MeUndies Hits 80% Containment and Increases Support Capacity by 3.7x During Busy Season

MeUndies Needed Scalable AI That Could Keep Up With Business

MeUndies tried AI for customer support before, but the tool they started with wasn't intelligent enough to handle the complexity of a membership business or flexible enough to keep up with constant changes to shipping, pricing, and policy. The team came away from that experience with a clear sense of what they actually needed from an AI solution for chat and email.

"We had tried AI before and it just wasn't built for the complexity of what we do. We needed something intelligent enough to handle a membership business and flexible enough to keep up when things changed. That bar was a lot higher than most vendors could meet." — Michelle Rascon, Senior Manager of Customer Experience, MeUndies

Staffing added another layer of complexity. MeUndies operates in a business with strong seasonality, which historically meant recurring cycles of hiring, onboarding, and offboarding 30+ agents. That process was time-intensive, and every new cohort of agents introduced risk to CSAT during their ramp period. The team needed a way to handle volume fluctuations without rebuilding headcount every few months.

That need showed up in unexpected places as well. When MeUndies relaunched its 3XL and 4XL sizing, the number of customer contacts spiked twice. First from customers asking why the sizes were unavailable and again once the products went live. Moments like that required surge capacity the team couldn't spin up overnight. Forethought gave them a way to absorb those spikes through automation and AI agents addressing high volume questions rather than hiring while keeping response quality consistent when member expectations were highest. 

At the same time, membership support has a rhythm that doesn't accommodate normal business hours. Membership orders are often placed overnight when no agents are available. The help center was there, but it wasn't doing enough on its own to resolve those requests without human involvement. With a medium-size team and peak season on the horizon, the gap between what the team could handle and what members expected was only going to grow.

MeUndies Built an AI Support Layer with Forethought 

MeUndies needed an AI solution that was intelligent enough to handle the complexity of a membership business, flexible enough to adapt to constant change, and capable of taking real action on behalf of customers without routing them to an agent. When they found Forethought, they started building AI Agent using Forethought Solve for chat and email, then added increasing personalization using custom APIs built in Action Builder.

Forethought Handles Chat and Email Volume Across Every Shift

MeUndies deployed AI agents across chat and email, the two channels where their members are most active. Forethought's omnichannel AI agents are built to understand customer intent, incorporate answers from a knowledge base, amongst many other information sources, and resolve requests fully without routing to a human – creating a personalized workflow for customers using API. 

Chat is MeUndies’ primary channel, and handles the bulk of incoming volume there. MeUndies also uses email. Because their email experience presents to customers as an interactive widget rather than a standard inbox exchange, the flows built for chat translated directly, with only minor adjustments needed to get email up and running.

What has made Forethought particularly effective for MeUndies is how quickly the team can adapt it when things change. Things like updates to the help center or changes in policies can be adjusted the same day. 

"We set out with a goal that everyone thought was too aggressive. We hit it by October, and then December came and we would have been completely overwhelmed without it. Forethought is the reason we made it through peak season." — Michelle Rascon, Senior Manager of Customer Experience, MeUndies

That same flexibility extends to how they manage deflection. The team set up three deflection paths, high, medium, and low, adjusting them based on incoming volume. When agents have capacity, more conversations are routed to humans. When volume spikes, AI agents take on more. Together, these controls have made it easier to optimize the customer experience without over- or understaffing.

MeUndies Builds Its Most Effective Customer Workflows in Forethought Action Builder

The MeUndies team has seen some of their most meaningful results with Forethought workflows. Forethought's Autoflows power what those flows can do, connecting to backend systems like Shopify and NetSuite to take real action on a member's behalf rather than simply providing information. For a membership business where so many interactions require actually doing something, that capability changed what support could look like.

Order Cancellation AI Agents

The cancel order flow is one of the clearest examples. Members often realize they need to cancel an order after the fact, sometimes when no agent is available to help. MeUndies built a flow that checks how much time has passed since the order was placed. If it falls within the first window, Forethought cancels it directly in Shopify. If more time has passed, it checks the third-party logistics provider and NetSuite to see whether the order can still be intercepted before it is physically picked and packed. Members can request a cancellation right up until that point, which is a significantly wider window than most brands offer.

"The cancel order flow is one I'm really proud of. A member can reach out at 3am and get it resolved instantly. That's the kind of experience that keeps people coming back." — Michelle Rascon, Senior Manager of Customer Experience, MeUndies

Subscription Retention AI Agents

Another example is their subscription management flow. When a member reaches out about their subscription, whether they want to change their order quantity, skip a month, or cancel altogether, Forethought walks them through an interactive experience that is more effective than the standard onsite option. The flow makes additional offers, and if a member is considering canceling, it can offer a discount or the option to skip instead. Today they see 25 to 30% retention through that flow. 

Forethought Gave MeUndies AI That Answers Questions and Takes Action

When MeUndies set out with a goal of 80% containment, the team thought it was aggressive, but by October, they had hit it. That number held through the holiday season, when shipping complications, carrier issues, and membership changes created the kind of volume that would have overwhelmed a team their size without it. The results across chat, email, and membership workflows speak for themselves:

• 80% containment rate on the chat widget
• 3.7x increase in support capacity; 11 agents handled Q4 2025 volume that previously required 41
• 25% member retention through the subscription management flow
• 113% increase in percentage of conversations originating in chat

For a membership business, every interaction is a moment where a member either stays or leaves. With Forethought, containment kept the team standing during their hardest month of the year, the subscription flow turned potential cancellations into retained members, and the flexibility to throttle deflection based on volume gave the team a lever that headcount couldn’t.

MeUndies Plans to Expand AI Agents Beyond Support

With containment where they want it, the team's focus is shifting toward what Forethought can do beyond support. Proactive AI agents are the next area they are looking at, where the same intelligence that resolves support interactions could be applied to how MeUndies communicates with members more broadly. 

They are also watching Forethought's product recommendation feature closely. For a brand where members choose their monthly items and exclusive prints drive loyalty, a recommendation layer that connects to their Shopify catalog equals an influence on what members discover and purchase. That kind of personalization, applied at scale, is a natural extension of what the support team has already built.

‍