Chart 007
The Navigator’s Journal: Observations from leaders navigating AI, strategy, and organizational change.
Where Does Your Idea Land?
What problem would you be solving if the word “AI” were taken off the proposal?
It’s a simple question, and it changes most of the conversations we have. Someone arrives with an idea such as a chatbot for the team’s procedure, a tool that reads maintenance logs, an assistant that drafts financial reports. The energy is real.
But when we ask “what happens today without it” or “who it affects” and “what it costs when things go wrong”, the idea often turns out to be a solution still looking for its problem.
That isn’t a failure. It’s exactly the moment a use case should be built.
In our white paper, Problem First, Tool Second, we set out a sequence for mariners and the leaders who support them. First, name the problem, then evaluate the tool, and keep a human in the loop throughout.
This Chart takes the next step. It walks through how to turn one idea into a use case that a captain, a department head or a budget holder can actually decide on. And introduces SpecNav.global’s free AI Assessment Workbook that does the arithmetic for you.
A Use Case Is A Decision, Not A Wish List
Most AI ideas stall in one of two places. Some ideas never get written down, so they live as hallway enthusiasm and quietly fade. Others get written down as a tool request such as “we need an AI assistant for maintenance management” which skips straight past whether there’s anything worth solving.
A use case sits between the two. It’s a short, honest record that answers four questions:
- What challenge are we addressing, and who feels it?
- What is it worth if we solve it?
- How ready are we, technically and as a team, to solve it now?
- What happens if it goes wrong?
Answer those plainly and you have something a decision-maker can say yes, no or “not yet” to. That’s the point. A use case doesn’t exist to get approved. It exists to get decided.
Decision Frameworks: A Shared Yardstick
How do you compare one idea that supports the engine room with another one, such as for the training office?
You need a shared yardstick. Without one, the loudest voice or the newest tool usually wins. A decision framework gives everyone the same questions in the same order, so ideas compete on evidence rather than enthusiasm.
There are plenty to choose from. Product teams often use RICE, created at Intercom, which weighs Reach, Impact and Confidence against Effort. Growth teams use ICE, created by Sean Ellis, which rates Impact, Confidence and Ease from 1 to 10.
Each has its place. RICE and ICE are good at ranking a backlog, but they say little about whether your team is ready or what happens when an output is wrong.
That’s why the SpecNav workbook uses a simpler starting point – the Four Dimension Evaluation Framework. It asks about four things, and only four:
Value: what it’s worth if it works.
Feasibility: whether the technology can deliver it today.
Maturity: whether your people and processes are ready.
Risk: what happens if it goes wrong.
It is not a full governance program but it is the first rung. A repeatable way to make a decision, write it down and put a name next to it. For many businesses, that’s where governance actually starts. And starting with an easy to use, repeatable framework helps build the knowledge, experience and trust in the process.
First, Name The Challenge
The workbook opens with the challenge, not the technology. You think about your business and processes, and pick a specific challenge that you are facing today. Quantify it and explain who is affected, from a single department to the whole team and potentially the clients too. Then you describe it in plain language.
A useful test is the difference between a complaint and a challenge.
“The paperwork for this is a mess” is a complaint.
“Watch handover notes are written three different ways, so the oncoming officer has to reconstruct what happened and occasionally misses something” is a challenge. It says what’s happening, why it matters and who carries the cost.
Notice that the second version doesn’t mention AI. If the challenge is real and AI really is the potential solution, it will survive the tool question. Then the four dimensions take over, each rated 1 to 5 with a line of evidence.
Step One: Value
What is it worth if it works? You name the main driver, such as revenue, cost, quality, people or capability and back the score with evidence. “It’ll save time” isn’t evidence. “Officers rebuild the log at every handover” is.
Some use cases don’t pay off on their own, like cleaning up a maintenance database or building basic AI literacy across a crew. The workbook flags these as enabling work, judged on what they unlock next.
Step Two: Feasibility
Can AI deliver this today? Not once a vendor’s roadmap lands. Today. You rate whether the data, infrastructure and skills are in place and name the biggest gap.
The SpecNav.global workbook also asks which category of tool you mean. As the white paper sets out, training and decision-support tools carry a very different risk profile from AI embedded in operational equipment.
A study aid that gets something wrong costs a team member some time. A misread sensor feed on the bridge is a different conversation, as is anything that might affect the operational outcome of the process.
Step Three: Maturity
Are your people, processes and culture ready? This is the dimension most often skipped.
Readiness isn’t the average of feasibility and maturity. It’s the lower of the two. A use case is only as ready as its weakest link.
A ship with a perfect engine and an untrained watch isn’t ready for sea. Neither is a superb crew with a failing generator. A capable AI tool dropped into a team that hasn’t agreed how to check its outputs isn’t a quick win. It’s a risk with a subscription fee.
Step Four: Risk
How likely is it that something goes wrong, and how bad would it be? The workbook uses a familiar 5 x 5 matrix. Multiply the likelihood and impact assessment for a score from 1 to 25, rated low, medium, high or critical.
Then list mitigations, the part people tend to rush. A human review step, a limited pilot, or a limit on what data the tool can see can mitigate the risk, reducing the impact, likelihood or both, making it a more viable solution.
This is where the BATHERS check earns its place. Before you settle on a score, ask the seven questions:
- Whose data trained it
- Who owns the fix
- Where the answer came from
- Who is in the loop
- Whether the use is sustainable
- Whether you’re ready to scale
- Who can see the data
If you can’t answer them, your risk score is a guess.
Where Your Idea Lands
Once all four dimensions are scored, the workbook calculates the result. No formulas to touch, no judgment calls buried in a spreadsheet.
- Quick Win: high value, high readiness, manageable risk. Name an owner, run a time-boxed pilot and track the value you claimed.
- Big Bet: high value, but readiness is lower or the risk needs work. Close the weakest link, release funds in stages and prove it small.
- Foundation: enabling work that passes the risk and feasibility tests. Fund it for what it unlocks.
- No-Go: the risk is too high, the readiness or technology isn’t there, or the value doesn’t stack up. Record why, and revisit when something changes.
A high risk score keeps even a valuable, ready idea out of the Quick Win box until it’s mitigated. A critical one stops it outright. It’s the discipline any watch officer applies: you don’t proceed until you understand what you’re proceeding into.
And a No-Go isn’t wasted effort. It’s a decision made with evidence, before anyone spent money.
The Gate: Someone Has To Own It
The last section turns analysis into accountability. It records a gate decision to proceed, hold for further research, or reject. This includes having a named risk owner, a review date and the reasoning behind the call.
That name matters. In Chart 006 we asked who is holding the helm once AI starts acting rather than answering. If nobody’s name sits next to the risk, nobody is managing it.
What you end up with is a one-page assessment you can attach to a proposal, a budget request or a conversation with a vendor. And when the vendor says “AI-integrated,” you already know what to ask. Which challenge on this page does it solve today?
Start With One Idea
Pick one idea you’ve heard in your organization, your department or around your office this month. Write the challenge down without the word “AI” in it. Then run it through the workbook and see where it lands.
If it’s a Quick Win, you have a pilot. If it’s a No-Go, you’ve saved a budget conversation. Either way, you’ve made a decision instead of collecting another tool.
What’s one idea your team keeps coming back to, and which box do you think it would land in? 🧭
Frequently Asked Questions
What Is An AI Use Case?
A short, honest record that answers four questions: what challenge you are addressing and who feels it, what solving it is worth, how ready you are to solve it now, and what happens if it goes wrong. It exists to get a decision, not an approval.
How Do I Evaluate An AI Idea Before Investing In It?
Start by writing down the challenge without using the word “AI”. Then score the idea from 1 to 5 on four dimensions (Value, Feasibility, Maturity and Risk), backing each score with a line of evidence.
What Is The Four Dimension Evaluation Framework?
SpecNav’s simple, repeatable way to assess AI ideas. It asks about four things only: Value (what it’s worth if it works), Feasibility (whether the technology can deliver it today), Maturity (whether your people and processes are ready) and Risk (what happens if it goes wrong).
How Is It Different From RICE Or ICE Scoring?
RICE and ICE are good at ranking a product backlog. They say little about whether your team is ready or what happens when an AI output is wrong. The Four Dimension Evaluation Framework adds maturity and risk, which matter most when adopting AI.
What Do Quick Win, Big Bet, Foundation And No-Go Mean?
They are the four places an idea can land once it is scored. A Quick Win is high value, high readiness and manageable risk. A Big Bet is high value, but readiness or risk needs work. A Foundation is enabling work judged on what it unlocks. A No-Go means the risk, readiness or value doesn’t stack up yet.
How Is AI Risk Scored?
The workbook uses a 5 x 5 matrix. Likelihood is multiplied by impact for a score from 1 to 25, rated low, medium, high or critical. A high score keeps an idea out of the Quick Win box until it is mitigated, and a critical score stops it outright.
Is The AI Assessment Workbook Free?
Yes. It’s a free Excel download from SpecNav.Global. You enter your scores and evidence, and it calculates where your idea lands, with no formulas to touch.
Read the white paper · Share a challenge · Book a free Navigation Call
AI-assisted: This article was researched and drafted with AI assistance and reviewed by a human before publication.
Cover image: original illustration by SpecNav.Global, inspired by REPMUS 2026 coverage in Ocean News & Technology (lead image credit: NATO). No part of the original photograph is reproduced.
Not ready to talk yet? Take one of our free AI Readiness Assessments to see where your business stands, or book a free 30-minute Navigation Call.
