You’re doing foresight. But do you actually have a foresight system?

Your team scans for weak signals. You track trends. You run scenario workshops. Maybe you maintain a trend radar and produce regular reports for leadership. Then the project ends.

Six months later, someone starts another scan. People dig out old presentations. Analysts collect sources again. Some of the people involved last time have moved on. New signals have appeared, but nobody has systematically connected them to the trends you identified. The assumptions behind the scenarios haven’t been revisited.

Individual foresight projects can produce great results. A foresight system makes the intelligence accumulate, embeds it into decision-making and compounds its value over time.

On 15 September, our CEO Panu Kause and Client Executive and Futurist Anna Grabtchak were in Frankfurt for the Innovation Roundtable Workshops hosted by Lufthansa. They facilitated two roundtable discussions with innovation professionals around two sides of the same question: Why does future-oriented decision-making fail, and what helps it succeed?

When they got back, I wanted to hear what came out of those conversations. As FIBRES’ Chief Product Officer, I’m particularly interested in where foresight breaks down in practice: what teams struggle to sustain, where technology genuinely helps, and where the real problem has little to do with tools.

So I sat down with Anna and Panu to continue the discussion.

Are you doing foresight, or building a foresight capability that compounds?

Dani: Let’s start with something that caught my attention from a product perspective. Plenty of organizations already have tools and methods for scanning, trend analysis, scenario work and reporting. Yet that doesn’t necessarily add up to a working foresight system. What is missing?

Anna: For me, the difference is how systematically foresight is organized and integrated into the organization. You can have well-organized individual foresight activities without having a system around them. You do a horizon scan for a particular project, build a radar, facilitate a scenario workshop. Then the project ends, perhaps with great results. But the real benefit depends on whether those results are used in decision-making.

A working foresight system creates a compounding effect. What happens to the signals you collected? Who monitors how the trends develop and what new signals emerge? Can you use that work in the next project? Can other people contribute or use the results? How is that futures intelligence created and used in the organization’s actual work and decision-making? Foresight shouldn’t only happen when someone commissions a foresight project. It needs a systematic place in the organization: in strategy, innovation, R&D, risk and other areas where people are making decisions. A lot of the challenge is practical: processes, responsibilities and continuous feedback loops.

Panu: I think that’s exactly the problem. A lack of a continuous and systematic practice is a major weakness in many foresight setups. I would look at three things together: the data, tools and methodologies; the mindset and culture; and the processes, roles and responsibilities.

If one of those is missing, the whole thing becomes much harder to sustain.

You can buy a tool, but if nobody knows who is responsible for the work, it won’t solve the problem. You can have a brilliant foresight team, but if management doesn’t know how to use foresight results, their impact stays limited. And you can have a good process, but if all your intelligence is scattered across presentations and people’s heads, you spend far too much time reconstructing what you already knew.

Dani: That’s an interesting distinction for me because organizations rarely have no tools. They usually have plenty. So is the problem less about individual tools and more about whether the foresight work actually connects across them and into decisions?

Panu: Exactly. You can have all the pieces and still lack the connections. A signal is in one place, trend analysis somewhere else, a scenario is sitting in a presentation, and the discussion about what it means happens in a meeting. Then the actual decision is made somewhere else again.

That’s not only an information management problem. It’s an impact problem. Foresight needs to travel through that chain and reach the people making decisions, from everyday operational choices all the way to top-level strategy. Otherwise you may surface excellent insights without changing anything.

Can good foresight still fail?

Dani: That’s the uncomfortable part. A foresight team can apparently do good work and the organization can still fail to act on it. Where does that disconnect happen?

Panu: There was an analogy at the event that I liked very much. Think about pioneers scouting far ahead of a caravan in the Wild West. Of course you need the scouts. Somebody has to go far enough ahead to see what the caravan cannot yet see. But successful scouting doesn’t mean riding out, seeing something interesting and coming back with a report. The scout needs to tell a story that changes the path of the caravan.

That’s the same test I would apply to foresight. Finding a weak signal isn’t the outcome. Identifying a trend isn’t the outcome. Even creating a very good scenario isn’t the outcome. The impact is the outcome.

Maybe you reconsider an investment. Maybe you prepare for a risk earlier. Maybe you change a strategic assumption. Maybe someone makes a different everyday decision because they understand where the environment is heading.

If foresight repeatedly produces insights but nothing in the organization changes, then from an organizational point of view much of that work is wasted.

Anna: And that connection isn’t always something the foresight team can solve by producing a better report. This is why the way foresight is organized matters so much. If foresight is isolated from the processes where decisions happen, you have to fight every time to make the results relevant to the organization. There can be a natural pull for futures knowledge. But if foresight doesn’t start by understanding decision-makers’ needs, blind spots and pain points, it easily turns into a push: the foresight team trying to convince people to use something they never asked for.

A systematic approach means thinking about those connections from the beginning. Where does foresight feed into strategy? Where does it enter innovation work? How does it inform risk? Who uses it in R&D? How do people elsewhere in the organization encounter and contribute to it? You need to understand those links.

Without those connections, you can easily encounter resistance to foresight itself. People focus on questioning the results instead of trying to use them. When foresight results are treated as forecasts, it usually signals a misunderstanding of what foresight is for. Traditional management doesn’t really teach us how to use this kind of information. We are used to facts, forecasts and ROI. Foresight offers a different kind of ROI: greater resilience over the long term, but also the ability to recognize and act on opportunities in the short term. It helps you work with uncertainty and build strategy against multiple possible futures, making the organization better prepared for change and more capable of acting on emerging opportunities.

How do you prove something about a future that hasn’t happened?

Dani: This came up as a tension between compelling visions and proof. Decision-makers understandably want evidence. But the future hasn’t happened yet. So what should they expect from rigorous foresight?

Anna: There isn’t one predetermined future waiting to be discovered. We work with multiple possible futures, so trying to “prove” one of them is the wrong starting point. The future isn’t fact-based in the same way as the present. That doesn’t mean foresight shouldn’t have an evidence base. You can show what you’re observing now and explore different ways those developments could progress. You can preserve the weak signals behind a trend. You can show the sources. You can explain the drivers of change and stress the assumptions behind your interpretation.

Panu: This is also a mindset issue. Most executives have learned to run businesses based on facts, KPIs, calculations and business cases. That’s completely understandable. But there are very few facts available about the future. So you need a different standard.

Instead of asking only, “Can you prove this will happen?”, ask what evidence points in this direction. Ask which assumptions would need to hold. Ask what this future would mean for the business. Ask what signals you would expect to see if the world started moving toward it. That is still rigorous decision-making. You’re simply not pretending that uncertainty can be calculated away.

Why do scenario discussions get stuck?

Dani: Scenarios were another recurring theme. One problem sounded almost contradictory: scenarios can be too vague to guide decisions, but they can also get buried in details. How do you avoid both?

Anna: You need to start with the scenario building blocks and create a narrative for the big picture first. Then you can go deeper: what would customers or stakeholders need in this scenario world? Where might new competitive advantages emerge? How would your current position hold up? Look for real-life examples that suggest how developments could move in these directions. Now you can connect the big picture with concrete implications for markets, customers and strategic choices.

However, scenario work sometimes gets stuck in the details, or remains at too high a level. People start asking for proof before they’ve even explored the big picture. Or they get stuck on risks: what else can go wrong? You add another risk, another exception, another detail and leave opportunities aside. If the scenarios are described at a very high level, people don’t know what to do with them either. A scenario needs to be descriptive enough to explain what that future would mean in your own context, but you need the big picture first.

This is also where embedding foresight matters. A scenario shouldn’t only be meaningful to the people who created it. Different parts of the organization need to be able to connect that future to the decisions they actually make. You can describe scenarios from innovation, risk management or strategy perspectives by adding the missing interpretation layers.

Panu: And then there is another problem: what happens when people simply don’t like one of the scenarios? All scenarios in a well-built scenario set deserve serious consideration. That doesn’t mean you think all of them are equally likely. The inconvenient scenario can be particularly useful because it stresses the established assumptions.

At the extreme, like we heard in some of our roundtable discussions, you fall victim to the “Titanic cannot sink” syndrome. There is a feeling of invincibility, so there is no real buy-in for examining futures in which the current model fails. That’s dangerous. The scenario that makes people uncomfortable can be exactly the one that reveals where you’re least prepared.

If AI can generate foresight, what are humans needed for?

Dani: I spend a lot of time thinking about where AI genuinely improves foresight workflows and where adding AI simply produces more output. We can already use it to scan huge volumes of information, find patterns and draft trends or scenarios. From the conversations in Frankfurt, where did AI seem to create real value?

Panu: Quite a lot of the mechanical work can now be automated or heavily supported. Scanning, structuring material, clustering, summarizing, drafting: AI can give foresight teams much more capacity. But this makes the human factor more important, not less.

If foresight can increasingly be “generated” by AI, producing the foresight content itself is no longer enough to create impact. The success stories kept coming back to human involvement. People need to discuss what they are seeing, make sense of it together and connect it to their own context. Then they need to connect that understanding to decisions.

And I mean decisions at all levels. Foresight isn’t only something for the annual strategy process or the executive team. If the organization really understands what is changing, that understanding should influence everyday choices as well as major strategic ones.

We humans build the narratives that connect foresight to decision-making. AI can support that process tremendously, but generating a trend report or scenario isn’t the same thing.

Anna: A one-liner in a presentation in Frankfurt said it well: “A scenario in a report is someone else’s opinion. Built with the team, it becomes their own plan.” That’s an important distinction. AI can produce scenarios for you very quickly. But reading AI-generated scenarios on your own isn’t the same process as discussing the scenario worlds together, bringing in different perspectives, challenging each other’s assumptions and working out what the scenario means for us.

So, I would use AI for the parts where it gives us speed and scale. For example, humans can select and combine the building blocks of a scenario, while AI helps develop the narrative, fill gaps and produce a clean first draft. You can also ask AI to propose combinations you didn’t consider yourself. One of those might even deserve a place in the final scenario set, helping challenge the team’s own biases.

Continuous horizon scanning is another obvious example: data search, news or patents monitoring, finding relevant material, structuring signals, drafting descriptions and helping identify connections. Even if AI is doing some of the heavy lifting, you still need to lead and manage the process and create space for people to make the important decisions along the way. And in the end, AI usually isn’t the one making use of the results. People are.

Dani: That’s a useful product design principle too: automate the work that consumes expert time without automating away the work that creates shared understanding.

Anna: Yes. If AI saves me hours of scanning and structuring, but I can still see the evidence, challenge the interpretation and work on it together with other people, that’s useful. The purpose isn’t to produce as much foresight content as possible. It’s to help the organization understand what is changing, identify what is relevant and use that understanding.

What if some of your best signals are already inside the organization?

Dani: One product problem we’ve thought about a lot at FIBRES is that horizon scanning tends to focus on external sources, while valuable signals inside the organization never make it into the foresight process. Did that come through in Frankfurt as well?

Anna: Yes. Potentially a lot of your useful intelligence is already inside the organization. Think about how many new developments, changes and surprises are discussed inside an organization every day, and how many interpretations people already make about where those changes could lead. Sales talks with customers. R&D follows scientific and technological developments. Procurement sees what is happening with suppliers. People working in different markets notice local changes. Customer service hears about new problems and changing demands.

Those observations can be very valuable signals. But they are usually scattered across discussions, documents and different teams. Some of them stay in people’s heads. If foresight is systematically embedded in the organization, you need ways for those observations to enter the foresight process. And it needs to work in the other direction too: the resulting futures intelligence needs to get back to the people making decisions. That’s why I think of this as an organizational system rather than simply a scanning process.

Panu: You need both views. You need to understand what is changing externally: markets, technology, regulation, society. Then you need to connect that with the inside view: your resources, capabilities, culture, processes and knowledge.

That’s part of the human role too. AI can collect and connect enormous amounts of information, but people understand what those changes mean in the context of this organization, this strategy and this decision.

What does a working corporate foresight system need?

Dani: This is obviously where the conversation becomes particularly relevant to me as CPO. If we’re designing technology to support a continuous foresight system rather than another isolated foresight activity, what does that system actually need to support?

Panu: I would come back to those three areas.

First, data, tools and methodologies. You need ways to continuously collect relevant information, preserve the evidence, structure signals and trends, connect things and turn them into something useful. You need to choose methodologies that support the decisions you’re aiming to make. And today you should certainly use AI where it can remove repetitive work.

Second, mindset and culture. Can people work with uncertainty? Are they prepared to question assumptions? Can they take an inconvenient scenario seriously? Does management understand what foresight can and cannot tell them? Are you prepared to embrace the fact that there are multiple possible futures?

Third, processes, roles and responsibilities. Who contributes? Who maintains the work? Who facilitates the sensemaking? When do you revisit the trends and scenarios? Where does the output go? Who is responsible for turning it into action? How do you connect continuous scanning with your key processes for strategy, innovation and risk intelligence?

And I would emphasize that last part. The system needs to connect all the way to impact. A foresight system that produces excellent intelligence but is disconnected from decision-making is simply incomplete.

Anna: Trust is one of the first things you need to build. That starts with explaining what foresight is and isn’t, and understanding the foresight needs of different teams. Planning a foresight system should start with a deep understanding of your organization’s foresight needs: the decisions people need to make and the information they need to make them. That’s why the word system matters when we talk about FIBRES too. We’re not trying to optimize one isolated step of foresight. The useful part is being able to connect the work.

Signal clusters detected during horizon scanning can become evidence for an emerging bigger shift. People should be able to assess and discuss that trend. It needs to become part of a shared view of the results, such as a living radar. Over time, those interpretations and conclusions build into futures knowledge that can inform strategy discussions. The results shouldn’t be abandoned. They should be continuously monitored, updated and available to the people who need them.

Our Foresight Agents can take on parts of the scanning, monitoring, structuring and drafting work. But people remain responsible for managing the process, interpreting what the findings mean in their context and deciding what to do with them. Technology supports the system. A systematic approach builds the foresight capability, but the ROI depends on how the organization uses the results.

How can you tell if you have a system or just a collection of activities?

Dani: When I talk with teams about how they work with foresight today, I’m often more interested in the workflow and the “jobs to be done” than the tool list. If someone wants to assess whether they have a real system, what should they look at?

Anna: I would start by looking at what happens in the major foresight projects, and ask how the results were used and by whom. Does the intelligence accumulate, or does the next project start by collecting much of the same information again? If somebody challenges a persistent assumption, is there a place for the discussion? If you spot a signal or potential trend, can you easily report it? Can you trace it back to the underlying signals, sources and reasoning?

I would also look at who gets to participate. If foresight depends entirely on what a small specialist team can personally scan, you’re missing a lot of benefits. Ideally, observations from your own experts can be compared to what you’re seeing outside the organization, and people have a place to assess, challenge and connect those findings together.

Then there is continuity, ownership and decisions. Someone needs to know who maintains the work and who is responsible for keeping it moving. If you’ve built scenarios, for example, are you monitoring the signals that indicate whether the world is moving toward one of the scenarios? Can your strategy survive or thrive in those scenario worlds? Does your trend radar reflect your latest observations regarding changes? Can you easily update it and share the results? Are foresight results used in both everyday decisions and major annual planning and strategy processes?

The goal is not to make everyone a foresight professional. It’s to make foresight available and relevant where people need it.

So when has foresight actually succeeded?

Dani: I want to finish with the most provocative claim from the Frankfurt discussions: foresight is not for insights, it is for future success. Panu, what do you mean by that?

Panu: It is deliberately provocative because I think this is the most important point. The value of foresight is in its impact, not in producing insights alone.

Organizations don’t invest in foresight because they need more interesting observations about the world. If you take foresight seriously, you should eventually see positive results in organizational performance. That doesn’t mean predicting the future correctly. That’s not the job. It means you saw a risk early enough to do something about it. You recognized an opportunity before it became obvious. You changed an investment because the evidence changed. You prepared for several possible outcomes instead of relying on one assumption.

And this is where the human factor comes back again. An AI can surface the signals. It can help identify the trends. It can even generate the scenarios. But someone still needs to make those things meaningful for the organization.

People discuss them. People connect them with the organization’s situation. People build the narrative around what this means for us. And people carry that understanding into decisions, from everyday choices to the top-level strategy. That’s why I like the caravan analogy. You need the pioneers scouting far ahead. AI can now help them scout much further and faster. But pioneering is only successful when what they find changes the path of the caravan.

Dani: That’s also where this discussion connects back to product for me. The interesting question isn’t how many foresight tools an organization has. It’s whether the whole chain works.

Someone spots something important. Where does that signal go? What existing intelligence does it connect to? Who sees it? Who interprets it? Is it still being monitored six months later? Where does the resulting understanding enter the organization? And if it becomes strategically important, is there a route from that first observation to a decision?

That’s the system.

If the answers depend on a spreadsheet, a slide deck or one particular person remembering what happened in the last project, there is probably still some system-building to do.

If you’re working through these same questions in your organization, book a personalized walkthrough to see how FIBRES supports the full foresight workflow.