<center> # Conversations are the New Oil <big> **Strategic considerations for protecting communication assets through systematic safeguards** </big> </center> <center> *Written by Max McCrea. Originally published 2025-02-19 on the [Monadical blog](https://monadical.com/blog.html).* </center> We are in the midst of a fundamental shift in how data creates value. While traditional structured data remains important, the rise of large language models has unlocked extraordinary value in *unstructured* data: everyday conversations, informal documentation, and the natural flow of work itself. Companies like Reddit and Twitter, which changed their API policies this past year to charge AI companies for the conversations happening on their platforms, have already recognized this. All company communications and processes as valuable data assets that deserve protection and strategic management. ## The AI-Driven Paradigm Shift in Data Value Traditionally, companies focused on structured data: sales figures, customer records, inventory counts. But tomorrow's AI can extract profound value from unstructured data, such as: - Internal chat conversations - Video calls - Email threads - Support ticket discussions - Company documentation and knowledge bases - Even water cooler conversations! Let's look at an example: > *Jo: that Japanese denim brand is asking about custom thread tension again > Sam: the one doing the vintage look? > Jo: yeah > Sam: run a test batch at 80% first - their fabric shrinks more in wash. Learned that one the hard way last season > Jo: got it, I'll have Sarah adjust the machines after the current run* This short conversation reveals: * Brand-specific material preferences * Learned industrial process adjustments * Production scheduling practices This one conversation along wouldn't be enough to build a competing company and undercut you. But the sum total of the interactions between your company's employees is tantamout to the company itself. Think about how every casual exchange contains tribal knowledge, hard-won lessons, and operational insights. Each conversation is a tiny fragment of your company's collective intelligence. What LLMs have shown us is that a staggering amount of structure can emerge from large collections of unstructured fragments of knowledge like this. ## AI is Double-Edged: Opportunity and Existential Risk ### The Promise of AI Leverage Even if LLMs stop improving tomorrow (they won't), we've only yet scratched the surface of what AI can do for us. If you're not interacting regularly with a chat interface like Claude, Gemini, or ChatGPT, do yourself a favor and give them a try today. Ask them to explain a concept in a field you're not very familiar with. The breadth of knowledge they are able to synthesize is breathtaking. With the lower of language models, all of the following is already achievable: - Institutional knowledge becomes indexed and searchable - New employees can quickly tap into years of accumulated wisdom - Response times and solution quality can improve dramatically as relevant knowledge can be summarized and presented to workers - Meeting summaries automatically capture and distribute key decisions and action items - Documentation can update itself based on employee conversations and actions - Product teams can extract customer needs from large databases of support calls, product reviews, or sales call transcripts - Factory floor conversations automatically surface equipment issues and maintenance insights that sensors and logs might miss And soon, we'll be able to do much more: - Real-time process optimization suggestions based on observed patterns - Predictive resource allocation based on historical project patterns - AI mediators that help resolve conflicts by providing objective historical context - Systems that can explain not just what the company policy is, but why it is that way, under what conditions it might need revisiting, etc - Cross-team knowledge pollination that happens organically through AI intermediaries All this is ignoring gains that will come from step changes in technology! ### The Existential Threat On the other hand, the same AI capabilities that make your data valuable make it dangerous in others' hands. Competitors with access to your data will eventually be able to replicate years of hard-won insights instantly, decode your operational advantages, and essentially download your company's collective expertise. A "business moat" around institutional knowledge and processes will become far more fragile. And unlike traditional data breaches, AI makes unstructured data immediately actionable. A single dump of historical communications could transfer core competitive advantages forever. Ultimately the moment data leaves your domain of control, there is no way to protect your competitive advantages from being lost. And at the center of all of this is a core problem of incentive alignment: **Many platforms are already quietly monetizing your data, and as the value of your company's data skyrockets with the power of AI, so too will the financial incentive to monetize your company's processes.** It doesn't take a malicious organization. There are accidental data leaks all the time, companies get hacked (and as the value of doing so rises, so too will the sophistication behind those attempts), or there could be one enterprising, unscrupulous system administrator with access to all of your company's data at a platform company selling your organization's secrets at a fraction of their value. ## Strategic Considerations ### Communications as Infrastructure Organizations need to fundamentally reimagine communications as critical infrastructure. This means developing systematic approaches to capturing and organizing valuable information exchanges. It's not enough to simply archive conversations; companies need intelligent systems for cataloging and accessing historical communications, with particular emphasis on preserving the context and reasoning behind key decisions. ### Data Security Self-hosted communication platforms aren't just safer, they're becoming essential for protecting your competitive advantage. While platforms like Slack claim data rights through evolving privacy policies, open-source alternatives like [Zulip](https://zulip.com/) offer full control over your intellectual property. Carefully consider your organization's information security model. Consider: - Could you describe your security risk assessment framework? - Who owns conversations and derived insights? - Who owns the systems through which your data passes? - How is data encrypted in transit and storage? - Which employees, systems, and vendors can access what information, under what conditions? How reliably can you test this? - What happens if vendors change their terms? Remember: Once data leaves your control, there's no way to guarantee its protection. A single privacy policy update could transform your company's conversations into training data for competitors' AI systems. ## Conclusion AI has fundamentally changed how we should think about company data. Every interaction, every conversation, and every solved problem now represents valuable data that could drive future innovation and competitive advantage. Companies that understand this shift and take steps to preserve and control their communication data will be better positioned to leverage AI for competitive advantage in the future. Own your data and own your business.