What it means
Many valuable decisions in business are made by experienced people who rely on judgement they find hard to explain. A knowledge engineer interviews these specialists, studies their documents and cases, and translates what they know into explicit rules, relationships and facts.
The result is stored in a knowledge base, which a program can use to reason about new situations. A classic expert system has two main parts: the knowledge base and an inference engine.
The knowledge base holds rules such as if the applicant's debt is above a set share of income and their history is short, then refer the case to a person. The inference engine works through those rules to reach a conclusion and, importantly, can explain how it got there.
In finance, early expert systems were used for tasks such as loan approval, credit assessment, tax planning, insurance underwriting and audit support. They allowed firms to apply the same standards consistently across many cases and to make the knowledge of a few senior staff available to everyone.
Rule-based tools remain widely used wherever decisions must follow published policies. Knowledge engineering has changed with modern machine learning, which finds patterns in data instead of relying on hand-written rules.
Even so, structured knowledge is still valuable, because rules are transparent and easy to audit. Many current systems combine the two, using learned models for prediction and explicit rules for regulation and policy.
The process has typical challenges. Experts may not agree, their knowledge may be incomplete or out of date, and the rules can become so numerous that they are hard to maintain.
A system is only as reliable as the knowledge placed in it, so testing against real cases and regular review are essential. For managers, the lesson is that automating a decision starts with understanding it.
Writing down how a good decision is made often reveals gaps, inconsistencies and unwritten shortcuts, and that insight can be valuable even before any software is built. Clear documentation also makes it easier to meet audit and regulatory expectations.
In practice
Real-world examples.
Example
A bank wants to speed up small business loan decisions. A knowledge engineer interviews its senior underwriters and records the rules they apply, such as minimum cash flow cover and limits on existing debt. The rules are built into a system that approves clear cases and refers the rest to a person.
Example
A tax advisory firm captures its specialists' knowledge of a country's rules for company deductions. The resulting system asks staff a series of questions and suggests the correct treatment. New employees can handle routine questions without waiting for a senior colleague.
Example
An insurer builds a claims triage tool from the experience of its best handlers. The tool flags claims that show the warning signs of fraud and routes them to investigators. Fewer genuine claims are delayed by unnecessary checks, and investigators spend their time on the cases that matter most. The insurer reviews the rules each quarter against the cases that were confirmed as fraud.
Case study
Seen in the real world.
Atlas Credit Union is an illustrative, fictional lender that struggled with inconsistent decisions on personal loans. Two experienced officers often reached different answers on the same application, and approvals could take five days.
A knowledge engineer spent six weeks interviewing the officers and reviewing 300 past decisions. They wrote 45 rules covering income, debt, history and purpose, and tested them against another 200 cases, where the system matched the final human decision 92% of the time.
The union used the system for straightforward applications and sent the remainder to staff. Decision times fell from five days to one for most customers. The illustrative lesson was that the exercise exposed rules the officers had never written down, and it improved consistency even before the software was in use. The union also set a rule that any change to the system had to be approved by the head of lending and tested on past cases first, which kept the knowledge base accurate as policies changed.
Watch out
Common mistakes.
- Assuming experts can easily describe how they decide, when much of their skill is intuitive and takes careful questioning to uncover.
- Building a rule set once and leaving it unchanged, when laws, products and markets change and the knowledge must be updated.
- Treating rule-based systems and machine learning as opposites, when many good systems combine the two.
Questions
People also ask.
What does a knowledge engineer do?
They gather expertise from specialists and documents, organise it into rules and structures, and work with developers to build a system that applies it.
What is an expert system?
It is a computer program that uses a knowledge base and rules to make decisions or recommendations in a specialist area, and can explain its reasoning.
Is knowledge engineering still relevant?
Yes, because transparent rules are valuable where decisions must be explained to customers, auditors and regulators, and modern tools often build on structured knowledge. A finance team that can show exactly why a system approved or declined a case is in a far stronger position when a customer or regulator asks questions.
From the founder's library
Take it further with the book.
Build your financial confidence beyond this definition. Shihan's full-length guide, Accounting Fundamentals, takes the same plain-English approach and turns it into a complete, practical playbook for non-finance managers, business owners and students - with chapter-end quiz answers and presentation slides included.
25% off with code MMHQ25, applied at checkout. Priced in USD - checkout may show the equivalent in your local currency.
View the book and save 25%Related