Technology, and today especially AI, often decides how competitive your company is and how it can grow. But if you don't have a technology background, it's hard to judge which solutions are possible today and which of them make sense.
To find out, build a portfolio of solutions, that is, a set of concrete changes to how your company works where, for each one, you know what benefit it will bring, at what cost and in what time, and so what return it will give. This lets you decide deliberately what is worth implementing at any given moment. The portfolio comes together in three steps. First, you define your company's context. Next, you pick a few of the most pressing problems and establish what they cost, where they come from and what they lead to. Finally, you choose solutions and calculate the return on each.
You don't have to describe the whole company right away. Start with a few of the most pressing problems, because that path is cheaper and faster. A description of the whole company, its full operating model, can come later, once you've solved the first problems and want to look for changes more broadly.
1. Define your company's context
First, establish where the company is heading and how it creates value. You'll later relate every problem and every solution to this context, so keep it in front of you at all times.
Start with the vision. Where should the company be in three or five years if everything goes well? Then put the mission into words, describing in one, two or three sentences how your company creates value. Finally, sketch the value creation process, meaning the stages the work goes through and the key areas without which that value can't be delivered.
2. Pick a few of the most pressing problems and establish their cost, causes and consequences
List the problems in your key areas. Look for ones that are already felt today. What takes up a great deal of time? Where do mistakes happen in your processes? Where does information fail to reach the person who needs it? Where do people understand the same task differently?
From this list, pick a few of the most pressing. Three questions will help. How badly does the problem hurt the company? How big is it? How hard does it hit the value creation process? The closer a problem sits to that process, the more it matters for the company's value and its competitive position.
Place each problem you picked in the process where it occurs. Why does this process exist, and how does it connect to the company's mission? Who performs it, and in which systems? It's worth recording the actual flow of the process on a map, for example in the BPMN 2.0 standard.
Then look at how information flows through that process. Take customer inquiries. To answer well, someone needs to know that a question came in, who asked it and who is supposed to answer. That person has to gather the information they need, prepare the answer and send it. If all inquiries land in one shared inbox and nobody knows who is responsible for which email, everyone has to work it out on their own. That costs the team time and energy.
Next, quantify each problem. How often does it occur, and how often are the activities of the process it occurs in carried out? How long do those activities take? What do the systems used in them cost, and what does people's time cost? Look at the distribution, not the average. You may find that only some customers' inquiries go unanswered, and the cause is an overloaded team, with more questions coming in than people can handle well.
The numbers also often show that the most important problems aren't the ones people talk about most. Unanswered inquiries may cost the company the most, even though nobody knows how many there are. Downloading reports by hand, which everyone complains about, may happen rarely and cost little.
Also establish where the problem comes from. You'll find the cause by asking "why?" until you reach the source of the problem. Usually it's a decision made in the company at some point, consciously or not. Maybe there was no time to think it through, maybe there was no need at the time, or maybe nobody knew what consequences it would have. Often these are silent decisions. That's just how things are done, but nobody ever asked why, or something came from the need of the moment and stayed that way. It's usually nobody's fault. The company decided with the time, information and money it had then, and today that decision has consequences it didn't have back then. Sometimes the source of the problem is an outside fact the company can't influence, such as industry regulations.
Then establish the consequences of the problem by asking the other way, "what does this lead to?" Working out who should answer an email takes the team's time. But if the email goes unanswered and the customer leaves for a competitor because of it, the company loses far more. That's why consequences need to be quantified too, in money wherever possible.
You need as much precision as the decision requires. Instead of analyzing every email from the past year, a well-chosen sample is often enough. It gives a reliable result and can cut the cost of the analysis by, say, 10 to 100 times.
After this step, you know where the problems are, what they cost, where they come from and what they lead to. Without it, a company usually asks for a quote on a scope defined from the visible symptoms. It's then unclear whether the company needs it, the scope isn't grounded in its business context, and the delivery risk rests entirely on the vendor's reputation.
3. Design solutions and calculate the return on each
Start with what you want to achieve in the area where the problem occurs. Work from the purpose of the process and from the company's vision and mission.
Then design solution options that remove the cause of the problem, not just its symptoms. If the cause is that nobody in the shared inbox is responsible for a given email, better search will speed up finding emails, but everyone will still have to work out on their own who should answer them. A rule that assigns every email to a specific person removes the cause itself.
Not every solution requires technology. Sometimes a change in organization is enough, for example appointing someone responsible for an area, a new rule or a convention. Sometimes it's enough to configure a program the company already uses. Sometimes you need a tool that speeds up the work, such as automatic assignment of emails to people or an AI agent. Such an agent answers simple inquiries on its own and passes the answers it isn't sure about to a person to check. Choosing solutions takes knowledge of what technology can do today, and in AI agents the market is still taking shape and experienced specialists are few.
Next, for each solution option, establish whether it's feasible, what it costs and how long implementation will take. Then estimate how much it will improve the company's work. You already know the scale of the problem and the design of the solution, so you know how the work will change and how much the company will gain. From that gain, the cost and the implementation time, you calculate the return on investment. Also take into account value that's hard to express in money, such as strategic value or brand image. That's how your portfolio of solutions comes together.
Finally, establish what budget you have and choose the solutions with the highest return. Don't reject solutions that give a small return on their own but make other solutions with a higher return possible. Assess them together with the solutions they enable, because without them the most profitable ones won't happen. Once you've gone through this cycle for a few problems, you'll get a sense of whether it's worth describing the whole company this way.
Let's talk about your company
If you'd like to go through these steps for your company, book a free call with Damian Mikła, founder of Workflows Lab. We'll talk about your company and its most pressing problems.
