2.3 CertiMind™ Grounded AI Test Studio

How to automatically generate grounded, robust test cases by scanning live web URLs or uploading PRDs and user stories.

Not enabled by default. Server-side CertiMind generation (URL/PRD scan, Plan Mode, Suite generation, the explorer) is off unless your organisation has it enabled. The recommended way to author tests today is your own coding agent (Claude Code, Cursor) connected to the CertiQA MCP server: it reads your code, creates the cases, runs only those cases, and repairs them from the failure report. Use the author_certiqa_tests MCP prompt. Contact your account team to enable CertiMind for your organisation.

Turning a URL or a PRD Into a First Draft

Hand-authoring dozens of test variations for a single checkout flow or onboarding wizard is hours of repetitive, error-prone work even on a no-code canvas – someone still has to walk the whole page, note every field and button, and decide what “correct” looks like for each one. CertiMind automates that first pass. Point it at a live URL and certiqa-lite, CertiQA’s fast DOM-scanning service, extracts a clean snapshot of the page’s interactive elements; point it at a PRD or a pasted user story instead, and the same generation pipeline reasons from the text directly. Either way, the LLM Gateway brainstorms realistic test scenarios for what it finds, then generates a full step sequence for each one.

Locators You Don’t Have to Fix Later

CertiMind doesn’t hand back a single brittle selector per element. Every generated step carries a ranked list of locator candidates in the same TESTID > ROLE > TEXT > CSS > XPATH priority order CertiHeal’s self-healing cascade already understands, plus a HIGH/MEDIUM/LOW confidence rating – so a generated step that resolved unambiguously reads differently in review from one the model had to guess at.

Why Generation Produces a Draft, Not a Live Case

On a 20-step user story, scenario brainstorming and step generation run concurrently across the backend’s Virtual Thread pool rather than one call at a time – but speed isn’t the reason nothing lands directly in your suite. Every CertiMind job, whether it started from a URL scan or a PRD upload, produces an AI job and a set of draft cases and steps – the same review model every other AI-authored path in CertiQA uses – so a person always reviews and promotes what the model proposed before it becomes a real, executable case.

Generating and Reviewing a Draft

• 1. Open “CertiMind Studio” from the left sidebar and choose a source: a target URL (e.g. https://store.acme-corp.internal/checkout) to scan the live page, or a pasted user story / uploaded PRD to generate from text.
• 2. Click “Generate with CertiMind” – brainstorming and step generation run in the background, typically returning well inside the time it would take to author the same coverage by hand.
• 3. Review each proposed case in the Copilot drawer, paying closest attention to any step whose confidence rating came back MEDIUM or LOW.
• 4. Click Save to promote the reviewed draft into a real case in your suite, carrying the same locator candidates and confidence data forward for CertiHeal and later audits to use.

Getting a More Complete First Draft

• A URL scan only sees what’s on the page when certiqa-lite loads it – for a flow behind a login or spread across a multi-step wizard, a user story or PRD upload that describes the whole flow in text often produces a more complete draft than the scan alone.
• A low-confidence step isn’t wrong by default – it’s the one most worth a second look, usually because the page had more than one plausible match for whatever the model was asked to find.