Multi Utility

5 min read

AI Business Case for a Utility Board

How to build and present an AI business case to a utility board: what boards want, the presentation structure, the numbers that matter, and objections.

See a 10-minute demo

Written by

Sewanti Lahiri

Published on

August 5, 2026

Updated on

August 7, 2026

An AI business case for a utility board wins on three things: a problem stated in the board's language, a number that shows the cost of doing nothing, and a low-risk path with a rollback. Boards do not approve technology; they approve a defensible return and a bounded risk. This guide gives the structure, the objections you will face, and how to answer them.

Why the AI Business Case Is Different?

Selling AI to a utility board is not the same as selling it internally. A board or council is accountable to ratepayers, cautious with capital, and skeptical of anything that sounds like a trend. That is why the case for a modern platform with built-in analytics and AI-assisted insight has to be made in outcomes and risk, never in features. The good news is that AI is no longer a trend to them. In a 2025 survey of 500 North American utility executives, Itron found that 81 percent of utilities already use AI and 96 percent view it as strategically important, yet only 26 percent have moved beyond proof-of-concept.

That gap is your business case. The board is not deciding whether AI is real; most peer utilities have already answered that. They are deciding whether to move from experiment to production, and whether your utility can do it without wasting money or taking on risk. Framing the ask that way, backed by the broader picture of AI in the utility industry, is far stronger than pitching the technology.

Are you asking your board to try AI, or to move from a pilot to production?

The second is a much easier approval, because it is a decision about scaling something proven, not betting on something new.

What a Utility Board Actually Wants to Hear

A board evaluates an AI proposal through a small number of questions, and every one of them is financial or risk-based, not technical. Answer these and you have a case; miss them and you have a science project.

What the board asks	What to show them
Will it save or make money?	A specific, measurable outcome with a baseline and a target
What does it cost, all in?	Total cost including implementation and change, not just the license line
What is the risk?	The downside, a rollback, and a phased path rather than a big bang
Has anyone like us done this?	A comparable utility reference, not a vendor demo
What happens if we do nothing?	The quantified cost of the status quo

The same discipline applies to any technology investment, which is why this builds directly on the general approach to building a business case for utility software to a board. The difference with AI is that you also have to defuse the hype objection, and the adoption numbers above are how you do it.

Build the Case on the Numbers That Matter

A board will forgive an imperfect solution before it forgives an unsupported number. The strongest AI business case is built on figures the utility can defend, not on a vendor's headline claims.

A baseline you measured, not an estimate, so the improvement is provable later
One or two outcome metrics tied directly to the use case, not a list of benefits
Total cost of ownership, including implementation and ongoing change, not the subscription alone
A payback period on the first use case, stated plainly
The quantified cost of the status quo, which is the number most proposals leave out

Resist the temptation to borrow impressive percentages from a vendor case study. A board can rarely verify them, and one unsupported number undermines the whole case. Building the cost side properly, including the total cost of ownership comparison, is covered in the guide to reducing utility software total cost of ownership.

Does your board know what the status quo actually costs each year?

How to Structure the Board Presentation

Run the presentation in this order. It moves the board from the problem to a bounded decision, and it keeps the technology in the background where it belongs.

Open with the problem in their language. Start with the operational or financial problem, not the technology. The board should recognize the problem before AI is ever mentioned.
Quantify the cost of doing nothing. Put a number on the status quo: staff time, errors, lost revenue, or risk. This is the baseline the decision is measured against.
Present the options, including doing nothing. Show that you considered alternatives. A single recommendation with no options reads as a sales pitch.
Recommend one, and say why. Make a clear recommendation and tie it to the problem and the numbers, not to the features.
Show the full cost and the payback. Total cost of ownership and a realistic payback period on the first use case, not a five-year projection built on assumptions.
Show the risk controls. A phased plan, a first use case chosen for low risk and high value, and a rollback if it does not work.
Make a specific ask. End with exactly what you need approved: the scope, the budget, and the first milestone. Not "support for an AI strategy", but a defined first step.

Choosing that first low-risk use case is the crux, and it is set out in the AI implementation roadmap for small utilities. A board approves a first step far more readily than a program.

Objections You Will Get, and How to Answer

Every AI proposal to a utility board draws the same handful of objections. Prepare an answer to each before you present.

Objection	How to answer it
"AI is just hype."	Point to adoption: most peer utilities already use it. Frame your ask as moving from pilot to production, not experimenting.
"It is too risky."	Start with one low-risk, high-value use case, and show the phased plan and the rollback.
"It will replace our staff."	Frame AI as removing repetitive work for a stretched team, not cutting jobs. Lean teams gain capacity, not redundancy.
"Our data is not ready."	Make data readiness step one of the plan, funded and scoped, rather than a reason to wait.
"It costs too much."	Compare it to the quantified cost of the status quo and the payback on the first use case.

For the finance-specific version of these questions, the municipal CFO's guide to utility billing software covers how a finance lead evaluates the same investment.

What to Avoid

The AI business cases that fail at the board usually make one of these mistakes:

Leading with the technology instead of the problem the board already cares about
Vendor claims a board cannot verify, which cost you credibility on every other number
A vague efficiency benefit with no baseline and no target attached
Asking for a program when you should be asking for a funded first step
No answer to "what if it fails", which is the first question a fiduciary board will ask

Frequently Asked Questions

How do I justify AI spending to a utility board?

Justify it with a measured baseline, one or two outcome metrics tied to a specific use case, the total cost of ownership, a payback period, and the quantified cost of doing nothing. Boards approve a defensible return and a bounded risk, not a technology. Framing the request as moving a proven use case from pilot to production is far stronger than pitching AI in general.

What should an AI business case presentation to a board include?

Open with the problem in the board's language, quantify the cost of the status quo, present the options including doing nothing, recommend one with reasons, show the full cost and payback, show the risk controls and rollback, and end with a specific, funded ask. Keep the technology in the background and the problem and numbers in the foreground.

How do I handle board skepticism about AI?

Address the common objections directly. For the hype objection, cite adoption: a 2025 Itron survey found 81 percent of North American utilities already use AI. For risk, start with one low-risk use case and show the rollback. For jobs, frame AI as adding capacity to a lean team. For data readiness, make it a funded first step rather than a reason to delay.

Should a small utility present a full AI strategy or a single use case?

A single use case. Boards approve a defined, low-risk first step far more readily than an open-ended strategy. Choose one high-value, low-risk application, prove it, and use that result to justify the next step. This is the core of a phased adoption roadmap.

See SMART360 in Action

SMART360 brings AI-assisted billing, metering, and asset insight into one cloud platform for water, electric, and gas utilities, with per-connection pricing and a phased implementation, so a first AI use case is a bounded, board-approvable step rather than a program.

Book a demo