PromenautUse case · Business analysis
Business Analyst Copilot
Business analysts spend most of their time reading: stakeholder interviews, process documents, tickets and email threads that rarely agree. The judgement is theirs, but the sheer volume caps how much they can cover and how fast. This puts an analysis workflow behind the analyst so the reading, cross-referencing and first-draft artefacts scale, while the calls stay human.
From raw input to signed-off output.
$ promenaut run business-analyst
The business-analyst workflow, end to end
Gather & structure
Ingest stakeholder interviews, existing process documents, tickets and free text, deduplicate across sources, and structure actors, systems, processes and rules into the knowledge graph.
Analyse & challenge
Cross-reference domain knowledge to surface gaps, ambiguities and contradictions between stakeholders, and map dependencies across the process so nothing is analysed in isolation.
Draft the artefacts
Generate the process maps, a structured business requirements document and a prioritised backlog, each traceable back to the source that justifies it.
Review & sign off
The analyst reviews the draft, resolves the flagged questions with stakeholders, and signs off. Every change is versioned, so a re-run shows exactly what moved.
How it actually works.
- Every finding is cited back to a source input, so the analysis can be traced rather than trusted.
- Contradictions are found by comparing stakeholders against each other, not by reading each document in isolation.
- The knowledge graph carries domain context across runs, so later analysis starts from what the last one learned.
- Decision-support by design: the workflow drafts and flags, but the analyst owns every call.
- Runs are versioned with delta tracking, so re-analysis after new input shows only what changed.
What it produced.
The business-analyst workflow, end to end
Whole corpus
Read and cross-referenced, not a sample
Gaps + conflicts
Surfaced with the source that justifies each
BRD + backlog
Drafted and traceable to source
Human sign-off
On every conclusion
A first draft the analyst edits, not an answer to accept. The workflow scales the reading and cross-referencing; the judgement stays with the analyst.
The parts that break naive tools.
Stakeholders disagree without knowing it
Two teams describe the same process differently. Comparing sources against each other, rather than reading each alone, surfaces the conflict before it reaches the backlog.
The context lives in people’s heads
Undocumented rules and assumptions are captured into the knowledge graph as the work runs, so the next analysis does not start from zero.
Volume caps coverage
An analyst can only read so much. Scaling the reading lets the whole corpus be covered, not a sample, while the calls stay human.
Analysis has to be defensible
Every gap, requirement and priority is traceable to the source that justifies it, so conclusions can be reviewed rather than taken on trust.
See it on a workflow you own.
Bring a real process. We'll show what your first agentic workflow looks like on the platform.
