Beyond Scenarios: Five Ways AI Is Reshaping Strategic Foresight

Most of us who think about the future—professionally, strategically, or just out of curiosity—have probably tried asking ChatGPT to generate scenarios about how the world might look in a decade. At first glance, the outputs can seem impressive—coherent, polished, and superficially convincing. Yet, upon closer examination, these scenarios frequently appear flat and overly formulaic, lacking the nuance, depth, and originality characteristic of expertly crafted foresight work. However, there are more fundamental and powerful ways to harness AI in foresight—ways that go far beyond simply prompting for scenarios. With rapid advancements in artificial intelligence, we’re witnessing and actively participating in AI’s reshaping of modern knowledge work.

In this blog, we explore five ways AI may transform foresight practice. As with anything related to the future, the goal here isn’t to make accurate predictions, but rather to provoke thoughtful consideration and encourage experimentation and adoption of cutting-edge methods that best suit your needs.

Automating Horizon Scanning and Signal Collection

In today’s fast-changing world, manually tracking signals of emerging change is increasingly impractical. Traditional horizon scanning methods often struggle with the sheer volume and speed of new developments. In his article “Why the Future of AI Marks the End of Scenario Planning,” Stefan Michel, Professor of Management at the University of Zurich, highlights this challenge vividly through a thought experiment:

“I researched a list of technologies that interact with the development of AI, resulting in 24 entries. This means we need to monitor trends across these 24 technologies. Calculating (using GPT-4) the potential interactions among these technologies, we discover there are 16,777,239 combinations of cross-fertilization we should be tracking.”

While tracking nearly 17 million interactions is clearly impossible for humans alone, AI isn’t merely the cause of this complexity—it could also be a key part of the solution.

Large Language Models (LLMs) offer immense potential to supercharge the horizon scanning process, automatically gathering, synthesizing, and continuously monitoring diverse signals and trends. They can detect accelerating changes, identify plateauing trends, and uncover emerging counter-trends early on, significantly enhancing practitioners’ ability to interpret complex, dynamic landscapes.

Numerous tech vendors and startups are actively developing solutions in this space. While comprehensively reviewing these solutions is beyond the scope of this blog, it’s an exciting area of innovation worth closely watching.

From Synthetic Facilitators to Synthetic Participants

Human interaction and face-to-face conversations remain central to effective foresight work. Yet, workshops can be costly, logistically challenging, and often suffer from diminishing returns as group sizes grow. As a result, even in futures-aware organizations, foresight workshops tend to be occasional retreats—often without all relevant stakeholders present.

AI introduces an intriguing alternative: synthetic facilitators that can manage, guide, and coordinate discussions across multiple breakout groups simultaneously—enabling real-time idea cross-pollination that would be impossible for a single human facilitator to achieve. A real-world example is ThinkScape, an AI-powered platform that enables large, networked groups of people to engage in productive real-time conversations, optimizing group insights and amplifying crowd intelligence. While these tools haven’t yet been widely applied to foresight workshops, one can easily imagine a foresight practitioner orchestrating a team of AI facilitator bots to run mass-scale engagements for enterprises or civic assemblies. Rather than replacing human facilitation, AI complements it, potentially transforming the depth, breadth, and interconnectedness of collaborative foresight.

Taking this concept further, AI could even embody synthetic stakeholders—simulated voices representing diverse perspectives—to help address the perennial question: “Who’s not at the table?” In asynchronous foresight dialogues, this approach could dramatically increase the scale and frequency of meaningful conversations. For example, each project or functional team could run its own foresight session, while AI-generated stand-ins bring in views from other departments—surfacing blind spots and second-order implications. As with facilitation, the goal isn’t to replace real-time meetings but to open new channels for communication, collaboration, and more inclusive thinking.

Experiential Futures: Using MCPs to Orchestrate Multi-Media Workflows

Many foresight professionals have experimented with AI-generated images or video. Recent developments in Model Context Protocols (MCPs) open a new frontier for experiential futures. MCPs give AI models a shared memory or persistent context across tasks and tools—enabling different applications to collaborate seamlessly within the same workflow. This means complex, professional-grade software—previously limited to highly trained users—can now be manipulated through simple text prompts, with AI translating intent into implementation.

You’re not a 3D artist, but imagine using Blender with an MCP-driven workflow to rapidly generate speculative 3D artifacts. You’re not a game designer, yet you can pair Unreal Engine with a coordinated set of AI models to create immersive virtual environments from narrative prompts. You’re not a GIS analyst, but with geospatial tools integrated into an MCP pipeline, you could simulate how climate adaptation policies might reshape neighborhoods—visually and dynamically.

These kinds of integrations bring previously specialized and resource-intensive methods within reach, enabling richer storytelling and deeper emotional engagement with possible futures.

From Insights to AI-Driven Action

As we’ve seen, AI can dramatically enhance how we generate, visualize, and communicate possible futures. But insight alone isn’t enough—some of the most frustrating outcomes in foresight come when scenarios never translate into action.

Emerging developments like OpenAI Operator, open-source browser-use agents, and Agent-to-Agent (A2A)communication protocols point to a future where AI doesn’t just produce insights—it follows through. These tools allow AI models to interact with external systems, perform tasks on the web, or even collaborate with other agents in coordinated workflows. In essence, they give AI the power to observe, decide, and act across digital environments.

For example, an AI agent could monitor real-time environmental data and autonomously update a scenario dashboard when a key indicator shifts. Or it could identify an emerging geopolitical risk and trigger a tailored briefing email to stakeholders, along with recommended response options pulled from previous foresight outputs.

This marks a shift from passive contemplation to proactive implementation—a future where foresight becomes more dynamic, responsive, and integrated with ongoing decision-making.

A Cautionary Note: Responsibility, Explainability, and Limitations

While this blog takes a techno-optimistic view of AI in foresight, ethical considerations and limitation should not be overlooked.

As we increasingly integrate AI into strategic foresight, critical questions about responsibility, reliance, and transparency inevitably arise. AI should not replace human judgment but serve as a partner—one whose contributions are transparent and understandable. Ensuring explainable AI and maintaining human oversight will be vital as we push the boundaries of what’s possible. Our ability to balance innovation with accountability will shape whether AI becomes a transformative ally or an inscrutable black box.

There are also limitations at multiple levels. AI capability exists on what researchers have described as a “jagged frontier”—a term from a recent Harvard Business Review study that illustrates how AI excels at some tasks while struggling with others that appear similarly complex to humans.

Some well-known issues, such as bias and hallucination, are technical in nature and steadily improving; we’re already seeing progress, and the gap will likely continue to close.

But other limitations are more fundamental and philosophical. Effective foresight isn’t just about producing outputs—it’s about igniting motivation, fostering ownership, and empowering people to shape the future. When humans imagine a day in the future, we feel excited or anxious, hopeful or afraid. The process of envisioning transforms us—it makes us care. Biologists describe living organisms as autopoietic systems—self-organizing and self-maintaining. Large language models, built on next-token prediction, exhibit none of these qualities. In other words: AI doesn’t care. It may help us pursue whatever futures we set our minds to, but defining what matters—what’s desirable, just, or meaningful—remains a deeply human task.

By thoughtfully exploring and responsibly integrating emerging AI capabilities, we can redefine what’s possible—making foresight not only more powerful, but fundamentally more accessible, actionable, and impactful.

Written by Wensupu “Wen” Yang, SAMI Associate and Founder, KWiM Tech 

The views expressed are those of the author(s) and not necessarily of SAMI Consulting.

Achieve more by understanding what the future may bring. We bring skills developed over thirty years of international and national projects to create actionable, transformative strategy. Futures, foresight and scenario planning to make robust decisions in uncertain times. Find out more at www.samiconsulting.co.uk

Image by Gerd Altmann from Pixabay