Last week I read about OpenAI rumors about building a productivity suite (Thanks Brian Madden for Sharing). Here are my thoughts about the topic, and my recent studies in the area of AI / Human collaboration.

Each technological revolution enables new ways of doing things. But we tend to build new technology in the image of the old first. Then, we realize that we can do things in completely different ways, and a new «standard» is emerging. I think the «winners» of next generation productivity suites will have AI workflows natively built in. Microsoft is trying this with Copilot in M365 but still adding AI on top of existing workflows. I find myself collaborating more and more with AI based on a case or subject, where I add more and more context as I work, and having AI complete the text based on our common understanding of the case. This works very well. I think future documents will be more live contextual parameters that we collaborate on. You chat with the document instead of reading through it. If you are «on the go» you talk or listen to your document.

We put Electric power on paper 20-30 years ago, now we make documents intelligent. The document has it’s own purpose and mind. It redefines what a document is. A container of information around a subject. Now it becomes «alive»

We have evolved form documents from physical (clay tablet 3200BC, papyrus and paper), to digital, and this moved os into the Microsoft office productivity suite, while cloud SaaS age, gave birth to more collaborative tools like Google Workspace and Office 365.

Now wiht LLM we stand before a new leap. Documents become contextual, and you can interact with in a new way. We move into contextual engineering. Contextual engineering is the discipline of architecting the entire “context window” for an AI—curating and structuring all relevant background data, tools, memory, and system instructions so the model can reliably and meaningfully complete a task.

If OpenAI is working on a productivity suite, they can redefine what a document is and how we collaborate, humans, machines around a subject, and the “document” becomes the context, the sum of parameters that we share with it, and the output can be what format we need, a paper, a chat interface, a podcast. It enables us to collaborate even more in the future.

This can make future organization more intelligent. Collective human intelligence + Machine intelligence. Vegard Kolbjørnsrud, Norwegian professor, has in his award winning paper “designing intelligent organizations” explained the principles behind how the future workforce may benefit from exactly this.

The collaboration principle focuses on interaction and the collaborative skills required from both human and digital actors. Effective human-AI collaboration depends on intuitive interfaces and AI literacy among employees

Imagine investigative journalist working together and sharing a huge amount of information, multiple journalists can share and draw information from a contextual model, learning faster, sharing faster, and producing multiple takes and views from the same model, to different platforms like podcast, video or physical paper. Or an crime investigation, a lawyer, an architect. The possibilites are enourmous. We need new productivity tools in the new era, and we need to learn contextual engineering.