Key Takeaways
Use foresight to shape present choices, not to claim certainty about tomorrow
Foresight is preparation. Strategic foresight gathers and interprets information about possible futures so organizations can make better decisions now. Its systemic approach, SFM, links the work rather than treating a single forecasting tool as an answer: scope the question, scan for change, imagine alternatives, set priorities, build strategies, act, and assess impact. Participation runs across the process. The book's cases also call the action phase intervention, while the general framework calls it implementation.
Repeat the cycle. For Singapore's robotics program, the work moved from market and technology research to scenarios, investment priorities, roadmaps, and commercialization choices. Evaluation matters because foresight consumes time and resources. Findings such as new partnerships or projects can feed the next cycle, rather than leaving a report on a shelf.
What's valuable here is the separation of foresight from prediction. A prediction asks which future is most likely; foresight asks what decisions remain sensible across several plausible futures. The second question is more useful when policy, technology, and markets affect one another. The cycle also makes responsibility visible: intelligence without action is research, while action without evaluation can become habit. The framework is demanding, however. Organizations need to narrow the scope early enough to act, without excluding stakeholders or uncertainties that could change the decision.
Scan broad forces, then test whether imported solutions fit local conditions
Start with the setting. Horizon scanning systematically searches for trends, threats, and opportunities that could affect an organization's goals. Singapore's robotics team considered healthcare and environmental services against national pressures: labor shortages, rising healthcare costs, and an aging population. But broad market demand did not establish which robot would work locally.
Context changes the design. Robots suitable for roomy overseas hospital corridors could be too large for Singapore's narrow wards, where crowds could also interrupt cleaning. Outdoor cleaners faced different curbs, rainfall, humidity, and plant debris. These constraints suggested compact, reconfigurable robots that could adapt to tight spaces, with potential relevance to other dense cities. The lesson is not simply to find a growing market. It is to identify the local operating conditions that make a solution useful, or unusable.
This case turns context from background detail into a design variable. A technology's usefulness depends not just on its capabilities, but on its fit with buildings, staffing, weather, and patterns of movement. That is a powerful corrective to market analysis based on global categories alone. Local constraints can also reveal exportable expertise, as the team reasoned about dense urban environments. The caution is that a local need does not automatically create a viable market elsewhere. Similar-looking cities may differ in procurement, infrastructure, and user expectations.
Search technologies by their building blocks, not their marketing labels
Translate between vocabularies. Market reports, scientific papers, and patent databases categorize technology differently. A search for “robot” or “gripper” can miss relevant patents because technical records may classify the underlying components instead of the familiar product. The Singapore team had to study grippers as narrower problems, such as in-hand manipulation and object slippage, then refine its search terms through repeated review.
Use more than one channel. Their search identified 12,808 patents related to robotic end-effectors and gripping, but even a large dataset could leave gaps. Patents from China, Japan, and South Korea were harder to retrieve through English-language searches. Literature reviews and visits to regional trade shows helped expose developments the database search might not show. The combined evidence revealed demand for grippers that could handle varied, non-standard objects.
The practical insight is that information retrieval is itself a technical task. A database does not provide a neutral map of a field; its categories reflect how inventions are documented. Joining patent records to research papers and conversations helps analysts notice what any single source makes invisible. The same process also discourages false confidence from impressive counts. A large search result can be mostly irrelevant, and a thin result can reflect terminology or language barriers rather than a lack of innovation.
Write scenarios to expose interacting uncertainties, not merely extend today's trends
Explore multiple futures. Scenario planning develops stories about possible future conditions, rather than extending one trend into a single forecast. In the book's soft-robotics workshops, researchers and industry experts explored healthcare and inspection settings, then considered how technology maturity and adoption might vary. Possible applications included rehabilitation gloves, assistive devices, and flexible robots inspecting aircraft engines or pipes.
Make uncertainty concrete. The stories helped participants connect technical development with needs, barriers, and potential users. A snake-like robot could inspect an engine without removing it; a soft robot moving through pipes could help locate leaks. These are not promises that deployment will happen on schedule. They are prompts for asking what capabilities, partnerships, and conditions each future would require, and for noticing assumptions that a single forecast could conceal.
Stories help make a technical possibility discussable across groups that do not share the same specialist vocabulary. A researcher can describe a capability, while an operator can imagine where it might fit into work. That shared picture is useful only if the story remains a hypothesis, not a disguised sales pitch. The cases ground imagined futures in practical problems, but even plausible applications can overlook adoption costs, safety concerns, or organizational resistance. Good scenarios keep those uncertainties in view instead of polishing them away.
Match participation to the problem: open calls widen search, invitations deepen focus
Choose the right funnel. After developing scenarios, Singapore's robotics team considered how to gather proposals for future research. An open grant call could attract a broad range of applicants and make selection criteria explicit. It could also produce many competing or duplicated proposals, consume evaluation resources, or attract few applicants when the subject is highly specialized.
Use focused invitations carefully. A bottom-up proposal process invites selected researchers to build a joint response to a defined problem. It can encourage integrated work and make sense when the expert pool is small. But the funder risks overlooking a better approach or capable researchers by choosing participants too early. The choice therefore depends on how specialized the challenge is, how well the funder knows the field, and whether breadth or coordination is the greater need.
The choice is not simply between democratic openness and expert efficiency. Each method allocates the risk differently. Open calls risk noise and duplication; invited teams risk a narrow view of who or what matters. The book's distinction is useful because it makes the funding mechanism part of strategy, rather than an administrative afterthought. A thoughtful funder should ask whether the problem is already well-defined. If it is not, a wide search may reveal alternatives that an invitation process would never surface.
Build roadmaps that connect market needs to technologies, skills, and timing
Link the layers. A roadmap connects where an organization is, where it wants to go, and the steps between. The robotics team combined market drivers and applications with technology levels, research programs, and required resources. For example, healthcare demand for smart hospitals could be connected to autonomous navigation, LiDAR, and the people and capabilities needed to develop them.
Test the map. The team asked 70 experts to assess impact, local competence, and likely timing. Forty-two percent saw autonomous vehicles as having strong social impact, while 74 percent saw LiDAR as influential for industry and economic growth. The survey also projected widespread use of gripper robotics later than the team expected. That disagreement mattered: expert input challenged assumptions rather than merely endorsing the roadmap.
A roadmap is most useful as a shared model of dependencies. It can reveal that a market opportunity needs several technologies, skills, and steps to line up before it becomes deployable. The Delphi survey adds structured challenge by asking experts to assess the same claims and reconsider them in light of others' views. Consensus, however, is not proof. Experts may share blind spots, and a timing estimate can be wrong even when the underlying technology trend is real.
Test customer pain before engineering a polished product
Turn assumptions into questions. Customer discovery tests whether a proposed product solves a real customer problem. The team presented a simple minimum viable product concept for a gripper combining computer vision, tactile sensing, and machine learning, then interviewed research performers, system integrators, and potential adopters. Hospital staff highlighted the expense of surgical robots and the absence of tactile feedback; other stakeholders emphasized safety, softness, cost, and handling different objects.
Translate needs into specifications. Quality function deployment, or QFD, turns customer requirements into technical characteristics. Its House of Quality matrix ranks needs and connects them to design features. For instance, safety could be linked to sensing capabilities. The proposed technology scored strongly on market potential but less strongly on technology readiness, a reminder that customer interest does not mean a product is ready to sell.
The sequence matters: first discover whether the problem is costly enough to solve, then determine what product attributes would address it. QFD provides a bridge between customer language and engineering decisions, reducing the chance that teams optimize features customers do not value. The interviews also show why several stakeholder types matter: users, integrators, and researchers see different parts of the problem. Still, stated interest is weaker evidence than actual purchase or sustained use, so interviews should guide tests rather than stand in for them.
Choose licensing, collaboration, or a spin-off by readiness and execution capacity
Fit the route to the work. Licensing lets an existing company commercialize intellectual property in return for fees or royalties. R&D collaboration can develop a technology further, while requiring agreement over ownership of new intellectual property. A spin-off gives a research team a separate company to build around its invention, but it must find funding, staff, facilities, and customers.
Technology alone is not enough. The gripper project showed promising market demand, but required more research before product commercialization. For an earlier-stage technology, collaboration may help close capability gaps; a mature technology with an interested company may suit licensing. A spin-off can offer focus and retain research links, but those advantages do not remove the strain of building a business. The route depends on both the invention and the organization's ability to execute.
Commercialization is often treated as the final step after research, but each route changes who bears cost, risk, and control. Licensing can reach an existing market faster, while a spin-off may give an invention dedicated attention. Collaboration can combine complementary capabilities, yet complicates decisions about future intellectual property. These are not abstract legal distinctions: they shape incentives and the resources available for development. The strongest choice is not necessarily the route with the greatest theoretical upside, but the one whose participants can sustain the work.
Revisit the roadmap as evidence changes, and measure what foresight actually changes
Keep plans provisional. Roadmaps can lose relevance when technology or market conditions shift. The robotics team noted that improved camera and computer-vision capabilities could change the case for investing in LiDAR. A roadmap should guide decisions, not force an organization to continue down a path simply because it was drawn earlier.
Look for consequences. The SFM impact phase evaluates what the work produces during the process, soon after it, and over a longer period. Useful signs include new relationships, coordinated projects, changed policies, and innovation outcomes. The book also points to data analysis and machine learning as possible ways to detect trends and weak signals. These tools can expand the evidence base, but interpretation and decisions still need accountable people and stakeholders.
This is the difference between planning as a commitment and planning as a learning instrument. A roadmap earns its value when it helps people make choices, and loses it when its authority exceeds the evidence. Measuring impact also protects foresight from becoming a ritual that produces attractive reports without changing behavior. The challenge is attribution: a new project or partnership may have several causes. Evaluation should therefore look for credible contributions, not claim that foresight alone created every favorable outcome.
Analysis
The book's central contribution is less a new forecasting instrument than a disciplined chain of translation. Market signals become research questions; research evidence becomes scenarios; scenarios become priorities; priorities become roadmaps; and roadmaps become customer tests and commercialization choices. The Singapore robotics cases make that chain tangible. They also reveal why a technically impressive invention is only one ingredient in innovation. A gripper has to fit a use case, satisfy users, reach a workable readiness level, and find an organization able to build and sell it.
The strongest insight is that foresight quality depends on connecting evidence that is usually separated. Patent databases show one view of technological activity, market reports another, and local users yet another. Treating any one as the whole picture creates predictable errors: taxonomic blind spots, imported solutions that do not fit, or investment in technologies without a credible route to adoption. The cases show how triangulation can expose these gaps, although the book's evidence is largely organizational and case-based. It illustrates a method more than it proves that the method reliably improves outcomes.
A useful extension is to treat each roadmap claim as a revisable hypothesis with an owner, an observable indicator, and a decision it could change. This would make uncertainty operational: teams could specify what evidence would justify continuing, pausing, or redirecting investment. It would also distinguish a meaningful forecast from an attractive timeline. Expert consensus can inform such decisions, but should not be confused with probability or truth, particularly when experts share institutional incentives.
The framework's breadth is both strength and burden. Its methods support systems thinking, but a full cycle can overwhelm small teams unless the question and time horizon are carefully scoped. The book's recurring practical message is therefore selective integration: use enough methods to connect the decision's important dependencies, then act and learn. Foresight is valuable not because it removes uncertainty, but because it makes assumptions, trade-offs, and adaptation more deliberate.
Glossary
Systemic Foresight Methodology
Connected future-planning cycleA framework for designing foresight as a connected, participatory process. Its stages include scoping, intelligence gathering, imagining futures, setting priorities, interpreting them into strategy, taking action, and evaluating impact. Interaction continues across the stages. The framework combines methods according to the question and context rather than treating any single technique as sufficient.
Horizon scanning
Search for emerging changeA systematic search for trends, threats, opportunities, weak signals, and other developments that could affect an organization's goals. It helps decision-makers understand changes in the environment and identify issues that merit attention. In the book's robotics cases, scanning combined market information with technology research and tested global findings against Singapore's needs.
Scenario planning
Explore plausible future conditionsA method for developing multiple possible future states and considering how interacting uncertainties could produce them. It differs from a single forecast by exploring alternative developments rather than assuming that current trends continue unchanged. Scenarios can be expressed as stories and used to identify strategic options, capabilities, and actions that may matter under different conditions.
Delphi survey
Iterative expert judgmentA survey method that gathers judgments from experts over repeated rounds. Responses are compiled and shared with participants, who can reconsider their views after seeing the group's responses. Anonymity and iteration are intended to support considered convergence. In foresight, Delphi surveys can assess technology timing, impact, capabilities, and priorities, but agreement does not guarantee accuracy.
Roadmapping
Time-based strategy mapA visual planning method that connects a desired future with the present and identifies steps toward it over time. A roadmap can align several layers, such as market drivers, applications, technologies, research programs, and required capabilities. It helps communicate dependencies and resource needs, but must be treated as revisable when conditions or evidence change.
Customer discovery
Test customer problem assumptionsA hypothesis-driven process for learning whether customers have a problem, whether a proposed solution creates value, and whether an application may fit a market. Teams present a prototype or minimum viable product concept, collect stakeholder feedback, and refine or change the proposition. The process aims to test assumptions before extensive product development.
Quality function deployment
Translate needs into specificationsA product-development method that translates qualitative customer requirements into technical characteristics and design specifications. It helps teams communicate customer needs across engineering and other organizational functions. A central tool is the House of Quality, a matrix that ranks customer requirements and relates them to design features. The matrix should be updated as customer evidence changes.
Technology readiness level
Measure technology maturityA scale used to describe how mature a technology is, from early research through demonstrated or deployed systems. The book uses readiness to distinguish an idea or laboratory proof of concept from technology prepared for commercial use. Readiness is only one part of commercialization potential; market demand, resources, and execution capability also matter.
Download PDF
Download EPUB
.epub digital book format is ideal for reading ebooks on phones, tablets, and e-readers.