insightful.cx

Opinionated qualitative research analysis for senior UX researchers — evidence-first, language-disciplined, defensible.

We turn 24 interviews and 22 research questions into a director-grade deck and report — without ever skipping a stage gate, collapsing the dual-axis scoring, or hiding the disconfirming voices. The tool proposes; the researcher disposes.

Members cohortRecurringMost participantsconfidence: high

Members consistently flag the renewal rate jump as the trigger to shop around — even when the cover itself still reads as competitive.

Prevalence
4/5
Intensity
5/5

Supporting evidence

1
  1. I will not stay if the next renewal is anywhere near this jump. The cover is fine — the surprise is what is killing the relationship.
    P14· Yvonne

Counter-evidence

1
  1. Honestly the rate change did not bother me — the cover is what kept us. They were transparent about why it was going up.
    P07· Sam
per_rq_synthesisclaude-opus-4-6tpl rq-syn@2026-04-3018,4201,372 tokMay 3, 2026, 07:42 PM

How the methodology works

The pipeline is fixed. Each stage produces a versioned artefact; each artefact requires explicit researcher review and lock before the next stage runs. There is no “skip” button and no auto-publish path.

Seven-stage qualitative analysis pipelineIngestion, speaker attribution, per-transcript coding, cross-transcript patterns, per-RQ synthesis, narrative drafting, reel curation. Each stage produces a versioned artefact and is gated by reviewer lock.1IngestionTranscript & metadata2AttributionSpeaker bindings3CodingCodes & quotes4PatternsThemes5SynthesisFindings per RQ6Narrative.docx & .pptx7Reel.mp4
Seven stages · six locked gates · one artefact per stage

Full methodology →

What it is

Stage-gated, human-in-the-loop pipeline

Ingestion → speaker attribution → coding → patterns → per-RQ synthesis → narrative → reel curation. Every gate produces a versioned artefact and requires explicit researcher review before the next stage runs.

Tier and dual-axis scoring, never collapsed

Every finding carries a tier (dominant, recurring, signal, outlier) plus independent prevalence (1–5) and intensity (1–5) scores. We never roll these into one number, because one number lies.

Full back-traceability from claim to source

Every claim in every output links to a participant ID, a verbatim quote, a transcript offset, and (where applicable) a video timestamp. If we cannot trace it, we do not say it.

What it produces

A .docx report, a .pptx deck, and an .mp4 highlight reel. Each one is a rendering of the underlying findings, not the work itself. Re-render is one click.

What it isn't

Frequently asked

Is this for me?

It's for senior UX researchers running director-grade engagements where defensibility is the first requirement. If you are looking for a flexible tag-and-quote tool, this is not it.

How is the methodology enforced?

The pipeline runs as seven stages with mandatory researcher review at every gate. The tool refuses to ship outputs that violate its principles — it will not auto-skip a gate or collapse the dual-axis scoring.

Where does the data go?

All structured data is in Supabase Postgres in Sydney; transcripts are sent to Anthropic's Claude API for analysis; nothing else leaves the stack. See /privacy.

Can I share findings with clients?

For now, you export the .docx, .pptx, and .mp4 and share them directly. Read-only client share links are on the roadmap.

How do I get access?

Invitation only. Write to jas@ortomate.ai with a one-paragraph note about a recent project.

Access is invitation only

We are working with a small group of senior UX researchers running real client engagements. If that's you, write to jas@ortomate.ai with a one-paragraph note about a recent project — we'll respond.

No form, no plans, no pricing, no waitlist. Read the principles first — they are the strongest filter we have.