Stackly achieved a 4.9/5 CSAT score with AI-powered frontline support that customers love
Key metrics
CSAT score
resolution rate
avg. response time
Priya Patel
CX Director, Stackly
Customers don't even realize they're talking to AI. The quality is that good.
Priya Patel
CX Director
Stackly is a developer tools company providing infrastructure monitoring and observability for engineering teams. Their technical customer base expects fast, knowledgeable support that understands complex technical environments.
The challenge
Stackly's developer customers are demanding — they expect fast, technically accurate answers and have little patience for generic responses:
- Technical queries requiring deep product knowledge
- Code-level troubleshooting that generic support bots couldn't handle
- Developer expectations for instant, accurate responses
- Growing user base outpacing their small support team
Previous chatbot solutions had been a disaster — developers could immediately tell they were talking to a bot, and satisfaction scores dropped. Stackly needed AI that could match the quality of their best human agents.
Why Stackly chose Duckie
Duckie's approach to knowledge was the differentiator:
- Deep knowledge base integration — Duckie ingested their technical docs, API references, and changelog
- Context-aware responses — the AI understands the user's environment and product version before responding
- Natural conversation — responses read like they're from a knowledgeable colleague, not a script
- Seamless escalation — when a question requires engineering input, the handoff is smooth and contextual
Stackly ran a blind test comparing Duckie's responses to their top agent's — evaluators couldn't consistently tell the difference.
The results
The numbers speak for themselves:
- 4.9/5 CSAT score — higher than their pre-AI baseline of 4.6/5
- 92% autonomous resolution rate for technical support queries
- Average response time under 2 minutes — down from 45 minutes
- Customer complaints about bot quality: zero since launch
The most telling metric: their NPS for support interactions actually increased after deploying AI — a result that's virtually unheard of in the industry.
Looking ahead
Stackly is now using Duckie's knowledge gap detection to systematically improve their documentation. Every question Duckie can't answer becomes a documentation ticket, creating a virtuous cycle of improving both AI and human support quality.
More Case Studies
Grid automates 91% of member support with Duckie, handling 70,000+ tickets a month without growing the team
Grid is a fintech app helping millions of Americans take control of their money, with cash advances up to $200, early access to tax refunds through PayBoost, credit score building, and the Grid Card. Built for people who live paycheck to paycheck and need money to move when they do, Grid has become a go-to financial lifeline in moments that matter.
CloudCart scaled to 45K monthly tickets while actually improving CSAT scores
CloudCart is an e-commerce infrastructure platform powering thousands of online stores worldwide. Their support team handles everything from merchant onboarding to payment disputes, requiring both speed and accuracy at massive scale.
FinServe Global achieved 100% audit compliance with full AI transparency across every interaction
FinServe Global is a financial services company providing payment processing and banking infrastructure to businesses worldwide. Operating under strict regulatory requirements, their support operations must maintain complete transparency and auditability for every customer interaction.