DoesQA AI
Optional AI across DoesQA: summaries, Automatic Elements, Suggestions, the Assistant, Learning, and links into MCP and the CLI.
DoesQA AI provides optional efficiency across the product: faster failure triage, quicker Element setup, clearer next steps for coverage, in-product answers, and agent-driven authoring when you want it.
DoesQA was built to be easy without AI, so every Flow, Element, and Run still works with no AI features at all. DoesQA AI defaults to On for new accounts. You can turn it off any time. See Control.
A shorter overview also lives under Platform → DoesQA AI.
Your data stays with DoesQA
Models and data used for DoesQA AI stay inside DoesQA.
Nothing leaves our ecosystem for DoesQA AI.
DoesQA AI is hosted in our UK datacentre.
How Learning works
Learning is optional. When enabled, DoesQA AI gets sharper for you and for everyone on your account, so summaries, Suggestions, and the Assistant stay closer to how your team already tests.
When something useful applies more widely, other customers only benefit from generic learning, not from your private account detail.
You
Notes and context that stay with how you work
Your account
Shared across everyone on that account
Generic
Broad improvements that do not expose your account’s private detail
You can opt out of Learning under Control.
Agent memory on MCP and the CLI is a different store. Agents save and recall durable facts with explicit scopes. That is not the same as Automation Intelligence Learning. See Memory for agents.
Layers
Each layer unlocks a different kind of speed. Use one, several, or all of them.
Test and Run summaries
When a Test Case or Run fails, DoesQA can show an AI Summary on the result: a short title and a clear write-up of what went wrong and what to try next. Open Assistant from the summary to continue in chat with that failure already in context. Triage moves from scanning every step response and screenshot to reading the summary, then acting, so failure review stays short as suites grow.
Automatic element selection
When you create an Element, set Element configuration to Automatic and write a plain-language Description, such as Yellow login button in the top right. DoesQA generates the Selector from that Description. You can switch to Manual later to edit the Selector yourself.
Describe the control the way a tester sees it, and DoesQA turns that into a Selector. First Flows land sooner, and when the page shifts you can refresh from the Description instead of rebuilding the Selector from scratch.
Your first Flow already uses this path. See Create and run your first Flow and Creating Elements.
Suggestions
On the home page, DoesQA can show Suggestions: short ideas for what to improve next, such as coverage to add or Selectors to harden. Open Assistant from a suggestion, or use Expand coverage, to turn the idea into a guided chat. Suggestions pull practical work from how your account already tests, so your next move is already on the table and expanding coverage takes less planning time.
Assistant
The Assistant is DoesQA’s in-app AI for questions about your Flows, Runs, and coverage. Open it from the main menu, or from a Flow, Test Case, or Run so the chat already knows what you are looking at. From home you can start topics such as expand coverage, get help, or ask anything.
The Assistant works with your account’s real context: the Flow or Run you opened from, how your team already tests, and the failures in front of you. Ask it to explain a failure, harden Selectors, or expand coverage, and it answers inside DoesQA with that context already loaded. While a reply is generating, use Stop in the composer to end that turn and keep control of the conversation.
An AI coding agent connected over MCP or the CLI can ask the same Assistant a question. The exchange shows up as an ordinary conversation you can read in the app. On that path the Assistant answers; it does not author the Test for the agent.
MCP and CLI
The DoesQA MCP and CLI connect an AI coding agent to your account. Each person creates an access token under Settings → User. That token is how the agent signs in as them. One connection exposes the platform as tools the agent can call: Flows and Test Steps, Elements, Run Recipes, Runs and Results, artefacts and traces, coverage, failure patterns, Suggestions, knowledge search, the Assistant, and durable memory.
The agent’s output is product data. You open a review link into the Flow, the Run, or the Assistant conversation and read the journey and the evidence from a real browser.
The Flow Builder stays the primary way people author Tests. MCP and the CLI are how an agent joins in.
How does DoesQA avoid AI hallucinations or unreliable results?
DoesQA was built so testing works without AI. Pass and fail for almost every Test Step come from real browsers, stored Selectors, and tried and tested Steps, not from generative guessing. DoesQA AI makes triage, setup, and authoring faster. It does not replace that execution model.
DoesQA AI is mature, stable, and trustworthy. It is tried and tested across millions of Runs on a wide range of platforms and systems. You can turn Automation Intelligence off any time and keep using Flows, Manual Elements, Runs, and Results as usual. See Control.
Runs stay deterministic
When a Test Case runs, DoesQA executes the Steps you configured. Automatic waiting, fixed Selectors, and shared Step behaviour are the same platform story as Maintenance and reliability. Generative AI is not what decides those outcomes.
The one intentional exception is AI Vision. That Step uses AI to judge a written visual expectation, so it can affect pass or fail when you add it. Use it when a plain-language visual Check is the right tool, and keep other Checks on ordinary Steps when you want a fully non-AI path.
AI that speeds you up
Summaries are correct and can be trusted for triage. They are built to explain the failure in front of you so you can act faster.
DoesQA generates a Selector from your Description as a starting point. The stored Selector is what Runs use. You stay in the loop when Selectors change. See Creating Elements and Maintenance and reliability.
Suggestions and Assistant
Optional helpers grounded in your account and the Flow or Run you opened from. They speed planning and questions; they do not rewrite Run pass/fail.
So: AI is optional acceleration on top of a deterministic product. Turn it off when you want. Keep AI Vision as an explicit Step when a visual expectation belongs in the Flow.
Control
DoesQA AI is powerful but fully optional. All of DoesQA keeps working with DoesQA AI off: Flows, Manual Elements, Runs, and results stay fully usable.
Automation Intelligence is the account switch for summaries, Suggestions, and the Assistant. It defaults to On for new accounts.
To change it:
Open Settings.
Find Automation Intelligence.
Set Enable Automation Intelligence to On or Off.
Automation Intelligence Learning is also optional. Leave it on so Intelligence learns from your account and stays closer to how you test.
Related
Last updated