UAT Studioby Salvion
AI test design

AI that does the tedious part of a UAT

Writing test scripts, chasing what an issue really means, summarizing fifty issues before the go/no-go: AI does the first draft in minutes. You check it and decide. Nothing changes without you.

Made for business testers, not ITYou approve every proposalNot used to train the AI
Try it in the free pilotSee what it costs
Before the test day

From user stories and processes to test scripts

The longest part of a UAT is writing the scripts. Read your user stories from Jira or Azure DevOps, or paste them, and get test cases written in the language of the business, with steps, test data and the expected result.

User storiesTest cases from your storiesOne or more proposals per story, including the unhappy paths: a wrong value, a missing field, a user without the right role.
  • From Jira, Azure DevOps, Excel or pasted text
  • Steps it had to assume are marked to check
  • Each test case stays linked to its story
ProcessesEnd-to-end tests from a processUpload a process diagram or work instruction and get the main path from start to end, plus the decisions and hand-overs between roles.
  • BPMN (Camunda, Signavio, Bizagi), Visio, Word or PDF
  • Lanes become the roles that test
  • Every decision gets its own test case
CoverageSee what is not covered yetWhich stories have no test case, and which issues belong to which story. Ask for proposals for just the stories that are missing.
  • Coverage per story, in the go/no-go report too
  • Issues go back to the story
During and after the test days

Issues the developer understands, and an overview for the go/no-go

Testers write what they see, in their own words. The AI turns an issue into steps to reproduce, the expected and the actual result, and says what is still missing, such as which record or which user role. Before the go/no-go meeting it summarizes all open issues: by theme, what blocks go-live, and which issues look like the same problem.

  • Write up with AI: use the proposal or discard it; the tester's original stays in the history
  • Summarize with AI: read it, click through to the issues, copy it for the minutes
  • Possible duplicates are suggested while a tester logs an issue, in every plan
Release readiness

Which test cases a CRM release touches

Salesforce has three releases a year, and Microsoft rolls out Dynamics 365 updates all year. Salvion keeps the changes of each Salesforce release and the items of the Microsoft roadmap for Dynamics 365, and checks each one. You tick once what your project uses; AI compares the changes with your test cases.

Review the releaseThree lists before the release reaches productionClick Check with AI on the projects page, for a Salesforce release or for a Dynamics 365 month or quarter.
  • Your test cases the release touches, each with the reason
  • Proposals for new test cases for what changes and is not covered yet
  • The required updates for your admin, with the date they are enforced
  • Paste the release notes of a package or your own release too
The roundA regression round with exactly those test casesThe Must test cases, the touched test cases you keep and the new ones you approve, copied with their steps, test data and expected results.
  • A reminder when the round is not signed off 3 days before production
  • A fail says which change of the release it came with, also in Jira, Azure DevOps and the API
  • The release report in the go/no-go report and the evidence pack

Only the changes of the release and the titles of your test cases with the action of each step go to the AI: no expected results, test data, results, issues or names of people. Release notes you paste are sent with their name, and nothing else. With your own AI model connected, the check uses that model. Release readiness is an add-on of €99 per month for every plan: add it under Plan & billing.

You stay in charge

A proposal, never a decision

1. You askOnly a test lead starts the AI, for the stories, the process or the issues they choose.
2. AI proposesTest cases, a write-up or an overview, marked as a proposal.
3. You decideEdit, approve or discard. Nothing lands in your project without you.
Not used for trainingWhat you send is not used to train the AI.
Only what is neededStories, steps and issue texts; no names of testers.
Copy and paste, if you preferReady prompts for the Copilot or ChatGPT of your company, for test cases, testers and issues, in every plan; then UAT Studio sends nothing to an AI.
Or your own AI modelConnect the model of your company: Microsoft Foundry / Azure OpenAI (also for Copilot users), OpenAI, Anthropic, Amazon Bedrock or Google. The texts then stay within your own contract.
Clear limitsA monthly allowance of user stories and fair use, so there are no surprises on the invoice.

AI test design uses Anthropic as a sub-processor, unless your workspace uses its own AI model; see For procurement and IT for where the data goes. Leave personal data of customers out of stories and issues.

What it costs

An add-on to any plan

PlanPriceIncluded
PilotFree, included as a trial10 user stories and 5 requests on issues
Project€199 once500 user stories for the run of the project
Team, Business, Partner, Enterprise€99 per month300 user stories a month
Need more?€49 once250 extra user stories
Release readiness, every plan€99 per monthThe review of each CRM release

A process description counts as one user story per 3,000 characters. AI on the issues and Release readiness use no user stories. Prices excluding VAT.

See it on your own storiesThe free pilot of 14 days includes an AI trial of 10 user stories.Start pilot