Knowledge and authority in AEA by Ivory & Co.Knowing something and having the authority to determine it are two different things.
Information, evidence, interpretation and an authorised determination are different kinds of knowledge. An enterprise needs to know which is which before it relies on any of them.
Somewhere in a business, there is a document that everyone uses.
It might be a pricing schedule, an operating procedure, a customer service policy or a template for approving expenditure. It has been copied between departments, attached to emails, uploaded to shared drives and referenced in training. People trust it because it is familiar. It has been used for years.
Then someone asks a perfectly reasonable question. Who approved this?
Nobody knows. There is no approval record and no clear owner. Nothing says whether the document is current. The person who created it left the business three years ago. And the organisation has been making decisions with it the whole time.
The document may contain excellent information. It may describe exactly how the business intends to operate. Its popularity still cannot establish its authority.
An enterprise can possess information, understand it, rely on it and act on it. Each of those raises a different question, and once knowledge starts to shape consequential decisions the differences matter a great deal.
The problem with information abundance
Modern businesses are exceptionally good at accumulating information: documents, emails, databases, dashboards, reports, meeting notes, policies, procedures and presentations. With collaboration platforms, cloud storage and AI-powered search, an enterprise can retrieve more about itself than at any point in its history.
Retrieval creates a problem of its own. Finding an answer does not establish whether that answer should govern a decision.
A manager asks an AI assistant, “What is our policy for approving customer refunds?” The assistant searches the organisation’s repositories and returns a confident answer, citing a document titled Customer Refund Policy — Final. The answer is clear and the document looks official. The policy was replaced six months ago.
The assistant retrieved the information successfully. The organisation had never established whether that information was authoritative for the question being asked. Without that, faster access simply accelerates the use of unreliable information.
Evidence can inform a decision without becoming the decision
A business reviews its supplier payment terms. The finance team examines cash flow, supplier contracts, payment history and industry practice. An analyst recommends extending standard terms from thirty to forty-five days. The recommendation is well researched and may be commercially sensible.
The analysis does not establish the new terms. The business still needs an authorised determination, and even after one is made, existing supplier agreements may continue to impose different obligations.
Five kinds of knowledge are present in this small example. The underlying records are evidence about the business and its obligations. The analysis interprets that evidence. The recommendation proposes a course of action. An authorised decision establishes what the enterprise has chosen, within the scope of that authority. The agreements and operating arrangements determine what can lawfully and practically be put into effect.
Each has value. Trouble begins when one is mistaken for another. A recommendation repeated often enough becomes accepted practice. A draft acquires the look of an approved policy. An interpretation circulates until nobody remembers that it began as an interpretation. Eventually the enterprise is governed by assumptions it never deliberately adopted.
Authority must have a source
Where does the authority of business knowledge come from? Sometimes from law or regulation. Sometimes from contractual obligations. Sometimes from the organisation’s own decision-making arrangements. Within a professional or intellectual discipline, it comes from the standards and governance through which that discipline determines and maintains what it holds to be true.
The source matters because authority has boundaries. A regulator establishes obligations within its jurisdiction. A board approves certain enterprise decisions. A manager acts within defined limits. A technical specialist can be the authoritative voice within a professional field without holding the organisational authority to approve every action that depends on that expertise.
These boundaries stop an enterprise treating every credible source as though it carried the same power to decide. They also stop authority being inferred from a person’s seniority, a document’s location or a system’s ability to execute an instruction.
Four questions test the authority of any piece of knowledge. Who or what established it? What does it apply to? Is it current? What conditions or limits govern its use?
A document is a representation of knowledge
Consider the familiar company policy. It has a title, a version number, an approval date and a handsome template. Those features help people recognise and use it. They do not create the authority behind it.
A document may accurately express an authorised policy. It may equally hold a draft proposal, an outdated instruction, an incomplete interpretation or information assembled from several sources. The same goes for dashboards, reports and AI-generated answers. Each can be useful, and the authority of the underlying knowledge cannot safely be read from the polish of the presentation.
This matters more as AI makes convincing material easy to produce. A well-written explanation may be entirely accurate. It may also be plausible, outdated or incomplete. The ability to produce an answer is becoming abundant. The ability to establish what may legitimately be relied upon remains essential.
A single source of truth: when the same knowledge appears in different places
An organisation keeps its customer returns policy in three places. The official policy is held in a controlled repository. Customer service has a simplified version in its knowledge base. The website carries a customer-facing explanation.
Each serves a different purpose, and they can legitimately differ in detail and expression. They must agree wherever they describe the same governing requirements.
Now the official policy changes. The website is updated immediately. The customer service knowledge base is missed. Two parts of the enterprise begin giving different answers.
The problem is larger than document maintenance. The business has lost coherence between representations that people are using to act. Fixing it means knowing which knowledge is authoritative, which representations depend on it and where the consequences of inconsistency land. Knowledge governance has to preserve the relationships between authoritative knowledge and the ways it is communicated and used.
Authority and currency are separate concerns
A document can be authoritative and outdated. It can also be recent and entirely unauthorised.
An operations manager writes a revised procedure to deal with a recurring service problem. The revision is excellent. It reflects current conditions and fixes weaknesses in the existing process. It has not been approved for use, and the previous procedure remains the authorised version.
The organisation now holds two pieces of knowledge with different qualities. One has established authority and may no longer be suitable. The other may be more accurate and more useful, and has not acquired the authority needed for implementation.
An enterprise that cannot recognise both conditions drifts towards one of two outcomes. It becomes rigid, applying unsuitable knowledge because it is officially approved. Or it becomes inconsistent, letting local improvements acquire authority without anyone considering their wider consequences. A disciplined approach makes room for challenge while protecting the integrity of consequential decisions.
Knowledge by repetition
Organisations develop habits. A particular report becomes the basis for monthly decisions. A spreadsheet is treated as the definitive financial position. An experienced employee’s reading of a policy becomes the accepted explanation. A procedure is copied into new departments because it worked somewhere else. People stop asking where the information came from, and familiarity starts to stand in for validation.
A company has used the same customer profitability model for several years. It shapes pricing, account management and investment, and it is widely trusted. Then a new analyst discovers that certain distribution costs were never included.
The model was applied consistently. Its conclusions were materially incomplete. The many decisions made with it are no evidence that those decisions were well informed.
This is why provenance matters: where information originated, how it was developed, what assumptions it contains and whether its conclusions are still supported. A history of use is relevant evidence. Validity and authority have to be established separately.
When AI becomes a participant
The distinction between knowledge and authority becomes sharper when AI moves beyond answering questions.
An agent is asked to assess a customer’s eligibility for a refund. It retrieves the purchase history, the product conditions and a document describing the returns policy. It determines that the customer qualifies and initiates the refund.
For that action to be legitimate, five things need to be true. The information must be appropriate for the decision. The policy must be applicable and authoritative. The agent must interpret the conditions correctly. The action must fall within the authority and operating conditions set for that participant. And the enterprise must be able to reconstruct what happened if the outcome is questioned.
Each can fail separately. The agent might apply an outdated policy. It might use a valid policy in the wrong context. It might assess correctly and act beyond its authority. It might act correctly and leave the enterprise unable to show the basis of its decision. These are different problems and they need different responses.
As agents become more capable, telling reliable information, authoritative knowledge and permitted action apart becomes part of the foundation for automation that can be trusted.
Knowledge should be open to challenge
Authority is essential, and authoritative knowledge must still be able to develop.
Suppose a professional standard establishes an accepted approach to a problem, and new evidence challenges one of its assumptions. The evidence deserves examination. It may eventually justify a revision. It does not rewrite the standard by existing. Identifying a possible deficiency and authoritatively deciding how to resolve it are separate acts.
Without a means of challenge, authority becomes rigid. Without a means of determination, knowledge becomes unstable. A developing discipline needs both.
This applies directly to Adaptive Enterprise Architecture. AEA is an evolving discipline. Its research may question existing propositions, and its application may expose limits or circumstances that need further examination. Those discoveries are valuable, and they do not by themselves change the discipline’s standards. The integrity of the Body of Knowledge depends on holding that line.
A discipline’s knowledge and a business’s knowledge
There is one more boundary worth understanding. The authoritative knowledge of a discipline is different from the knowledge produced by applying that discipline to a particular enterprise.
A business uses AEA to examine its customer service arrangements. It establishes that its current model depends on particular employees, technologies and providers. It identifies weaknesses and proposes a different arrangement. Those findings concern that enterprise. They do not become universal principles, and one effective way of applying AEA does not become the only legitimate way.
Disciplines develop through both theory and practice. Application supplies evidence, reveals limits and generates research questions. A successful example, however persuasive, does not establish doctrine.
Ivory & Co., as steward of Adaptive Enterprise Architecture, maintains the distinction between the discipline’s governed foundations and the knowledge produced through its application. That is what allows the discipline to develop without any single implementation, commercial product or interpretation quietly redefining it.
What an enterprise should be able to know about its knowledge
Return to the document everyone uses, the one nobody remembers approving.
Imagine the business can establish its origin, its purpose, whether it still applies and how it relates to the authoritative knowledge it represents. It can tell whether the content is approved, proposed, superseded or subject to open questions. It can separate the evidence supporting a conclusion from the determination that gives the conclusion effect. It knows where the knowledge is being used and what will be affected when it changes. When someone challenges it, the organisation can examine the challenge without losing track of what is currently established.
That is a very different capability from storing and retrieving documents. It supports more reliable decisions, reduces ambiguity between functions, makes change easier to examine and gives a sounder footing to technologies that increasingly consume knowledge and act on it.
The more capable our systems become at finding answers and taking action, the more it matters which answers deserve authority. Knowledge becomes powerful when we can understand it. It becomes consequential when we act on it. Authority determines the legitimacy of that action.
Publication and intellectual property
Ivory & Co. holds both published and unpublished intellectual material. A public summary does not release the Body of Knowledge behind it.
On this site
The published standards, and the public pages of this site.
The platforms
The full Body of Knowledge. Paid access, through certification and the platforms. Not published openly.
Not released
Detailed models and research Ivory & Co. has not released, even through certification.
Where to go next.
Body of Knowledge
The Body of Knowledge already exists and is stewarded by Ivory & Co. It is not published on this site. The standards here are public. The full Body of Knowledge is available through certification and the platforms.
Standards
The AEA Standards establish the architectural foundations of Adaptive Enterprise Architecture: its governing concepts, the distinctions between architectural concerns, and the conditions under which those concerns are understood together.
Consequences
A proposal, an authorised decision, its assessed consequences and what actually changes are four different things. Keeping them apart lets a business see what a choice will reach before it commits, and learn honestly from what follows.