An experienced manager is away when a commercial property group’s major tenant asks to renew. The contract is stored in one system, the rent schedule in a spreadsheet and the account history in the CRMCRM stands for Client Relationship Management. A CRM system is a digitised process of keeping your customer interactions and details.... The team can see what was agreed last time. They cannot easily see why a concession was made, which service obligation changed with it or who should approve a similar offer now.
The team calls the manager and waits. The delay reveals a common problem in established businesses: critical knowledge lives in people, while the systems used by the wider team contain only part of the story.
Quick answer: To capture business knowledge, start with a recurring decision that depends on an experienced person. Review real cases, record the evidence and reasoning behind each outcome, define who can approve exceptions, and place the current guidance inside the workflow where the decision is made. Give the guidance a named owner and review date so staff and AIArtificial Intelligence (AI) is a very large and broad spectrum of technologies which most people would be familiar with through... tools can distinguish trusted rules from old examples.
What Business Knowledge Is Usually Missing from Systems?
Most systems are designed to record transactions, documents, dates and status. A contract shows the final terms. A work order shows a completed job. A credit note shows the amount authorised. The decision criteria behind those outcomes may be spread across email, conversations and individual memory.
The missing knowledge often includes:
- Why an exception was approved and whether it should be repeated
- Which facts an expert checks before applying a rule
- What changes an approval threshold or escalation path
- Which customer history matters to the decision
- Which source is authoritative when records disagree
- When a policy or past decision stopped being current
This knowledge shapes customer experience, margin and risk. When teams cannot access it at the point of work, they repeatedly ask the same people for help, make inconsistent calls or delay routine decisions. The business also becomes more exposed when a key person takes leave, changes role or departs.
IBM’s discussion of knowledge fragmentation describes how information can exist across an organisation while remaining difficult to use coherently. The leadership question is whether a person can find the right information, understand its context and make a sound decision without reconstructing the history from memory.
How Do You Capture Expert Judgement Without Flattening It?
A useful starting point is a decision that is frequent enough to study and consequential enough to matter. It could be a renewal concession, a credit exception, a supplier substitution or a customer escalation. Choose one with a clear business owner and a measurable cost when the decision is slow or inconsistent.
Ask the experienced people involved to walk through three types of case: a straightforward decision, a difficult one and a past decision they would handle differently today. For each case, record the question that triggered the decision, the evidence checked, the rule applied, the trade-off considered, the approval required and the final rationale. This reveals conditions that a generic procedure often misses.
Then test the proposed guidance with another experienced person and with someone who was not part of the original decision. If they interpret a threshold differently or cannot find the required evidence, refine the guidance. Leave room for human judgement in unusual cases, with a clear route to review and record the exception.
The output should be concise enough to use during work. A long interview transcript may be valuable source material, but a current decision guide or structured record is easier to act on.
A Practical Example: A Commercial Lease Renewal
Imagine a property group managing several commercial sites. A major tenant asks for a lower annual increase in exchange for a longer lease term. The account team sees commercial value in the request. It also affects service obligations and the group’s future flexibility.
The latest contract, correspondence and budget are available, but the decision depends on questions the business has handled informally. Has the tenant met its current obligations? Do planned site works change the economics? Does the longer term support the concession under current margin targets? What can the account manager propose, and when must finance, property or legal teams review it?
An experienced commercial manager works through these factors and knows that a similar concession in an older agreement was a one-off response to unusual circumstances. That distinction belongs in the team’s decision process. Without it, a past outcome can look like a current precedent.
A structured decision record might capture:
| Field | What it should show |
| Trigger | The tenant’s requested change and decision deadline |
| Current sources | Approved contract version, property plan, account history and current financial view |
| Decision rule | The normal concession range and how it is calculated |
| Exception reason | The circumstances that justify review outside the normal range |
| Approval owner | The person or group authorised to accept the commercial change |
| Outcome and rationale | What was decided, why, when and with what follow-up |
The record gives the next decision-maker a reliable starting point. It also helps leadership see which rules are used frequently, where exceptions cluster and whether a policy needs revision. This is an illustrative example, not an Aerion client result.
Where Should Captured Knowledge Live?
Put the guidance as close as practical to the decision. If a renewal is prepared in a CRMCRM stands for Client Relationship Management. A CRM system is a digitised process of keeping your customer interactions and details... workflow, the current criteria, relevant sources and approval route should be accessible there. A separate knowledge base can hold fuller explanations, but staff should not have to search several repositories to discover which rule applies.
Each guide or decision rule needs an owner, an effective date, a link to its source and a route for corrections. Permissions should reflect the sensitivity of commercial and customer information. When a policy, market condition or customer contract changes, the owner should review the guidance and mark superseded versions clearly.
This is where software design becomes important. Connected records, clear source authority and an audit trail make knowledge usable across teams. Aerion’s custom software development work can help when existing tools cannot present the right context or route approvals reliably.
How Does Better Knowledge Improve AI Use?
An AIArtificial Intelligence (AI) is a very large and broad spectrum of technologies which most people would be familiar with through... assistant can retrieve documents and draft a decision brief. Its usefulness depends on the business context around those documents. A previous tenant concession may be a one-off exception; a draft policy may no longer be current; an account record may be visible to one role but restricted for another.
IBM’s analysis of enterprise AI context highlights source authority, lineage, business definitions and permissions. Those are practical design requirements for an AIArtificial Intelligence (AI) is a very large and broad spectrum of technologies which most people would be familiar with through... tool used in a commercial workflow. The assistant should be able to show which records it used, surface uncertainty and send material decisions to a person with the authority to approve them.
In the lease renewal example, AIArtificial Intelligence (AI) is a very large and broad spectrum of technologies which most people would be familiar with through... could gather the current contract, summarise tenant history, identify the relevant concession rule and list missing evidence. The manager would review the brief, make the decision and record its rationale. The next case then has a more complete history to draw on.
McKinsey’s 2026 work on developing expertise in the age of AI also points to the need to connect knowledge management, role design, learning and coaching. Capturing decision logic can help newer staff develop judgement while allowing experienced people to focus on the cases that genuinely need them.
A Four-Step Starting Plan
1. Select one high-value decision. Choose a recurring case with a clear owner, meaningful commercial effect and visible delays or inconsistencies.
2. Capture the reasoning from real cases. Compare ordinary, difficult and outdated examples. Document sources, criteria, exceptions and approval rights.
3. Validate it in the workflow. Ask other decision-makers to apply the guiddecisionmakers Check whether they can find current evidence and reach a consistent result.
4. Maintain and measure it. Assign ownership and review dates. Track decision time, repeat questions, rework, exceptions and relevant commercial outcomes.
This limited pilot gives leadership evidence for a wider programme. It can show whether the main need is clearer policy, better-connected systems, a revised approval workflow or carefully governed AIArtificial Intelligence (AI) is a very large and broad spectrum of technologies which most people would be familiar with through... support. Aerion’s DevReady process provides a way to examine those options before committing to a larger build.
Make Expertise Available Where Work Happens
Established businesses have built valuable judgement through years of service, exceptions and commercial decisions. When that reasoning is captured, maintained and connected to daily workflows, teams can act more consistently and leaders can reduce dependence on a few individuals. The same foundations make AIArtificial Intelligence (AI) is a very large and broad spectrum of technologies which most people would be familiar with through... assistance more useful and governable.
Aerion helps businesses turn critical knowledge into secure, scalable software and clearer processes. If important decisions depend on information your team cannot readily find or verify, book a DevReady consultation with Aerion to identify where a practical change could create the most value.
FAQs
What is tacit knowledge in a business?
Tacit knowledge is practical know-how that people apply through experience but may not have written down. It often includes judgement about exceptions, customer context, warning signs and trade-offs. Capturing it begins by examining real decisions and the reasoning behind them.
How can a business reduce key-person risk?
Identify decisions and relationships that depend heavily on one person. Record the sources and criteria they use, assign a second capable reviewer, put current guidance into the workflow and test whether others can handle routine cases. Review that guidance as the business changes.
Is a knowledge base enough for AIArtificial Intelligence (AI) is a very large and broad spectrum of technologies which most people would be familiar with through...?
A knowledge base helps when its material is current, well-owned and accessible to the right people. AIArtificial Intelligence (AI) is a very large and broad spectrum of technologies which most people would be familiar with through... used for business decisions also needs context about source authority, definitions, permissions, exceptions and approval boundaries. Human review remains important for consequential outcomes.
Which decisions should be captured first?
Begin with decisions that recur, affect revenue or risk, and currently cause delays or inconsistent outcomes. A focused pilot is easier to validate and measure than a broad attempt to document everything at once.
