ChimpAgent - Your Exploratory Testing Assistant

🤖 AI browsing agents bring exciting possibilities - especially for QA. But, as with everything, success depends on "how" we use the tech... 🔍

Some argue for replacing plain-old scripts with AI test bots. There are major downsides to that:

🚨 Too slow for CI/CD – AI bots can’t match the speed of Playwright scripts, making them impractical for pipelines running thousands of tests.

🚨 Expensive – Bots consume tokens & CPU cycles, making large-scale testing a costly endeavour. 

🚨 Non-deterministic, prone to hallucination - imagine if the last gatekeep of your production release is a bot that hallucinated and gave the thumbs up!  “The AI said it was OK” isn’t going to be a good excuse when you’re explaining what went wrong.

We took a different approach and built ChimpAgent - an exploratory testing companion that outputs plain-old test scripts, instead of running agentic tests every time. 🚀

🛠 How it works:

1️⃣ Describe a scenario in plain English or upload a Gherkin file.

2️⃣ ChimpAgent executes the scenario, like a human, in the browser. 

3️⃣ Our SDKs capture the agentic session and generates a complete test script (in playwright), with detailed assertions included.

💡 We maintain Gherkin versions of tests alongside the script, which enables our agent to brainstorm related scenarios, and come up with accurate steps for them (which you can refine).

👉 The agent then executes the new scenarios (just like in step 2 above).

🔁 Rinse and repeat. You get a thorough test script suite that is fast & cheap to execute, and is deterministic - unlike an agentic tester.

Here is a quick preview of our ChimpAgent in action:

#AIinTesting #ExploratoryTesting #QAEngineering #TestAutomation #BrowserUse #AIAgents #QualityAssurance #Playwright #Cypress