin squeeze — ZBB returns to the agenda. Every time it runs into the same operational reality: it takes far more time and data than most finance teams have, and the justification burden is enormous when applied at scale across a large organisation.
What's different in 2026 is that agentic AI has changed the operational feasibility calculation. Not the philosophy of ZBB — that hasn't changed and shouldn't — but the manual burden that made it impractical for all but the most resource-rich finance teams. This article examines what AI actually changes about zero-based budgeting, what it doesn't, and where the risk lies in conflating the two.
Zero-based budgeting starts from zero rather than from last year's actuals. Every line of expenditure has to be justified on its own merits, each cycle, rather than being carried forward with an incremental adjustment. The logic is straightforward: incremental budgeting perpetuates spending that was once justified and never re-examined. ZBB forces the question every period: if we didn't have this budget line, would we create it today?
That's a powerful question in theory. In practice, it generates an enormous amount of analytical work — assembling spend data from across the organisation, building justification packages for each budget owner, reconciling submissions into a coherent view — and most of that work has historically been done manually. Which is exactly why ZBB tends to get adopted enthusiastically and abandoned quietly: not because the philosophy is wrong, but because the operational cost of running it rigorously exceeds what most finance teams can sustain alongside their other responsibilities.
Before addressing what AI changes, it's worth being clear about what it doesn't — because conflating the two leads to ZBB programmes that use AI to automate the wrong things.
AI doesn't eliminate the need for human judgment in ZBB. It eliminates the manual burden that prevented human judgment from being applied where it actually matters.
Three things remain unchanged regardless of how much AI is deployed in the process:
The justification decision is still a human decision
Whether a budget line is justified — whether the activity it funds creates enough value to be funded again this year — is a judgment call that requires understanding the business, its strategy, and its trade-offs. An AI agent can surface the data, model the alternatives, and flag activities where the cost-to-value ratio looks questionable. It cannot make the call. That responsibility stays with the finance team and the business leaders who own the spending.
Organisational will is still the prerequisite
ZBB is politically difficult. It asks budget owners to defend expenditure they've previously taken for granted, and it often surfaces uncomfortable conversations about activities that are funded more by inertia than by value. No amount of AI-assisted data preparation changes the cultural and political work required to run ZBB with real rigour. The finance function still needs executive sponsorship, clear governance, and the authority to challenge justifications that don't hold up.
The strategic framing still comes from leadership
ZBB produces better outcomes when the question "would we create this budget line today?" is answered against a clear strategic backdrop — which activities does the business prioritise, which are investment bets, which are operational necessities. AI can model scenarios against different strategic assumptions. It cannot supply the assumptions themselves. Those still come from leadership, and ZBB programmes that run without clear strategic anchors produce cost-cutting without direction.
With that foundation clear, there are four things AI-driven connected planning genuinely changes about ZBB — each of which addresses one of the operational reasons the methodology has historically been abandoned before completion.
The specific failure mode to watch for in AI-assisted ZBB is using AI to make the justification decisions rather than to inform them. It's a subtle distinction with significant consequences.
An AI model that surfaces activities with low cost-to-value ratios and recommends defunding them is doing analysis. An organisation that acts on those recommendations without human review of the strategic and operational context is outsourcing judgment to a system that can't fully see what it's recommending against. ZBB programmes that rely on AI to make cuts rather than inform them tend to produce short-term cost reduction and medium-term damage to the activities that were quietly generating value in ways the model didn't capture.
The right architecture for AI in ZBB is the same architecture that governs agentic AI in any consequential planning context: the AI prepares, models, and surfaces — with a named human reviewing and deciding, and a full audit trail behind every decision. That's not a constraint on AI's usefulness. It's what makes the outputs trustworthy enough to act on.
In practice, a well-designed ZBB programme with AI-driven connected planning has a clearly different rhythm from the traditional approach. Rather than an annual data gathering sprint followed by weeks of justification package building followed by months of review cycles, the process becomes:
A continuous data foundation that keeps spend, activity, and output data current without a manual assembly exercise at budget season — so the ZBB cycle starts from a current, trusted picture of the business rather than last quarter's actuals
AI-generated analytical scaffolding for justification packages — spend history, benchmarks, scenario models — that analysts review and challenge rather than build from scratch, compressing the analytical phase from weeks to days
Connected scenario modelling that lets leadership test the full-year financial impact of different funding decisions in real time, without rebuilding the model for each scenario — so strategic trade-offs are visible before decisions are made, not after
A decision audit trail that records what was justified, by whom, against what criteria, and what action followed — so ZBB outcomes are accountable and revisable when conditions change mid-year
The outcome isn't a different kind of ZBB — it's the same methodology, executed at a scale and cadence that was previously impractical without the operational overhead that caused most programmes to be abandoned.
Zero-based budgeting's core logic hasn't changed and doesn't need to. The case for justifying expenditure from first principles rather than perpetuating last year's allocations is at least as strong in 2026 as it was when the methodology was introduced. What has changed is the operational feasibility of doing it rigorously — and doing it more than once a year.
AI removes the manual burden that made ZBB impractical at scale. It doesn't remove the judgment, the organisational will, or the strategic clarity that ZBB requires to produce genuinely better resource allocation rather than just faster cost cuts. Finance teams that understand that distinction will use AI to run better ZBB. Teams that don't will use it to automate the wrong decisions faster.
The opportunity isn't AI-driven budgeting. It's AI-enabled budgeting — where the analytical burden that prevented rigorous ZBB is handled by connected, agentic systems, and the judgment that actually determines whether it works stays where it's always belonged.
What is zero-based budgeting and why does it keep coming back?
ZBB starts every budget cycle from zero rather than from last year's actuals — every expenditure has to be justified on its own merits rather than carried forward with an incremental adjustment. It returns during cost pressure cycles because it forces the question: if we didn't have this budget line, would we create it today? That question is perennially valuable. The operational cost of answering it rigorously has historically been what limits adoption.
What does AI actually change about zero-based budgeting?
AI removes the manual operational burden that made ZBB impractical at scale: data assembly across the organisation, building justification packages for hundreds of budget lines simultaneously, running scenario analysis without rebuilding the model for each scenario, and maintaining a continuous rather than annual view of cost-to-value ratios. The methodology and the judgment requirements stay the same.
What are the risks of using AI in zero-based budgeting?
The primary risk is using AI to make justification decisions rather than inform them. An AI model that flags low cost-to-value activities and recommends defunding them is doing analysis. An organisation that acts on those recommendations without human review of strategic and operational context risks cutting activities that were generating value in ways the model didn't capture. AI should prepare and surface — humans should decide.
Can AI make ZBB a continuous process rather than an annual one?
Yes — this is one of the most significant changes AI enables. With a connected data layer continuously reconciling spend and activity data from source systems, and driver-based models updating automatically as assumptions shift, organisations can maintain a live view of cost-to-value ratios across the budget year-round. Budget lines that lose their justification mid-year surface when the signal appears, not at the next annual cycle.
Does AI eliminate the need for organisational buy-in for ZBB?
No. ZBB remains politically difficult regardless of AI investment. It asks budget owners to defend expenditure they've previously taken for granted. That requires executive sponsorship, clear governance, and the authority to challenge justifications that don't hold up — none of which AI provides. The operational feasibility improves. The cultural and political prerequisites stay the same.
Krystal Sync AI provides the connected planning foundation that makes ZBB operationally sustainable — not by replacing the judgment ZBB requires, but by removing the manual burden that has historically caused it to be abandoned before it delivers.
Continuously validates and reconciles spend data from source systems — so the ZBB cycle starts from a current, trusted picture of the business rather than a manual data gathering exercise.
Driver-based modular blueprints that let scenario analysis run without a manual rebuild — so leadership can test the full-year impact of different funding decisions in real time, before committing.
Full audit traceability behind every justification decision — what was funded, by whom, against what criteria, and what action followed — so ZBB outcomes are accountable and revisable when conditions change.
Run ZBB at a scale and cadence that was previously impractical.
Book a demo and see how Krystal Sync AI gives your ZBB programme the connected data foundation it needs to deliver durable results — not just first-year cost cuts.
Book a Demo →
.png)
