Accurate Data. Wrong Decision. What’s Missing?
Why AI needs more than trusted data and context to make sound business decisions
Your AI has access to complete, accurate customer data. It understands what the data means and how it relates to other information.
So why might it still make the wrong business decision?
Because knowing your customer data isn’t the same as knowing how your business works.
Dell Technologies recently announced an approach to making enterprise data more useful to AI agents by providing trusted context.
One example particularly caught my attention. Arthur Lewis, president of Dell’s Infrastructure Solutions Group, used a customer record to illustrate the importance of understanding what data means, not simply making it accessible.
I smiled when I read that example.
In an earlier LinkedIn post, I asked a seemingly simple question:
“How many customers do we have?”
It is a question many organizations struggle to answer consistently. Sales, Finance, Customer Service, and executive management may each produce a different number.
Often, the problem is not that the data is wrong. It is that the people asking and answering the question mean different things by customer.
Dell’s emphasis on context is an important part of addressing that problem. But I believe it also points toward a broader challenge that organizations need to confront as they adopt AI.
Even if AI understands what our business data means, does it know the business rules that govern how we use it?
A customer is not always the same customer
Consider the word customer.
Sales might recognize a customer when a contract is signed. Finance might count only customers who generated revenue during a reporting period. Customer Service might work with individual account holders, while management reports count corporate groups.
Each perspective may be legitimate.
The difficulty arises when everyone uses the same word without making the distinctions explicit.
If we ask AI, “How many customers do we have?”, it needs more than access to a customer database. It needs to know which meaning of customer applies to the question.
That requires business vocabulary and definitions that people across the organization understand and agree upon.
The same issue arises with words such as product, account, active, eligible, and revenue. Their meanings may appear obvious until different departments try to use them together.
Making data available to AI does not, by itself, resolve those differences.
Understanding the customer is only the beginning
Now consider a different question:
“Which customers should receive a renewal discount?”
Imagine that our AI agent has access to a complete, accurate customer record. It knows the customer’s identity, purchase history, contract details, and payment status.
That is useful information.
But can the agent make the correct recommendation?
To do so, it must know the organization’s business rules.
For example:
- Which customers are eligible for a renewal discount?
- Which products and contracts qualify?
- Does an overdue payment affect eligibility?
- Can a renewal discount be combined with another offer?
- Who has authority to approve an exception?
These are not merely questions about the contents of a customer record.
They are questions about how the organization has decided to conduct its business.
An AI agent could have accurate data and still recommend the wrong discount because it applied an incorrect, outdated, or incomplete business rule.
That is a different problem from data quality.
It is a business knowledge problem.
Business knowledge is more than data and documents
Organizations have spent years investing in databases, systems, data governance, and document repositories.
Those investments are important. But the knowledge that gives them business meaning is often scattered.
Definitions may live in people’s heads. Policies may be written in documents. Business rules may be embedded in application code, spreadsheets, procedures, or individual decisions.
Different departments may have different interpretations of the same concept or policy.
AI can retrieve information from these sources. But retrieving a statement does not establish whether it expresses the organization’s intended meaning, whether the rule is current, or whether it applies to the situation at hand.
This is why we need to recognize business knowledge as an organizational asset in its own right.
Business knowledge includes the vocabulary and concepts an organization uses, the relationships among those concepts, and the policies, business rules, and decisions that guide its operations.
For AI to use that knowledge reliably, the organization must make it explicit, understandable, validated, and maintainable.
Who is responsible for teaching AI the business?
Technology can help us connect information, retrieve documents, and make knowledge accessible.
But technology alone cannot decide what an organization’s business policies ought to be.
That responsibility belongs to the business.
Subject-matter experts must clarify the meanings of important concepts. Business owners must validate the rules that govern decisions. The organization must establish who is responsible for keeping that knowledge current.
This is not work that begins and ends with an AI implementation.
It is an ongoing business capability.
And it benefits people just as much as AI.
When definitions are clear, departments can reconcile conflicting reports. When rules are explicit, employees can make more consistent decisions and explain why those decisions were made. When business knowledge is shared, onboarding and collaboration become easier.
The same foundation that helps people work together also helps organizations teach AI how their business operates.
From trusted data to trustworthy decisions
Dell’s announcement is a useful reminder that AI needs context, not just access to data.
I would take that thinking one step further.
Context helps an AI system understand what information means.
Business rules help it understand what the organization permits, requires, or prohibits.
Both matter.
As organizations prepare for AI, the questions should not stop at:
“Can our AI access the right data?”
We should also ask:
“Can we explain what that data means in our business?”
“Have we made the relevant business rules explicit?”
“And can we trust that those rules reflect how our organization intends to operate?”
These are questions BRS has been helping organizations address for more than 30 years.
AI has made the need more urgent. It has not made the underlying work less important.
Before You Build AI… Teach AI Your Business.
Explore the RonBot Learning Experience
At Business Rule Solutions, we help organizations develop the shared vocabulary, clear definitions, and explicit business rules needed for consistent decisions and AI readiness.
The RonBot Learning Experience helps people learn how to think critically about business knowledge, examine definitions and rules, and develop shared understanding.
Reference: Dell Technologies, “Dell Technologies Turns Enterprise Data Into Trusted Context for AI Agents.”
Gladys S.W. Lam is Co-Founder and Principal of Business Rule Solutions and co-author of Building Business Solutions: Business Analysis with Business Rules.
Back to all resources