Collaborative Knowledge Management: Intelligent Organizations That Learn
Tom has always been interested in collective intelligence, particularly in how technology can help people get smarter together. His first company, Intraspect, was an attempt to do this for the enterprise. The problem is simple yet gnarly: Human intelligence doesn’t naturally scale past groups of the size that we evolved to live in (extended families and tribes). Two heads are better than one — but what happens when we combine the minds of dozens, hundreds, or thousands of people into an organization that needs to think as one? Unless they are extremely well organized, most large “teams” of size N are not N times more knowledgeable or competent than groups that can fit around a table.
Tom has training in both AI and Cognitive Science. AI is good for making technology that thinks, and cognitive science is good at describing how people think. As he examined the problem of scaling human intelligence, Tom wondered what it would be like to treat a modern thinking organization — large groups of knowledge workers collaborating on projects within and across corporations — as a single cognitive unit. What would it need to be smarter? At the very least, an intelligent organization needs to be a learning organization. Today people scramble to know what the organization knows, and as people move into and out of a group, or change companies, the organization doesn’t have a way to remember what it learns. Essential to learning is memory.
Tom created Intraspect as the first step toward collective intelligence: an enterprise-scale group memory. The easy part about augmenting memory is that computers do it quite well. We have documents and databases and all sorts of clever ways of searching them. But most of the knowledge that knowledge workers create is not found in corporate repositories and databases. It is found in all the little documents that are passed around; in the tasks, projects, promises and reports that constitute communications among people in organizations. When Intraspect was founded in 1995, at the dawn of the Web, most documents lived on individual PCs and most communication was in isolated email threads. In order to create a group memory, Intraspect had to solve the problem of how to get the knowledge out of individual silos and into the group memory.
The solution was to create a suite of collaboration tools, based on the ideas emerging from the early Web about threaded discussions, hyperlinked content, and really good search. These tools allowed people who were just beginning to experience email in the workplace to collaborate in virtual spaces in which everything they did was kept and could be shared. As a consequence of working online together, they contributed to a collective knowledge of what they knew and knew how to do. This collective knowledge could be browsed like Wikipedia (which hadn’t been invented yet) and could be searched as well as Google (which was just getting started). This meant that people could always benefit from prior work on any task they faced, and — just as important — discover other people in the organization who could help. In effect, the group memory grew as a by-product of people working together online, and the organization learned.
The idea proved to work. Over eight years, hundreds of thousands of users enjoyed an enterprise scale product that was implemented and refined by over 200 Intraspect employees. Over 100 of the Fortune 500 were customers. Companies used Intraspect to create group memories and associated enterprise applications for colleagues working in financial services, marketing services, and professional technology services. Wherever a group of people brought their collective knowledge to bear, Intraspect’s applications could help them do it better.
On the cusp of an IPO, the company fell victim to the macroeconomics of 2001, when the dot-com recession destroyed the appetite for new enterprise software. Enterprise search, document management, portals, and collaboration software all consolidated and then imploded. Today, we see the resurrection of something similar in the ecosystem that integrates Slack, Microsoft Teams, Google Docs, and Dropbox/Box. Tom continues to advise companies that are completing the vision such as Cognizer, a startup creating a new category of semantic search.
Intraspect’s 2001 white paper.
Intraspect’s 1998 white paper.
First public demonstration of prototype at the Workshop on WWW and Collaboration, Sept 11, 1995, MIT/W3C, Cambridge, Massachusetts.
Early Talk on organizational intelligence that makes the analogy between modern organizations and herds of prey animals – as agents of collective intelligence. And it has a Scott Adams cartoon. A real period piece.