We have long lived in the Information Age, a “sub-era” of The Internet Age defined by the abundance, accessibility, and distribution of information – making information publicly available at a global scale. But something new is happening now. With the rise of AI, we’ve entered what I like to call the Information Compression Sub-Era, a phase where the vast ocean of data humanity has produced is being distilled, compressed, and transformed into instantly usable insights.
In the Information Age, companies raced to make information as accessible as possible – and they thrived because of it. Google organised the world’s knowledge and made it searchable. YouTube and Spotify democratized access to multimedia, letting anyone watch, listen, and learn from anywhere. Wikipedia built a free crowd-sourced encyclopedia. Even businesses whose models weren’t directly about public information embraced the same principle internally: they built search tools, knowledge bases, and internal archives so employees could find and reuse insights easily. The idea was simple: the more information people could access, the more knowledge they could apply to solving problems. In that era, information itself was valuable, but making it accessible was what created a lasting advantage. Companies that hoarded information or maintained a culture that made it hard to reach were left behind. Take Nokia, for example. Once the undisputed leader in mobile phones, its downfall wasn’t simply about missing the smartphone wave – it was also about how information moved inside the company. Nokia’s teams were deeply siloed [1]; engineers, designers, and executives often worked in isolation, with little open exchange of insights or honest feedback. Critical information about software limitations, market trends, and user experience never flowed upward in time to influence key decisions. Meanwhile, competitors like Apple and Google built open, collaborative ecosystems where information and ideas circulated freely, accelerating innovation [2]. In the end, Nokia didn’t just lose to better technology, it lost to better information culture. But that era of mere information abundance is shifting. With AI, we’re moving from making information accessible to making it intelligible and actionable; compressing oceans of data into concise, usable intelligence.
We had already been seeing glimpses of this information compression trend in culture long before AI took center stage. YouTube Shorts and TikTok made bite-sized content not just popular once more, but this time, are genuinely useful for learning and problem-solving as studies have shown that these kinds of contents can offer such possibilities [3]. Long-form articles accompanied by short summaries or quick-read versions are far appealing than those without them. People increasingly value distilled insight over endless information.
The information compression we’re seeing in the AI era is fundamentally different from what existed before. Unlike pre-AI summaries or short-form content, this new compression is user-driven: the user asks for what they want, and information is intelligently tailored to their perspective, context, and comprehension level.
AI can personalise insights, no just shorten information. It can highlight what’s most relevant, reorganise complex ideas into digestible insights, and anticipate connections – all on the fly.
The implications of this are profound: A project manager can examine a live user engagement dashboard and see in real time how a new feature he managed is impacting metrics. A designer can use the same dashboard to assess how a recently implemented theme change has influenced user behavior. Meanwhile, a marketing manager can view the dashboard and receive insights tailored specifically to her marketing goals. This level of information superpower is certainly a productivity boost for any organisation.
So here’s how organisations can harness it:
-
Invest in insight-generation systems across every key decision-making level of the organisation. This means building systems that deliver both proactive insights (recommendations on what to do next) and reactive insights (analysis of what has already happened). From internal, hyper-personalised ChatGPT-like assistants trained on company data to advanced, intelligent analytics tools. While their sophistication will naturally depend on the technical expertise of those who build them, the goal should be about creating systems that generate data-backed insights and can be adapted to “speak” the language of your organisation.
-
Establish insight-sharing workflows. Encourage teams to share, discuss, and collaborate on their discoveries, and ensure your systems make that easy. Insight should not live in isolation (don’t be another Nokia). Every AI-generated report, analysis or AI-derived insights should be accessible, commentable, and referenceable within your internal platform, allowing others to build upon it. The goal is to turn individual findings into collective intelligence that compounds over time.
-
Build better data models. Legacy organisations are notorious for poor or inconsistent data structures – but this is fixable. Modern system designers use techniques like model layering, where raw, messy data is mapped onto a cleaner, logical schema (often called the analysis model) that’s used for reporting, inference, and AI-driven insights. This approach creates a bridge between legacy data and modern intelligence systems. AI models, in particular, require intuitively explainable data structures, and without them, you leave room for misinterpretation and hallucination.
-
Create a framework for applying insights. Crucial or eyebrow-raising insights should never be taken lightly. This doesn’t mean AI becomes an all-knowing oracle whispering to decision-makers, however, it means recognizing AI-driven analytical tools as directional agents of some sort; they may not always be right about the problem or opportunity they surface, but they often point to where is worth looking. I believe that that alone has immense value. To make the most of this, organisations should have a clear framework for turning promising insights into actionable projects – complete with milestones, ownership, and requirements. And just like the insights themselves, these projects should remain open, transparent, and easily accessible to every stakeholder.
-
Measure, measure, measure. No system is complete without feedback. Continuously measure how effective your insights are, both the proactive ones (what the AI suggests you do) and the reactive ones (what it helps you understand after the fact). Ask yourself: is it worth acting on what the AI says? One unconventional but powerful metric I use is the “wow factor”; how many genuine moments of insight or surprise does the system produce? If the results consistently make you say “wow,” then it’s worth doubling down: invest in improving the system’s capabilities and trust it enough to follow where its insights lead.
Building internal systems that deliver all these capabilities is no small feat – it’s in fact a contemporary research and engineering challenge among the brightest systems minds today. Designing such systems requires deep thought across data architecture, human–AI interaction, and organisational knowledge flow. We’ll explore the methods and techniques behind this in other posts. I hope you found this one insightful 😉. Thanks for reading.