PromenautRequirements Intelligence Report
The combined, stakeholder-ready report of the run: readiness, gaps, conflicts, quality, dependencies, what was generated, risks and next steps, produced as HTML and PDF automatically.
Requirements reach engineers before anyone has checked them
A requirements set arrives as a spreadsheet or a stack of tickets. Gaps, contradictions and duplicates surface during the build, when they cost most to fix, and nobody can say whether the set was ready to start.
Every requirement parsed, scored and cross-checked, with its stories and tests written and one readiness verdict, before anyone builds.
Turns a set of requirements into analysed, structured and testable work: every requirement parsed and scored, gaps, dependencies and duplicates found, stories, acceptance criteria and tests written, and a readiness verdict before anyone builds.
Goes in
A requirements set as a spreadsheet, tickets or free text, with the intake details
Comes out
A readiness verdict, a report, and stories, acceptance criteria and tests for each requirement
Who decides
Your people. Work stays private until its author promotes it, and publishing is a choice.
To start
A set of requirements you already have, and the objective, problem, markets and sponsor behind them.
01What you get
A report people use, and the work engineers build from.
User stories, acceptance criteria and tests
Written for each requirement, ready for the engineers and the developer agents who build from them.
Traceability and readiness
A matrix from each requirement to its stories and tests, and an assessment of whether the set is ready.
A quality score and verdict for each requirement
One score per requirement with a verdict of ready or needs revision, and a scoring deck and spreadsheet that show them.
02The team you get
Specialist agents, each with one job, working as one team.
Every agent this module brings, by the team it works in. Teams marked side by side work at the same time; the rest work in turn.
Intake
Capture the submission from the intake form, and record the assumptions and dependencies the submitter gave.
Intake form capturerCodeAssumptions and dependencies recorderCode
Analysis
Parse each requirement into a common structure, and score its quality.
ParserModelQuality scorerModel
Cross-referenceSide by side
Find gaps against what the organisation already knows, map the dependencies between requirements, and catch duplicates and contradictions.
Gap analyserModelDependency mapperModelDuplicate detectorModel
Structure
Merge overlapping requirements into one, keeping their sources, and split an over-broad requirement when asked to.
Structure checkerCode
Impact
Identify the applications, journeys and process steps each requirement touches, and flag significant new functionality.
Impacted applications finderModel
Quality
Combine each requirement's scores into one quality score, with a verdict of ready or needs revision.
Composite quality scorerModel and code
GenerationSide by side
Write user stories with story points, test scenarios and acceptance criteria for each requirement.
Story generatorModelTest generatorModelAcceptance criteria writerModel
Validation
Trace each requirement to its stories and tests, and assess whether the set is ready to build.
Traceability checkerModelReadiness assessorModel
Report
Compose the findings into the Requirements Intelligence Report.
Report composerModel
Scoring deck
Compose a scoring deck, one slide per requirement, and a spreadsheet of the scores.
Score deck composerCode
Backlog generation
A requirements set grouped into capabilities, broken into features, epics and user stories with acceptance criteria, and the whole tree prioritised.
Epic generatorFeature generatorCapability generatorBacklog prioritiser
Continuous improvement
Feedback signals from earlier work prioritised into an improvement plan, and the plan written up for the people who act on it.
Improvement narrative writerImprovement prioritiser
Demand intake
A raw business demand turned into a structured problem statement, classified and scoped, with its stakeholders, impacted applications, risks, assumptions, issues, dependencies, actions, questions and decisions, in one intake summary ready for a business case.
RAID capturerDemand classifierIntake summariserStakeholder identifierProblem statement writer
Traceability
The dependencies between requirements mapped, each requirement traced to the related requirements around it, and every change followed through the dependency graph to whatever it would affect.
Requirement tracerChange propagation assessor
Gap and impact
The current and target state of a process read from the process map and the estate, the gaps found, the applications and journeys the change touches, and the difference between the two states computed exactly.
Target state analyserCurrent state analyserSystem impact assessorProcess impact assessor
Readiness for release
Operational readiness judged from open blockers, automation and cycle-time signals, then business readiness from sign-off, value and training, into one release verdict.
Business readiness assessorOperational readiness assessor
Evaluating one initiative
The submission captured with its business context, the initiative's context assessed, each requirement parsed and structured, acceptance criteria written, impacted applications found and quality scored, in one evaluation dossier.
Initiative context assessorQuality gate scorerDossier composer
Analysis and readiness without the generation
Requirements parsed, classified and cross-checked, stakeholders and impacted applications found, and the set assessed for readiness, in one report.
Requirement classifier
Analysis with the organisation's objectives in view
Requirements parsed, scored and cross-checked, ranked by how well they serve the objectives with OKR Intelligence, then stories, tests and criteria written and traced.
OKR prioritiser
A quick triage
Requirements split into functional and non-functional parts and classified, checked for ambiguity and contradiction, clarifying questions written for the business, and a go or no-go verdict.
Functional and non-functional splitterAmbiguity and conflict detectorClarification question generatorReadiness scorer
Test generation
Acceptance criteria written, tests and end-to-end scenarios generated, product verification tests drafted against the organisation's testing principles, and the coverage of every acceptance criterion checked exactly.
Product verification test drafterTest scenario generatorAcceptance validator
ImprovingSide by side
Propose improvements to a requirement for its writing quality, its uniqueness and its alignment to objectives.
Writing quality proposerUniqueness proposerAlignment to objectives proposer
In conversation
People can also ask questions about requirements, or have a set evaluated, in conversation.
Requirements adviser
03The team at work
Choose a job and watch the team do it.
Each requirement captured with its business context, parsed, cross-checked, judged for impact and quality, and made buildable, with one report and a scoring deck.
People decide
Your people. Work stays private until its author promotes it, and publishing is a choice.
04Results
What it produced.
Measured on a tier 1 bank implementation programme
98% faster
Two people and more than 30 hours of manual review, done in 28 minutes.
Hidden conflict caught
A production-blocking contradiction across documents that the human review had missed.
Readiness call
A verdict on whether the set is ready to build, with reasons a person can act on.
Traced end to end
Stories, tests and acceptance criteria traced back to their requirement, with orphans flagged.
People sign off. The team finds and drafts; the people who own the requirements decide what goes forward.
05People stay in charge
At every point where the work leaves one person's hands.
Submit
The submission is written into the person's own requirements workspace, which colleagues cannot see.
Review the report
The finished report card shows the verdict and what stands in the way.
Promote
Promote commits the submission as a canonical version, still inside that private workspace.
Publish
Publish makes it available in a shared library, only when the person chooses.
Refused until scored, and while it needs rework; a library that requires approval holds it for sign-off
06What it keeps
A requirements submission, held as one unit.
A submission and everything the team found about it travel as one unit: shared together or not at all, and only into a library that has said it accepts that kind of submission.
ObjectiveProblem statementImpacted scopeAlignment to objectivesSuccess measureRequirementsAcceptance criteriaAssumptionsDependencies
07What it does not do
Designed limits.
- Approve or reject requirements. It flags what is missing or under-specified; the analysts and product owners who review requirements decide.
- Check requirements against systems outside the supplied material and the organisation's knowledge graph. It assesses them as written.
- Hold the analysis for approval part way through. Oversight applies to what is done with the output: promoting and publishing it.
Put this team on your own work.
A set of requirements you already have, and the objective, problem, markets and sponsor behind them.
