When organizations confront complex technology problems, their first instinct is often to search for a better system. The assumption is that a technological deficiency requires a technological remedy.
That assumption is frequently wrong.
Technology operates within an organization whose processes, responsibilities, and objectives determine what its systems must accomplish. A failure attributed to technology may originate in the system itself, but it may also arise from conflicting business requirements or processes that were never adequately defined.
Solving the problem therefore requires more than technical knowledge. It requires an accurate understanding of the organization and the relationships among its systems.
That understanding begins with discovery.
The Complexity of Replacing a Legacy System
A hospitality company recently engaged Oxer to help replace its legacy reservation and property management system.
The existing system served as a central component of the company’s operations. It exchanged information with telephone systems, televisions, and electronic door locks. It also integrated with customer relationship management and data collection systems.
Replacing the central application without understanding these dependencies would risk disrupting systems that continued to function correctly.
Oxer began by documenting the existing architecture and tracing the flow of information among its components. Equally important was determining how each department used the system and what it required from a replacement.
The investigation extended beyond current functionality. Future operational requirements had to be established before alternative products could be evaluated.
The engagement remains in progress. Oxer has identified a replacement SaaS platform that satisfies all future requirements identified during discovery.
The significant result was not simply the selection of a new product. It was establishing what the product needed to accomplish.
The Problem May Not Be Technical
Some of the most consequential findings emerged from discussions among departments rather than from examining software.
Each department understood its immediate responsibilities and the limitations of the existing system. Yet their requirements had developed independently, sometimes in direct conflict.
No one had systematically reconciled them.
Once these differences became explicit, several longstanding operational problems were resolved through relatively simple changes. No additional technology was required.
This illustrates a broader principle. An organization is not merely a collection of departments using common software. It is a system of interdependent processes. A decision that improves one process may impair another.
Technology cannot resolve a contradiction in business requirements that the organization itself has not addressed.
Discovery must therefore examine not only how individual departments operate, but how their requirements interact.
The Importance of Asking the Right Questions
Discovery is fundamentally an exercise in disciplined inquiry.
Its quality depends on determining which questions matter, identifying who possesses the relevant knowledge, and recognizing when an answer is incomplete.
These abilities depend heavily on experience.
An experienced engineer knows that an apparently straightforward answer may conceal an assumption. A requirement may describe an existing procedure rather than an actual business necessity. A technical limitation may be the consequence of an earlier decision whose rationale no longer applies.
The most productive questions frequently reveal something the organization had not previously considered.
Experience also broadens the range of possible solutions. Oxer’s work across different industries provides exposure to architectures and operational practices that may be unfamiliar to a particular client.
A problem encountered in one industry may already have an effective solution in another. Recognizing that equivalence requires understanding the underlying problem rather than merely its industry context.
AI Changes the Economics of Discovery
Discovery has always required extensive documentation. Historically, the cost of producing and maintaining that documentation limited its practical depth.
Artificial intelligence has changed that constraint.
Oxer conducts detailed meetings with the stakeholders responsible for each business function, typically using Zoom or Microsoft Teams. AI tools analyze the transcripts and help construct structured descriptions of workflows, business rules, and system requirements.
The documentation also captures why earlier decisions were made. This distinction matters because the reason for a decision may disappear while the resulting procedure remains in place.
AI assists in identifying inconsistencies and unanswered questions. Those findings guide subsequent discussions, allowing the documentation to become progressively more complete.
The process is iterative. Each round of inquiry refines the understanding of the organization and exposes questions that could not have been formulated at the outset.
Detailed diagrams make the resulting processes easier to examine. A commercial lease transaction, for example, may involve six parties and more than a dozen steps before the lease is executed.
The diagram illustrates a general difficulty with complex business processes. Each participant may understand an individual responsibility without understanding the complete sequence or its dependencies.
Making those dependencies explicit allows the process to be evaluated as a whole.
Documentation of this depth once required substantial manual effort. AI now makes it practical to produce more comprehensive records and maintain them as systems evolve.
The improvement is not merely faster documentation. It is the ability to investigate and preserve details that would otherwise remain undocumented.
Understanding the Business Before Examining the Code
Legacy software presents a related problem.
Its source code may be available even when the original developers have departed and the business rules governing its behavior are no longer documented.
AI can assist in analyzing that code, identifying defects, and detecting security vulnerabilities. It can also help reconstruct the program’s structure and describe its observable behavior.
But understanding what software does is not equivalent to understanding what it should do.
The distinction is essential.
A program may execute correctly while implementing an obsolete business rule. Conversely, behavior that appears incorrect to an outside reviewer may serve a legitimate requirement that has not yet been explained.
For this reason, Oxer first establishes the relevant business rules and then evaluates the software against them.
Only then is it possible to judge whether the existing system should be corrected, modernized, or replaced.
The same method extends to cloud infrastructure and SaaS platforms. AI can assist in examining their actual configurations through administrative interfaces and command-line tools, producing documentation grounded in the systems as they exist.
The governing principle remains unchanged. Technical behavior must be evaluated against defined requirements.
Tools Do Not Replace Judgment
AI has substantially increased the amount of information that can be collected, organized, and analyzed during discovery.
It has not eliminated the need to determine which information matters.
An AI system can identify an inconsistency in two documented requirements. It cannot independently establish which requirement best serves the organization’s objectives. It can explain how software operates, but the correctness of that operation depends on business context.
These are questions of judgment.
Two firms using identical tools and following similar procedures may reach different conclusions because their engineers differ in experience and understanding.
The value of discovery is therefore not proportional to the volume of documentation it produces. Its value lies in the accuracy of the conclusions that documentation supports.
AI makes a more thorough investigation economically practical. Experienced engineers determine what to investigate and what the findings mean.
Understanding Precedes the Solution
Complex technology problems often persist because organizations attempt to solve them before establishing precisely what is wrong.
The consequences can be substantial. A replacement system may reproduce the deficiencies of its predecessor. An expensive integration may automate a process that should have been changed. A technically successful implementation may fail to satisfy the people who depend on it.
These outcomes are not necessarily failures of technology. They are failures to define the problem adequately.
Effective discovery reverses that sequence. It establishes the business requirements, examines the existing environment, and identifies the causes of its deficiencies. Only then does it evaluate possible remedies.
The method is straightforward. Its successful application requires experience, technical competence, and disciplined inquiry.
Does your organization depend on systems that no one fully understands, or face technology decisions whose requirements remain unresolved? Oxer can help solve them.