How much time did you save? It is a reasonable question. It is also a remarkably small window through which to look at what AI can change.
I want to know whether a decision became better and arrived in time. Whether someone finished the day with energy left. Whether an idea reached the room because its owner finally had a way to express it. Whether we made something that would otherwise never have existed. And whether we stopped doing something we no longer needed to do.
These seven value vectors are places to look, questions to ask and possibilities to explore. They grew out of my work with organisations adopting AI, and they keep changing as that work changes. Use them with a real situation in mind. A framework earns its place through what it helps you notice.
Efficiency AI. Opportunity AI. Both.
Efficiency AI
How can we do this better, faster or with less effort?
Improve something we already do. Reduce waste, repetition, delay and avoidable mistakes. These gains matter, especially to the person who carries the burden today.
Opportunity AI
What could we now do that we previously couldn’t, wouldn’t or hadn’t considered?
Explore new capabilities, services, experiences and ways of working. Question the constraints that shaped the old process, including whether parts of it need to exist at all.
They coexist and are intertwined. Making something dramatically easier can make an entirely new use possible. Exploring a new opportunity can reveal a simpler way to do the work we already have. This is no ladder with efficiency at the bottom and opportunity waiting at the top.
Both belong in every vector. Finding an old report faster is useful. Bringing relevant experience into a decision before anyone remembers to search changes the information flow. Discovering that the report and its weekly meeting have become unnecessary changes the work itself.
Two I care about a little extra
If you ask me to pick favourites, they are number two and number six: cognitive capacity and wellbeing, and access, inclusion and agency. There is something deeply human here. Someone may have the insight, the curiosity or the ambition already, while language, memory, confidence, disability or an awkward system gets in the way of expressing it or acting on it.
AI can help bridge some of those gaps. The person does not have to become a different person to contribute. The environment can become better at meeting them where they are. And having enough energy left for the rest of life is a valuable outcome in its own right.
Vector 01
Decision quality and judgment
Can we make a better decision, at the speed the situation requires?
Decision velocity matters. A carefully considered answer can still arrive too late. When circumstances move faster than our ability to respond, waiting becomes a decision with consequences of its own. If you snooze, you lose.
AI can help assemble context, compare options, expose assumptions and explore consequences. It can also help us notice that we are answering the wrong question. The opportunity reaches into how decisions happen: who has the information, who has the authority and which steps are adding understanding rather than delay.
Good judgment includes choosing the pace. Sometimes we should make a small, reversible decision and learn. Sometimes the consequences justify slowing down. More confidence in an answer is not, by itself, evidence that the answer improved.
A possibility to explore
Imagine a team that spends its weekly meeting reconstructing what happened. A shared, source-linked brief could let people arrive with the context already understood. The meeting can then focus on the unresolved choice, or disappear when someone already has the authority and information to decide.
What would count as value?
Look for shorter avoidable delays alongside better reasoning, appropriate escalation and outcomes you can revisit. Ask what changed because the decision was both better and timely.
Explore decision quality and judgment with your AI
Help me explore decision quality and decision velocity in my work. Start by asking me about one recurring decision, who it affects and what makes it difficult or slow. Ask one question at a time. Help me separate missing information, genuine uncertainty and unclear authority. Explore how AI could improve the judgment, shorten avoidable delay and change how information reaches the decision-maker. Identify anything we could stop doing. Challenge my assumptions and distinguish reversible choices from consequential ones. Then help me design one small experiment with evidence of both decision quality and timeliness. Do not invent facts or make the decision for me.
Vector 02
Cognitive capacity and wellbeing
What could become lighter, so we have more attention and energy for what matters?
A day can be full of small tasks that are individually easy and collectively exhausting. Remembering unfinished threads. Finding the right version. Switching between systems. Rebuilding context. Translating an idea into the format someone else needs.
AI can carry some of that load and help preserve continuity. The opportunity is to arrange work so that attention is spent deliberately: on a difficult conversation, a creative question, careful judgment, learning or simply being present.
This is one of my favourites because the benefit can be felt in a person's day. Leaving work less depleted counts. We do not have to spend every reclaimed minute on another task. And we need to check the whole experience: an assistant that produces endless material to review may create more cognitive work than it removes.
A possibility to explore
Instead of repeatedly checking several channels in case something needs attention, imagine a bounded briefing that brings together relevant changes, unresolved questions and their sources. Its usefulness depends on what it reliably leaves out as well as what it includes, with a clear way to check and correct it.
What would count as value?
Ask people about interruptions, mental effort, after-hours catch-up and energy left at the end of the day. Include the effort of checking and managing the AI. Protect thinking and learning that are worth doing yourself.
Explore cognitive capacity and wellbeing with your AI
Help me explore cognitive capacity and wellbeing in my everyday work. Begin by asking about a recent day that left me mentally drained, and what I wish I had energy for instead. Ask one question at a time. Help me identify avoidable remembering, searching, switching and coordination, while preserving thinking I value. Explore both AI assistance and changes that need no AI, including work we could stop. Account for the effort of checking the assistant. Do not diagnose me or assume that freed capacity should become more output. Help me choose one small experiment and a simple way to judge whether my day actually feels and works better.
Vector 03
Shared knowledge and collective intelligence
How can what we learn become useful to someone else, at the moment they need it?
Organisational memory starts with retaining what happened and why. Knowledge activation takes the next step: making that experience useful in the work now being done. An archive has potential; the value appears when someone can understand, question and act on what it contains.
AI can help connect decisions, conversations, experiments and corrections across time and teams. This opens a possibility beyond better search: knowledge can travel with the work instead of relying on someone remembering who to ask.
The system still needs care. An old assumption is not a current fact. A summary is not the original conversation. Sources, dates, disagreement and corrections need to remain visible. And not everything someone knows can be captured in a document; people still need routes to one another.
A possibility to explore
Imagine a service observation reaching the people designing the next product, connected to similar observations and the decisions they informed. The person who raised it can see what happened next. We have changed an information flow, not merely made another searchable folder.
What would count as value?
Look for useful reuse, fewer repeated mistakes, clearer handovers and people finding context they did not know existed. Check that the knowledge is current, permissioned and connected to its sources.
Explore shared knowledge and collective intelligence with your AI
Help me explore shared knowledge and collective intelligence in my team. Ask me about one situation where useful experience failed to reach the person who needed it. Ask one question at a time. Map how the knowledge is created, captured, passed on, questioned and updated. Explore how AI could help it travel with the work, including a different information flow and a handover or report we might retire. Keep source dates, disagreements, permissions and human expertise in view. Do not assume you can access our internal systems. Help me design one small experiment and show how we would know the knowledge was used well, rather than merely stored.
Vector 04
Foresight and resilience
What can we notice earlier, prepare for better and respond to more effectively?
Proactive compliance and risk mitigation remain important here. An overlooked obligation, a recurring quality problem or a fragile dependency can have real consequences. But the opportunity is broader than checking whether we followed the rules.
AI can help connect weak signals, examine assumptions, compare scenarios and rehearse a response before circumstances force one on us. It can support a move from periodic inspection to timely attention, with people deciding which signals deserve action.
Foresight is not prediction dressed up as certainty. A useful scenario makes us better prepared even when events unfold differently. Resilience also means having a way to detect mistakes, recover and keep operating when the AI itself is unavailable or wrong.
A possibility to explore
A team reviewing suppliers could explore what would happen if a key delivery failed, what early signs it could reasonably observe and which alternatives it could prepare. The useful result might be a changed dependency or a rehearsed response, rather than a longer risk register.
What would count as value?
Look for earlier relevant warnings, fewer distracting false alarms, better preparation and more effective recovery. Distinguish AI helping us manage risk from the safeguards needed to use AI responsibly across all seven vectors.
Explore foresight and resilience with your AI
Help me explore foresight and resilience in my work. Start by asking what important outcome could be disrupted, and by what kinds of change. Ask one question at a time. Help me examine dependencies, obligations and overlooked assumptions without inventing risks as facts. Explore useful signals, alternative scenarios and practical responses, including a dependency or routine we could remove. Separate established evidence from speculation. Include false alarms and failure of the AI itself. Help me choose one bounded experiment with a clear owner, a response to uncertainty and evidence that we are better prepared. For specialist obligations, identify what needs qualified verification.
Vector 05
Learning and capability development
What can we become capable of, individually and together?
There is a difference between completing a task with AI and learning something from doing it. Both can be useful. We should be clear about which one we are trying to achieve.
AI can be a patient explainer, a practice partner, a critic or a source of feedback. Learning can meet us inside a real task, at the moment a question becomes meaningful. We can try a difficult conversation, explore an unfamiliar subject or test an idea before it meets the world.
The opportunity includes developing others. A useful method can travel through a team; people can share what they tried and what surprised them. Onboarding, coaching and career development belong here, but learning is much bigger than an employee lifecycle. Curiosity does not need a job title.
A possibility to explore
Someone preparing for a difficult conversation could practise with an AI, pause to examine an assumption, try a different approach and reflect afterwards. The value is in the understanding and judgment they develop, not the existence of a polished script.
What would count as value?
Look for better questions, transferable understanding, useful practice and the ability to explain or apply what was learned. Distinguish supported performance from independent skill; both matter, but they are different outcomes.
Explore learning and capability development with your AI
Help me explore learning and capability development. Ask what I want to become better at, why it matters and what I can currently do. Ask one question at a time. Help me distinguish getting today's task done from developing lasting understanding or skill. Suggest a small practice activity in which I do meaningful thinking, receive feedback and try again. Explore how AI could make practice possible where it was previously difficult, and what unhelpful training routine we could stop. Adapt to my starting point. Finish by helping me test what I can explain, transfer or do, and what still depends on AI support.
Vector 06
Access, inclusion and agency
Who gets to participate, express themselves and act when a barrier becomes smaller?
Some barriers are easy to mistake for a lack of ability. A person may understand the issue but struggle to write about it. They may have the experience but not the vocabulary. They may know what they want to make but have no practical route to building it.
AI can help bridge language, expression, sensory, cognitive and technical barriers. It can also help services meet people through an interface, pace or form of communication that works for them. This is my other favourite: the moment someone becomes able to act on an intention that previously had nowhere to go.
Access alone is not enough. Agency means having meaningful choice and control: being able to understand, question, correct or decline what the system does. A tool that works beautifully for me may still exclude somebody else. We need to ask the people affected and learn from their actual experience.
A possibility to explore
A colleague with a strong idea but difficulty writing could use voice and AI to shape a proposal, check that it still says what they mean and take it into the conversation themselves. The contribution belongs to them. The organisation also needs to listen; a better document cannot, on its own, remove a power barrier.
What would count as value?
Ask who can now contribute or act, whose effort has decreased and who remains excluded. Check affordability, usability and control with the people concerned. Assistance should expand their choices.
Explore access, inclusion and agency with your AI
Help me explore access, inclusion and agency in a situation I will describe. Start by asking what someone wants to do and what gets in the way. Ask one question at a time. Do not assume a diagnosis, a lack of ability or that my own experience represents everybody. Explore barriers in language, expression, interfaces, resources and decision-making power. Consider how AI and changes to the environment could help, including a requirement we could remove. Preserve the person's ability to understand, correct, choose and decline. Help me design a small experiment with the affected person and evidence that their practical possibilities and control have increased.
Vector 07
Creation and destruction
What should we bring into existence, and what can we deliberately bring to an end?
AI gives us new ways to make products, services, tools, experiences and art. It can shorten the distance between an idea and something we can try. A person who could imagine a useful tool may now be able to build it. An organisation may discover a service it could not previously offer.
Creation is only part of this vector. I also want us to be willing to stop doing things: a report nobody uses, a handover that exists because systems cannot talk to one another, a routine built around a constraint that no longer applies. Sometimes the most valuable thing we create is a way to make that work disappear.
Destruction is a deliberate word. Its objects need to be named. What are we retiring, whose work or life does that affect, and what useful function must survive? A seemingly inefficient conversation may build trust. A manual step may carry judgment we have not understood. We need to learn why something exists before congratulating ourselves on removing it.
A possibility to explore
Imagine replacing a standard report with a service people can question when a decision arises. That might create a new offer for customers and remove repeated preparation work. First test whether people get what they need, including the context and conversations the report used to support. Retirement is a design decision, not an automatic reward for building the replacement.
What would count as value?
Look for something newly possible or useful, a genuine need met and work that has actually ceased. Include transition costs and effects on the people involved. Producing a prototype does not establish demand, and hiding a task in someone else's workload is not eliminating it.
Explore creation and destruction with your AI
Help me explore creation and destruction in my work. Begin by asking about a need I would like to meet, an idea I could not previously realise or a recurring activity whose purpose I question. Ask one question at a time. Explore what AI makes newly possible and what we could deliberately stop doing. For anything we might retire, investigate why it exists, who depends on it and what valuable function must survive. Consider affected people and transition costs. Challenge novelty for its own sake. Help me design one small, reversible test with evidence of usefulness and explicit conditions for keeping, changing or retiring the old arrangement.
Start with something real
Choose one piece of work, one ambition or one frustration. You do not need seven projects. Follow the vector that helps you see it more clearly, then notice where the others become relevant.
For that situation, explore both efficiency AI and opportunity AI. What could become better or easier? What could become possible? What could we stop doing? Describe what a useful change would mean, for whom, before choosing the tool.
The vectors overlap. One change may free attention, improve a decision and let more people contribute. Notice those connections without counting the same benefit several times. Include the costs of checking, maintaining and changing the system, and ask who gains the value and who carries the burden.
Try something small enough to learn from. Keep what works, change what does not and let the result improve the next question. These are lenses for inquiry, not seven promises that AI will deliver.
A framework that keeps developing
My framework began as Five Value Vectors in the context of M365 Copilot in October 2025. Access and Inclusion became the sixth in July 2026. This seven-vector version, published on 19 October 2026, broadens the language, makes efficiency AI and opportunity AI explicit together, and adds Creation and Destruction. The wording and prompts were developed through AI + JJ. Five Ways AI Changes Work remains a separate, related framework.
- Previous version: Six Value Vectors
- The Four Jobs: doing, improving, learning and developing others
- Interface Sovereignty: services meeting people through their chosen interface
The previous page remains available in its six-vector form, with a link to this version. The examples here are possibilities to explore, not claims of measured client results.
Editorial explanation and conversation prompts · AI + JJ