A large energy sector operator needed to replace a fragmented, paper-heavy document approval process spanning multiple facilities with a single web platform, backed by an AI assistant that let teams query thousands of pages of regulatory documentation instead of waiting hours for an analyst's answer. Azati's QA engineer joined the client's testing team to validate complex document lifecycles, an extensive role-based access model, and the AI documentation assistant itself.
volumes of technical and regulatory documentation underpinning the system's requirements
distinct user roles a single document could pass through across its full lifecycle
months of testing across functional, regression, and requirements coverage
Large energy sector operators running multiple facility types, generation plants, central engineering units, thermal stations, generate an enormous volume of bureaucratic process: document approvals, deadline assignments, routing tasks to the right executor, tracking who is responsible for what at every stage. At the scale this client operates, that process had been running largely on paper and across disconnected systems, and it was buckling under its own volume.
The fix was to move documentation and the processes around it into a single web platform. Alongside that core system, the client also built a companion tool, closer to an AI agent than a simple chatbot, that let teams query the project's documentation directly instead of routing every question through an analyst. With more than 20 volumes of technical and regulatory documentation underpinning the system's requirements, that assistant solved a real bottleneck: a question that used to mean a two-hour wait for an analyst's reply could now get answered immediately.
The platform's role-based access model was unusually extensive for a system of this kind. Entire sections of the requirements documentation were dedicated to roles alone, and testing had to go well beyond checking that a feature worked:
A single document moving from creation to archival could require action from roughly ten different users in sequence, each responsible for a specific step according to their role. Validating that flow meant understanding not just the immediate task, but the document's entire journey:
Effective testing depended on real fluency with the underlying documentation, not just the application itself. Knowing where to look for a specific requirement, and noticing when a document had been revised, was a skill in itself given the scale involved:
This engagement placed a single Azati QA engineer directly inside a testing team of roughly 10 to 15 testers plus a test lead, working alongside a development team of around 10 engineers and two further test leads for backend and frontend, plus a separate analyst group. Integrating productively into a structure that size, rather than working as an isolated contributor, was itself part of the value delivered.
Testing a system backed by more than 20 documentation volumes is not primarily a tooling challenge, it is a comprehension challenge. Azati's QA engineer built real working knowledge of the requirements base, including how to track revisions, which mattered as much as any test script for catching defects rooted in a missed documentation update.
Verifying an access model with this many roles, permission combinations, and edge cases requires methodical, unglamorous work that is easy to under-scope. Azati treated the role model as a first-class testing target in its own right, not an afterthought to functional coverage.
Document control systems in heavily regulated industrial environments carry real consequences when something is missed. Azati's QA engineer approached document lifecycle testing with the rigor that a regulated, multi-facility operational context demands, the same discipline behind Azati's other enterprise document and portal work in the energy sector.
Extensive role-based access models, multi-stage document workflows, and dense regulatory requirements all need a tester who can keep up with the documentation, not just click through the UI. Let's talk about your platform.
Stress-test my document workflowAzati tested the full document lifecycle, from creation through every approval stage to archival, validating that each of the roughly ten users involved in a typical flow could perform their assigned actions correctly, in the right sequence, with the right status transitions. Where one document's status affected a linked document, testing covered both sides of that relationship.
With an unusually large number of distinct user roles, Azati systematically verified each role's access level, visibility, and permitted actions against the requirements documentation. This included checking for permission gaps, overlaps, and edge cases where a role's behavior depended on document status or context.
Test cases were built directly from the project's technical and regulatory documentation rather than from the interface alone, with active tracking of documentation updates to keep coverage current. Regression testing protected previously validated functionality as the platform continued to evolve across its full development cycle.
Azati's QA engineer tested the internal AI assistant built to help the project team query the documentation base directly, verifying that responses reflected the actual content and intent of the underlying volumes rather than producing plausible-sounding but inaccurate answers, an important distinction in a documentation set this dense and consequential.
| Area | Azati contribution |
|---|---|
| Document workflow | Tested multi-stage approval flows spanning roughly ten user roles per document lifecycle |
| Access control | Verified an extensive role-based permission model against requirements documentation |
| Test design | Built requirements-based test cases from 20+ volumes of technical documentation |
| Regression coverage | Maintained regression testing across ongoing releases |
| AI tooling | Tested the team's internal documentation assistant for response accuracy |
| Documentation | Produced acceptance test documentation for delivered functionality |
The platform operates in a heavily regulated industrial environment where document control, access traceability, and procedural correctness carry direct operational consequences. Role-based access control governed who could view, edit, or progress each document, with permissions tied closely to documented procedural responsibilities. Testing treated the access model itself as a security-relevant surface, verifying that no role could act outside its documented scope.
Azati provided a single senior QA engineer who worked entirely within the client's own testing organization rather than as part of a separate Azati delivery team. The broader project included a testing team of 10 to 15 testers plus a test lead, a development team of around 10 engineers with two additional test leads for backend and frontend, an analyst group of 4 to 5 people, and project management.
The engagement ran for roughly 6 to 8 months and concluded naturally once the full scope of work was completed. Delivery followed a blended methodology:
Azati's QA engineers are comfortable inside dense regulatory documentation and large access control models, not just clicking through a UI.
Talk to a QA engineerDocument approvals, deadline tracking, and executor assignment across multiple facility types moved from a fragmented, largely paper-based process into one coherent web platform, with testing ensuring that consolidation held up under the platform's full operational complexity.
A question that previously meant a wait of around two hours for an analyst's response could be answered immediately through the project's internal AI assistant, freeing analyst time for higher-value work and keeping the broader team unblocked.
Systematic testing of an unusually large access model gave the client confidence that user permissions matched documented procedural responsibilities precisely, a meaningful assurance in a regulated, multi-facility operational context.
The engagement concluded because the complete scope of testing work was finished, not because of a fixed calendar deadline, a marker of a well-defined engagement executed to completion.
What stands out about this engagement beyond the testing scope itself:
Most QA scopes treat permissions as a secondary check layered onto functional testing. Here, the access model was large and consequential enough to warrant its own dedicated testing discipline, complete with its own slice of the requirements documentation. Recognizing that distinction, rather than folding access checks into generic functional test cases, is what catches the permission gaps that matter.
On a system with 20-plus volumes of source documentation, the tester who can navigate that material directly, and notice when it has changed, finds defects that a tester working purely from the UI never will. That fluency is a deliberate capability, not an incidental byproduct of time on the project.
Feeding a large documentation base into an internal AI assistant turned a recurring two-hour wait into an immediate answer. It is a small-scope build next to the core platform, but it is exactly the kind of targeted AI application that removes a specific, measurable friction point rather than chasing a broader automation narrative.
Enterprise construction cost estimation platform for oil & gas projects with complex cross-module calculation dependencies.
Oilfield reservoir analytics platform with 4 modules in 4 months, computational core rebuilt in 10 days, and ~1000X reduction in numerical precision loss after enterprise QA.
Sharepoint migration and portal consolidation for oil and gas, 30+ portals consolidated, 100,000+ content entities migrated, 15%+ customizations beyond initial estimation.
Optimized Oil&Gas web platform with streamlined workflow and fuel card management.
AI-powered service for automated oil & gas meter data reading and processing.
Data bus integrating SAP ERP with document generation for optimized workflows.
Last updated
