Browse categories
English
$
USD
Connect to your AI
I will test your ai chatbot and ai agent for bugs and broken workflows
QA Engineer, AI Chatbot Testing, API and Automation
I help teams find release-blocking bugs in web apps, SaaS products, APIs and AI-powered workflows. With 5+ years in QA, I offer functional, exploratory, regression and API testing, plus automation wit...
About this Gig
Launching an AI chatbot or agent? I will test real user conversations, find broken workflows and deliver a bug report your developer can act on.
Coverage: happy paths, invalid inputs, fallback loops, multi-turn context and instruction consistency. Standard and Premium also cover tool/API actions and workflow error handling. Answers are checked against your supplied reference material.
Your report includes scenario IDs, test inputs, pass/fail results, expected vs actual behavior, reproduction steps, severity, screenshots or transcripts, and a coverage summary.
Basic: 10 scenarios. Standard: 25. Premium: 50, plus a regression checklist and one scoped retest. One scenario is one objective, up to 5 user turns, run once. Each package covers one chatbot or agent.
Message me with your demo and main user journey; I will confirm the right scope before you order. Provide expected behavior, reference answers and test access. Functional QA only; agent development and security testing require a separate scope.
Testing application:
Software
Device:
PC
My Portfolio
FAQ
What counts as a test scenario?
One scenario has one defined objective and expected behavior, up to 5 user turns, run once. Each package covers one chatbot or agent. Repeated runs for response variability, additional agents and longer workflows require an agreed custom scope.
How do you check AI answer accuracy?
I compare outputs with your reference answers, knowledge sources and acceptance criteria, and flag unsupported or inconsistent responses. Ambiguous results are marked for review. I do not claim every AI answer can be proven correct.
What do I receive and what does a revision cover?
A coverage summary, reproducible defect report with severity, and screenshots or transcripts. A revision clarifies the report within the original scope. Premium includes one retest of reported issues; other retests can be scoped separately.
Do you build agents or perform security testing?
This gig tests existing chatbots and agents. Agent development, automated evaluation pipelines, penetration testing and voice/telephony testing are separate scopes. Please message me before ordering if those are your goals.
