OluKai scales support with Sierra's AI: 70% auto-resolution, 4.5/5 CSAT
Footwear brand OluKai reports that its "Day Makers" support team now resolves about 70% of customer inquiries autonomously using Sierra's AI-driven platform. Despite the automation, they're maintaining a reported 4.5/5 customer satisfaction score.
The AI handles structured tasks like return policy exceptions and educating customers on Happy Returns. Human agents stay focused on higher-complexity cases, where context, empathy, and judgment drive outcomes.
What's actually being automated
- Return policy exceptions (clear rules, edge-case handling, and approvals where applicable)
- Happy Returns education (guiding customers through the process and setting expectations)
- Similar routine inquiries that follow policy and can be scripted without losing quality
Why this matters for support leaders
- High containment without a major drop in CSAT suggests customers accept automation when it's fast and accurate.
- Agents get time back for escalations, VIPs, and nuanced situations that impact loyalty.
- Omnichannel coverage helps keep responses consistent across chat, email, and social.
- Lower cost-to-serve and tighter SLAs without adding headcount.
How to apply this playbook
- Map top intents by volume and value. Start with return flows and policy-driven questions.
- Codify exception logic. Define thresholds for approvals, credits, and replacements.
- Integrate returns tooling and content. If you use Happy Returns, sync policies and statuses so answers are live and accurate. Happy Returns
- Set routing rules. Escalate on sentiment spikes, account flags, or unresolved multi-turn threads.
- Measure fast and often: containment rate, CSAT gap vs. human, handle time, and recontact rate.
- Create a feedback loop. Review transcripts weekly, tag failure modes, and update policies or prompts.
- Train agents on "AI + human" workflows so handoffs feel seamless to the customer.
KPIs to watch
- AI containment rate: share of tickets fully resolved by the AI agent
- CSAT delta: compare AI-handled vs. human-handled interactions
- Average handle time and first contact resolution
- Deflection to self-service and recontact within 7 days
- Escalation accuracy: are the right cases reaching humans at the right time?
Risks and how to mitigate
- Policy drift: lock policies to a single source of truth and version changes.
- Edge cases: define hard stops where AI must hand off; audit exceptions weekly.
- Tone control: set style guides and test across channels to avoid robotic or overly casual replies.
- Compliance and privacy: restrict data access and log decisions for review.
What this signals for ops and investors
Automating a large share of support without degrading satisfaction points to real efficiency gains. If similar outcomes repeat across more brands, it supports recurring revenue growth, higher retention, and a stronger position among AI-powered customer experience platforms.
For support teams, the takeaway is simple: focus AI on clear, policy-driven work, protect the edge cases with smart routing, and constantly tune the system based on actual conversations. That's how you keep quality high while scaling volume.
Helpful resources
- CSAT essentials and benchmarks: What is CSAT?
- Upskill your team on AI for customer support: Complete AI Training - courses by job
Your membership also unlocks:
