Developer stress-tests 30 self-built support chatbots with 1,800 chats
11/20/2026 — 11/20, 04:18·1 sources (all)·1 reports (all)
Key takeaways
- A developer built 30 fictional small-business support chatbots on the same model, spanning 10 industries and 4 guardrail levels, then ran th…
Background & analysis
On 2026-11-20, DEV Community · AI published a developer's write-up titled "30 Chatbots, 1,800 Conversations: Why They Failed", describing a stress test of self-built customer-service bots. The developer used the same model to build 30 fictional small-business chatbots, spread across 10 industries, with 4 levels of safeguards. Testing was organized into 12 tiers and covered 1,800 conversations in total. All of the tested bots were self-built rather than real company support systems, and the stated goal was to identify the most common failure modes of no-code and low-code chatbots.
The article opens by noting that most guides on LLM red teaming are aimed at teams that ship Python, recommending tools such as promptfoo, garak, PyRIT or Giskard. Those tools work by requiring users to write test configurations or code, wire them into CI, and keep them maintained. The author argues that a different group builds many of the chatbots that real customers interact with every day, including a freelancer using Make to deliver a booking bot to a dental clinic, an agency using n8n to ship a support agent to an online shop, and a salon owner who built an assistant in Voiceflow. These builders do not fit that Python-oriented workflow, which is why the author turned to no-code and low-code approaches for the bots and the stress test described in the post.
The article is now public. No further developments or follow-up reports on the project have appeared, so the effort currently rests at the stage of this single published summary.
AI-generated from 1 reports · updated 2 days ago
- Heat index
- 25
- All sources
- 1
- All reports
- 1
- First seen
- 2 days ago
Reports on this story headlines open the original
A developer built 30 fictional small-business chatbots across 10 trades and stress-tested them through 12 test levels and 1,800 conversations. Using the same model for every bot, with only the instructions changed, the experiment looks at how low-code and no-code assistants fail. The bots were purpose-built for the test, not real company deployments.
DEV Community · AIAI score 53
Other stories people are talking about
- Heat index 1143RisingAnthropic Reaches $1.6 Trillion Market Value in First Trading Week After IPO21 sources
- Heat index 1011Anthropic Mathematician Says Claude Formalized the Generalized Riemann Hypothesis21 sources
- Heat index 622Anthropic IPO total raise reaches $93.5 billion after greenshoe exercise20 sources
- Heat index 596RisingThinking Machines Lab Open-Sources Inkling 2 Mini with 32GB Model Card Recommendation10 sources
- Heat index 490RisingSEC probes Nvidia accounting practices as shares post seven-week slide9 sources
How is heat calculated?About the methodHide
Heat counts how many independent sources covered a story in the last 48 hours: one source counts once no matter how many posts it published, decaying with a 24-hour half-life. What ranks first is what many people are talking about.
This page aggregates public feeds. Headlines and summaries are machine-organized and remain the property of the original authors; verify important facts at the source. If you believe a headline or summary infringes your rights, email the contact address in the footer with the page URL and basis of your claim, and we will remove or replace it after review.
- Surge
- Discussion rising fast
- New
- First report within 6 hours
- Rising
- Still gathering discussion

