Challenge
Kameleo’s engineers were its support team, working across a split system that leaked time.
Support ran on a Zendesk inbox with documentation scattered across Zendesk and GitHub, all maintained by hand. Every release meant rewriting docs, and because updates lagged, documentation quietly drifted out of sync with the product. The same technical questions about Selenium and Playwright integration came in repeatedly, and answering them pulled the people building the product away from building it, with customers stuck waiting for business hours.
Nothing in the stack learned. A hard-won answer an engineer wrote at midnight helped exactly one customer and then disappeared into a closed ticket, so the team solved the same problems again and again.
“Our support tickets are basically code reviews, answering them well needs an engineer, so support was literally stealing hours from the roadmap. And nothing we used remembered anything. We’d solve the same Playwright edge case five times and start from scratch every time.”
Kameleo needed a platform that could actually resolve deep technical questions, keep documentation honest, and remember how problems were solved.
Solution
Kameleo replaced its Zendesk + GitHub setup with Helply’s complete platform, built around the Support Agent.
The Support Agent resolved genuinely technical tickets end to end, investigating across documentation, past resolutions, and logs to reason about root cause and return precise answers on Selenium, Puppeteer, and Playwright workflows rather than pointing at a doc. Connected to Linear and Slack, it filed clean engineering issues and alerted the right person when something truly needed a human. The Assist Agent rode alongside the engineers, investigating in the background and recommending actions as approve-or-edit cards so responses went out faster.
Managers built Skills in plain language, describing a recurring integration scenario and the steps to handle it, no code required, and living documentation mapped where content was missing, revealing the knowledge base was only about 70% complete and drafting the articles to close the highest-impact gaps. Every engineer edit became company memory, so a fix written once sharpened every future investigation.
“It doesn’t deflect technical questions, it resolves them, and it finally remembers how. We describe a workflow in plain English and it becomes a Skill. Living docs showed us exactly where our documentation was lying to customers, and fixed it.”
Kameleo joined the first cohort migrating fully off Zendesk to Helply’s platform.
Results
Kameleo automated the majority of its technical queue, kept its docs honest, and got its engineering hours back.
The Support Agent resolved 79% of inbound questions end to end, and because customers no longer waited for business hours, responses ran roughly 3x faster. Engagement doubled as customers asked more, knowing they’d get an instant, accurate technical answer, and more conversations flowed through Helply than had ever come through the old Zendesk queue.
Platform highlights
- Fragmented Zendesk + GitHub setup replaced with one platform
- 79% of technical questions resolved end-to-end with root-cause investigation
- 3x faster responses; customer engagement doubled
- 30+ plain-language Skills built for recurring integration scenarios
- Knowledge base mapped from ~70% complete and closed by living documentation
With the repetitive technical load handled and the learning loop keeping documentation current, Kameleo’s engineers went back to shipping product. The support system became a forcing function that got sharper with every resolved ticket.
“Helply forced us to run support properly. It resolves the deep technical tickets, keeps our docs honest, and remembers every fix, so the whole thing compounds. That’s the difference between support being a drag and being an asset.”
More customer stories
Marketing Software
How Sender.net turned support conversations into pipeline with Helply
~4,000 monthly conversations handled, 71% resolved end to end
Research / Healthtech
How Covidence automated researcher support without sacrificing accuracy with Helply
Zero drop in accuracy while resolving 64% of support end to end (72% at peak)
Sales Software
How Proposify unified two products on one support platform with Helply
~220 fewer tickets a month reaching humans, with 52% resolved end to end
Get the latest from Helply.
Ideas on AI agents, customer support, and building AI-native customer experiences.