Case study • Laundry Sauce

How Kodif helped Laundry Sauce 11X their AI support performance vs their previous provider.

80%
of tickets handled
77%
resolution rate across all intercepted tickets
11x
more resolved tickets vs. their previous provider

Challenges

Before Kodif, Laundry Sauce worked with another AI CX provider. The goal was to take load off human agents and keep subscribers happy. What they got was harder to measure.

The reported numbers didn’t match the tickets
The previous provider’s dashboard reported that it was “assisting” with 68% of tickets. When the Laundry Sauce team audited the data alongside Kodif and cross-referenced it with Gorgias logs, actual resolution was closer to 7% of total tickets. In high-priority categories the gap was similar: only 11% of subscription cancellation tickets were actually resolved.
Conversations that started but didn’t finish
The AI would offer a resolution or a deflection, such as suggesting a skip instead of a cancel, and then stop if the customer didn’t reply. There were no reminders and no automated next steps. Human agents ended up manually closing tickets the AI had only half-handled. That added to the team’s workload and frustrated customers who expected a finished interaction.

Solution

Phase 1: Full-cycle conversations that end in an action

The Kodif AI Agent was set up to own the entire support cycle: start the conversation, decide what to do, take the action, and confirm it was completed. That covers things like:

Cancelling a subscription
Rescheduling a shipment
Addressing a missed delivery

If the customer doesn’t reply, the agent follows up automatically and takes the next best action.

Phase 2: Automation you can verify

Every decision comes with an audit trail, clear reporting, and live visibility, so the numbers Laundry Sauce sees match what actually happened in Gorgias.

Phase 3: Control at the level of each use case

Instead of an all-or-nothing setup, Laundry Sauce controls exactly which issues the AI handles, how it handles them, and when it escalates. Each scenario is configured separately, with its own messaging, SOPs, tone, and logic:

Subscription save logic is tuned independently from cancellation handling
Lost package responses follow a different tone and escalation path than order status tickets
Shipping delay handling can include proactive refund or gift options without affecting unrelated returns

Phase 4: A hands-on partner

From onboarding through optimization, Kodif worked like an extension of the team. That meant rigorous QA of past agent performance, live syncs to adapt to business changes, and custom paths for subscription saves and deflection.

Results

The AI now handles 80% of Laundry Sauce’s tickets end to end, with a 77% resolution rate across every ticket it intercepts.

80%

of tickets handled by Kodif AI

77%

resolution rate across all intercepted tickets

11x

more resolved tickets vs. previous provider

“Compared to our previous provider, Kodif is easily 11x more helpful in handling our CX workload. We’re automating more, we have better visibility with audit trails, and the team’s been thrilled with the hands-on support.” (Ian Blair, CEO)

“The enterprise-level customizability made a huge difference. We weren’t forced into a one-size-fits-all flow, we got exactly what we needed.” (Ian Blair, CEO)

The takeaway for subscription and DTC brands

Laundry Sauce’s story starts with a measurement problem, not an AI problem. “Assisted” and “resolved” are different numbers, and a dashboard that blurs them hides where customers are still waiting. The audit against Gorgias logs is what showed the real gap.

For CX leaders evaluating AI, two questions matter most. First, does the vendor’s resolution number reconcile with your helpdesk data? Second, what does the AI do when the customer goes quiet halfway through a flow? Follow-through on unanswered conversations is where subscription saves are won or lost.