Over the course of an average workweek, executives and workers alike say they spend a quarter of their time looking up the information needed to do their jobs. Even more than that, half, say they’ve seen teams unwittingly repeat work that someone else already did.
Businesses are wasting time, talent and resources looking up information they already have and solving problems that have already been addressed.
AI initiatives can make this reset loop more visible. As a team finishes a pilot, they should be passing along far more than their final result to the rest of the organization: who needed to be involved, where bottlenecks slowed progress, which data sources mattered, what informal workarounds were required. But if that context stays in meetings, chat threads and individuals’ memories, later teams start from scratch, look up the same information and solve the same problems.
This is where the distinction between tribal knowledge, institutional knowledge and institutional intelligence, a central concept to the AI-enabled operating model proposed in our recent eBook, is important.
What is tribal knowledge?
Tribal knowledge is the everyday lived experience people use to get work done, but it is rarely written down or captured.
It is the decision-making, operational context and subject matter expertise that resides in a person’s head or a team’s meeting notes. The person who’s always asked first, before asking the key approver in a workflow; the process step that always takes longer than expected for the same reason; the workflow that is always abandoned at the same point.
Most organizations already have that knowledge. The problem isn’t that it exists, it’s that it often isn’t shared beyond the people who learned it.
This has always been an issue, but with the advent of AI, it has become far more important to address. AI does not have any of this context unless it is explicitly given it. It can’t access shared experience or informal conversations.
To make work better, AI needs governed access to captured context so that the knowledge normally trapped in people’s heads, meetings and informal workflows can become queryable and useful by the right people across the organization.
Documentation is only half the answer
Most businesses already have plenty of documentation. They have the policies, process maps, standard operating procedures, project notes, knowledge bases and system records that explain how the work should happen, but they rarely show the whole picture. Institutional knowledge is important, but incomplete.
While a process document might show the official approved path, it doesn’t necessarily show the person everyone informally checks in with before approval is requested. A project record can show the final decisions, but fail to include the tradeoffs and delays that shaped it. Workflow diagrams show intended handoffs, but not the bottlenecks everyone already knows to plan around.
The difference between the official record of what ideally should happen and what actually does happen is a gap that AI can’t cross on its own. It needs both tribal knowledge and institutional knowledge to be effective.
Making captured knowledge useful
When tribal knowledge and institutional knowledge are captured and structured in a way that makes them accessible to anyone within an organization that needs them (while remaining compliant with appropriate governance and regulatory guardrails), institutional intelligence becomes possible. Lessons learned by previous teams are able to be applied by new ones. Problems that have already been solved become solutions for the next round.
And when AI is applied to institutional knowledge that is current and connected, it becomes easier to search, understand and apply. An authorized user can ask a question in their own natural language and get valuable context back:
- Why was this decision made?
- What have similar teams done in my position?
- Where has this process broken down previously?
- Who do I need to speak to before this project can move forward?
Compliance and regulation are still important. Some information is sensitive and some knowledge only belongs with specific teams or roles. Institutional intelligence still requires governance, permissions and controls.
Even with those appropriate guardrails, the right people should still be able to surface the right knowledge and context the moment they need it.
Putting institutional intelligence to work
Teams get a head start with fuller context when tribal knowledge is transformed into institutional intelligence. They see what’s already been tried, what’s failed and what works. They know who to talk to from the beginning of an approval workflow, instead of getting delayed two days waiting for someone to get back from PTO.
AI can make institutional knowledge easier to access, apply and improve. When the reality of how work gets done, the context behind the decisions and the relevant operational data are captured and governed in a way AI can use, employees can ask practical questions and receive relevant answers quickly. That reduces the need to dig through file trees, search old documents or chase down colleagues for answers the organization has already learned.
But AI is only as useful as what it can access, and if this foundation of institutional intelligence isn’t built, knowledge remains scattered and teams keep repeating work.
This is how AI-enabled operating models help enterprises move beyond isolated pilots, learn from each effort and pass those learnings along to other teams and projects that need it. Our eBook, "Converting Tribal Knowledge into Institutional Intelligence: An AI-Enabled Operating Model", explains how to start capturing context and structuring it in a way that makes it queryable for the entire organization.
If your teams are spending more than a quarter of their working weeks looking up needed information, or repeating work already done by others, then your organization likely has tribal knowledge waiting to be put to better use.
