AI for real work · 1 of 5 · Discover
Where AI can add value to your business and how to estimate its cost
When a tool can do things that once seemed difficult, it is natural to wonder how you could use it. You start exploring, discover possibilities and soon have more ideas than you can pursue. One essential question still needs an answer: what do you want to achieve with them?
Artificial intelligence opens up a wide range of possibilities. Giving them value means connecting them to a purpose. For a small or medium-sized business, or an independent professional, that purpose concerns the work people need to do, the resources available and the results worth producing.
That is why I suggest starting with the work you want to improve. This is where we can begin to identify an opportunity, then evaluate and develop it.
Start with what you want to achieve
Thinking about technology through its purposes is an idea that runs through Techne and the writing on Idoneos. A technical capability becomes valuable when it enables a change we want to make. To judge its usefulness, we need some clarity about that change.
What would you like your business to do better? The answer may involve time, quality, responsiveness or something that is currently too costly to attempt. It helps to describe the change clearly enough to recognise the difference between the current situation and the one you want.
Defining that difference is valuable in itself. It makes us examine the intended result, who it matters to and what stands in the way. From there, a conversation about AI has a direction.
Look at the work with a different question
Once the purpose is clear, we can look at how the related work is done today. We need to understand the whole sequence: what starts it, what information it requires, who makes decisions and when it is considered complete.
Looking closely reveals effort that familiarity can make invisible. Finding something already known, interpreting information again, correcting inconsistencies and waiting for answers all consume resources. Because these activities are familiar, we may have learned to accept them without asking whether they could be handled better.
That is the question worth recovering. What part of this effort is necessary to achieve the result? What part could be reduced if people had a different capability available?
The answer may involve AI, better organisation of information, a change in practice or a combination of tools. Discovering an opportunity includes understanding the kind of intervention it needs. The more clearly we describe the difficulty, the better we can choose how to address it.
Connect effort to value
Once we recognise a possible improvement, we can ask about its potential. Answering means connecting a change in a task to what that change makes possible for the business.
Reducing effort can free up time. Improving quality can reduce rework. A new capability can make work feasible that was previously out of reach. Its value depends on how important that result is and how well the organisation can use it.
We therefore need to look at what happens next. If time becomes available, what worthwhile work could it support? If delays are reduced, what changes for the people involved? If capacity increases, is there a need that allows the business to use it?
These questions help us distinguish an improvement in a task from its economic effect. Available time is a resource; how it is used determines the benefit it produces. An estimate needs to explain that connection and make its assumptions visible.
At this stage, we can develop an initial estimate. Its purpose is to guide the investigation and show what we need to measure to become more confident.
Understand what it takes to make the improvement possible
When considering the cost of AI, it is easy to focus on the price of a tool. An improvement also requires understanding the process, preparing information, developing what is needed and supporting the people who will use it.
There is an effort to put it in place and another to sustain it. The first includes preparation, development and integration. The second includes operation, review, support and updates. People contribute to both, even when their time does not appear as an additional invoice.
Looking at that whole sequence gives us a fuller comparison between the potential value and the resources needed to achieve it. It also helps identify the greatest uncertainty. The estimate may depend more on the quality of information or the variety of the work than on the cost of running an AI model.
We can then express a range of effort and cost, explaining which conditions would increase or reduce it. That estimate becomes more precise as we learn about the problem and test the solution.
Give the idea a form we can discuss
By this point, the opportunity has a useful shape. We know its purpose, understand a difficulty, can describe an improvement and have an initial sense of its value and cost.
Some questions will remain unanswered. What matters is being able to identify them. A well-defined opportunity separates what we know from what we expect and helps us choose what to investigate first. The AWS guidance on AI adoption also connects this exploration to business objectives and available capabilities.
With that foundation, we can compare ideas. Discovering possibilities expands our choices; choosing among them makes it possible to act.