UAT
Business users describe scenarios in plain language and run acceptance tests without scripting.
QAI · AI-Powered Quality Engineering
QAI plans, writes and runs UAT, functional, end-to-end, smoke and regression tests on Playwright — and keeps the evidence — across Oracle CX Sales, Oracle CPQ and other web applications.
- Delivered for Oracle CX Sales and Oracle CPQ teams
Proven results
- reduction in test design effort
- 70%
- faster UAT execution cycles
- 50%
- increase in automated test coverage
- 3x
Test types
One platform for business users and QA teams. Every test is written as a Gherkin scenario and an executable Playwright test.
-
-
Functional & end-to-end
Full user journeys through your application, planned from requirements and the screens QAI explored.
-
Smoke & sanity
Fast checks that a build is fit to test before the full suite runs.
-
Regression
Suites replayed against an approved baseline, with every difference reported.
-
Migration parity
The same suite run on a legacy and a rebuilt product, compared part by part.
-
API compatibility
Checks that an API change stays backward-compatible with its consumers.
-
Cross-browser (Roadmap)
The same tests across Chromium, Firefox and WebKit.
-
API functional testing (Roadmap)
Executable tests for REST, GraphQL and SOAP endpoints.
Product walkthrough
A simplified run-through of the QAI workflow, step by step.
QAI · UAT cycle
Oracle CX Sales · REQ-OPP-118
Illustrative demo
Sample data · not a client result
QAI turns a plain-language user story into a Gherkin scenario and an executable test, creates happy, edge and negative test data, runs it end to end in Oracle CX Sales with Playwright and links every result to its evidence.
Illustrative QAI UAT walkthrough for Oracle CX Sales. User story REQ-OPP-118: “As a sales rep, I can create an opportunity with a revenue line, so the forecast reflects the deal.” Acceptance criteria: The opportunity is created on the account; Amount rolls up from quantity × price; A negative quantity is rejected. QAI extracts the actor (Sales rep), object (Opportunity) and child object (Revenue line), then writes 2 Gherkin scenarios with 6 steps in opportunity_revenue_line.feature (Create an opportunity with a product revenue line; Negative quantity is rejected) and turns them into an executable test. The test data factory creates 6 synthetic opportunity records on the Demo - Northwind account (2 happy · 2 edge · 2 negative); Oracle must reject the negative records, and any that saves is flagged as a defect. Playwright runs the test end to end in the Oracle CX Sales Redwood UI and passes all 8 steps in 5.6 seconds (illustrative timings): Sign in as sales rep; Open account “Demo - Northwind”; Create opportunity; Add revenue line 20 × 150.00; Expect amount = 3,000.00; Enter quantity −5; Expect the line is rejected; Clean up test records. The result links the requirement REQ-OPP-118 (opportunity with a revenue line), test cases TC-1181 · TC-1182 (happy path · negative), run RUN-0457 (8 / 8 steps passed), evidence 8 screenshots · video (ready for sign-off).
Describe
A user story and its acceptance criteria, in plain language.
Generate
QAI writes a Gherkin scenario, then turns it into an executable test.
Provision data
The test data factory creates happy, edge and negative records.
Execute
Playwright runs every step end to end in Oracle CX Sales.
Trace
Results link the requirement, test cases, run and evidence.
How it works
- 01 QAI Explorer
Explore
Walks your application in a real browser, signed-in areas included, and learns its screens, forms and fields.
- 02 QAI Planner
Plan
Reads requirements, Jira stories and documents, and plans suites and cases, each tagged with its requirement.
- 03 QAI Author
Write
Writes each case as a Gherkin scenario and an executable Playwright test, with happy, edge and negative data.
- 04 You
Approve
A person approves every plan and every generated test, and can edit or regenerate it, before anything runs.
- 05 Playwright
Run
Runs every step end to end, with a live view, a screenshot per step and a video of the session.
- 06 QAI Healer
Heal & trace
Repairs a broken step only if the re-run passes, and links every result to its requirement and evidence.
Capabilities
AI-powered quality engineering for Oracle CX Sales, Oracle CPQ and other web applications — QAI writes UAT, functional, end-to-end and regression tests, runs them on Playwright and keeps the evidence.
-
01 Automated Test Generation
Generate UAT, functional and end-to-end tests from requirements, user stories, business documents or plain language — as Gherkin scenarios and executable Playwright tests.
-
02 Intelligent Test Data Factory
Generate realistic, reusable test data across happy-path, edge and negative cases, with rules that keep test records safe on shared environments.
-
03 Requirement Traceability
Link every requirement to its test cases, runs and evidence, and see which requirements have no test yet.
-
04 End-to-End Execution
Run tests end to end on Playwright in Oracle CX Sales, Oracle CPQ and other web applications, with a screenshot of every step and a video of every run.
-
05 Oracle CPQ Product Learning
Learn a configurable CPQ product from its own screens — every attribute, option, nested option and rule — into a product map a person approves before tests are built from it.
-
06 Self-Healing Tests
When a screen changes and a step breaks, QAI proposes a fix and keeps it only after the test passes again.
-
07 Baseline & Parity Regression
Replay suites against an approved baseline, or on a legacy and a rebuilt product side by side, and report every difference.
-
08 Signed Evidence & Audit
Every run produces a verifiable evidence package, backed by a tamper-evident audit trail.
Use cases
-
UAT for an Oracle CX Sales release
Business users need to sign off a release, but every UAT cycle waits on QA to script the tests.
- Describe
- Generate
- Provision data
- Run
- Report
Business users run acceptance tests from their own requirements, and negative tests flag a defect when Oracle accepts a bad value.
-
Regression-check an Oracle CPQ migration
A configurable product is being rebuilt, and every option, rule and price must still come out the same.
- Learn both products
- Build the suite
- Replay old and new
- Compare
The same suite runs on the legacy and the rebuilt product, and every difference in parts and totals is reported.
-
Functional coverage for a web application
Requirements live in Jira and documents, and nobody knows which ones have a test.
- Explore
- Plan
- Write
- Run
- Trace
Smoke and end-to-end suites run with self-healing, and uncovered requirements export to CSV.
Bring one user story or requirement from Oracle CX Sales, Oracle CPQ or another web application. We'll turn it into an executable test in front of you.
AI agents
Each agent does one job. A person approves every plan and every generated test before anything runs.
-
QAI Explorer
Maps your application
Learns your application's screens, forms and fields. For Oracle CPQ it learns every attribute, option and rule of a configurable product.
-
QAI Planner
Plans the coverage
Plans suites and test cases from your requirements, user stories and documents, each tagged with the requirement it covers.
-
QAI Author
Writes the tests
Turns each case into a Gherkin scenario and an executable Playwright test, with happy-path, edge and negative test data.
-
QAI Healer
Keeps tests green
When a screen changes and a step breaks, proposes a fix and keeps it only after the test passes again.
Security & governance
-
A person approves every test
Plans and generated tests wait at a review gate. Failures QAI can't repair go to a human triage queue.
-
Your own private model
Assign a model to each AI role — cloud or your own private endpoint — and restrict high-security workspaces to private models only.
-
Evidence you can verify
Every run produces a signed evidence package, backed by a hash-chained, tamper-evident audit trail.
-
Credentials kept out of tests
Credentials are injected at run time from a vault and redacted from evidence; single sign-on and role-based access control who does what.
Integrations
- Applications under test
- Oracle CX Sales Oracle CPQ Any web application
- Execution
- Playwright Chromium
- Requirements & source
- Jira PDF Word Excel GitHub GitLab Bitbucket
- AI models — cloud or private
- OpenAI Claude Gemini Azure AI Amazon Bedrock Vertex AI Your private model
- Platform
- Docker Kubernetes PostgreSQL Single sign-on (OIDC)
On the roadmap
Part of the QAI architecture and in development.
-
Oracle Fusion Suite Coverage (Roadmap)
End-to-end execution across further Oracle Fusion modules beyond CX Sales, such as ERP, HCM and SCM.
-
Defect & Release Traceability (Roadmap)
Link test results to defects and releases in your issue tracker and ALM tools.
-
Scheduled Regression (Roadmap)
Run regression suites on a schedule or from your CI/CD pipeline.
-
Compliance-Grade Test Data (Roadmap)
Masked test data derived from production patterns, aligned to your compliance policies.
Also see CortexIQ, InsightMesh and AIM
Get started with QAI
Book a demo and we'll take one of your requirements from plain language to an executed test, with the evidence your release board needs.