Poplin Automates More Than 55% of Customer Support, Its AI Vendor Says

Short answer: A Poplin co-founder publicly says he used AI and automation across the company, including customer service, as part of work that cut overall operating costs in half. That's backed by a detailed case study directly from the vendor involved, Champ AI, which says it now automates more than 55% of Poplin's support volume, including late-delivery workflows that Champ says can be diagnosed and resolved without human review or customer outreach when the delay is determined to be customer-driven. The same case study quotes Poplin co-founder Nachshon Fertel saying most customers don't even realize they're talking to an AI. In practice, our own testing found support began through an automated in-app chatbot before human escalation. Multiple customers describe that human layer as outsourced, though we couldn't independently confirm Poplin's staffing arrangements. Poplin doesn't publish a general customer-support phone number; general requests go through its Help Center and ticketing system, though it does publish a phone line specifically for U.S. customers age 65 and older.

You don't usually get to see a company's cost-cutting strategy in its own words. In this case, you do, because the person who ran it posted about it.

What Poplin's own co-founder says he did

Nachshon Fertel co-founded what became SudShare in 2017 alongside his father, fellow co-founder Mort Fertel, later rebranded as Poplin. His publicly listed role history at the company runs: Technical Co-Founder/CTO (2017-2022), Head of Product (2022-2023), Head of AI Operations (2023-2026), and Board Member/Co-Founder (2026-present).

In his own words, describing the Head of AI Operations role on his public professional profile, his priority was efficiency and unit economics at "the largest laundry service in the U.S." He says he rebuilt internal operations from the ground up, integrating AI across customer service and automating workflows, as head of the company's largest department with 60 direct reports. He credits that broader AI and automation work, spanning customer service and other internal operations, with cutting the company's overall operating costs by 50%. His profile doesn't say customer service specifically accounted for that full reduction, so we're attributing the cost claim to the company's operations as a whole, with customer service as one piece he explicitly named.

He's since stepped back to a board seat and founded a new company, Day54, built explicitly around selling that same playbook to other businesses: finding manual workflows, automating them into a client's existing tools, and cutting operating costs the way he says he did at Poplin. In a recent public post promoting that new venture, he describes spending years using AI and automation at Poplin to cut operating costs in half, and says he now helps other companies do the same.

This isn't a leaked memo or a disgruntled-employee account. It's one of the accomplishments the founder highlights publicly about his own work, and it provides unusually direct context for the automated support experience customers describe.

A named vendor confirms the automation directly

Separate from Fertel's own account, there's a full, on-the-record case study from the vendor itself, published directly on Champ AI's website, not just referenced secondhand. It names Poplin as the customer, credits Nachshon Fertel by name and title, and lays out the arrangement in detail: Champ says customer support is one of Poplin's biggest expenses, and frames the program's goal as increasing support capacity without increasing cost alongside it. Champ AI now automates more than 55% of Poplin's support volume, up from roughly 20% with a previous AI vendor Poplin had tried and dropped. According to Champ, Poplin would like to push automation to 70-80% going forward.

This isn't a simple FAQ chatbot. According to the case study, Champ's system can pull a customer's order record, review prior messages between the customer and their Laundry Pro, check support-ticket history on both sides, communicate directly with either party, log internal notes, and carry a multi-step case through to resolution without a person handling it. One example the case study walks through: when an order runs late, the system checks delivery and payment status and reviews prior context. If it determines the delay was "clearly customer-driven," it records its own reasoning and closes the matter without ever reaching out to the customer. In other cases, it messages both the customer and the Laundry Pro directly. That's a system determining whether a delay was customer-driven and, in some cases, resolving the alert without human review or customer outreach.

The most notable line in the case study, in our view, is a direct quote from Fertel: "Most of our customers don't even realize it is an AI." Champ says Poplin tracked CSAT during the rollout and found the AI's performance aligned with its human agents. Between Fertel's own account of AI-driven cost cutting and this case study's operational detail, Poplin's support system runs substantially on automation, on the record, from the company's own vendor and co-founder, not just as something inferred from customer complaints.

What this looks like when you actually need help

Poplin doesn't publish a general customer-support phone number for most users. Its Contact page directs general support requests to its Help Center and ticketing system, while separately publishing a phone line for U.S. customers age 65 and older. In our own testing, the in-app support flow began with an automated chatbot, and multiple customer accounts describe a similar sequence:

1. An in-app AI chatbot handled the first response in our testing and in the customer accounts we've reviewed. We haven't systematically tested every category of issue, so we can't say it covers "nearly everything," but it was the first point of contact in every instance we've directly seen.
2. If the bot can't resolve it, it may connect you to a live agent. Multiple customer reports describe that layer as outsourced rather than Poplin's own in-house staff, though we couldn't independently confirm who actually employs those agents.
3. If the live agent can't resolve it either, the issue may get escalated to email. Customer reports also describe that layer as handled by outsourced support staff, though the same caveat applies: we don't know the employer behind it.

We found no description in Poplin's public materials of any of these layers as staffed by Poplin's own in-house employees. That's a statement about what we could find, not proof that no in-house staff are involved.

Why this matters for a claim, not just a chat

We've written before about Poplin's Protection Plan and its "sole discretion" language for approving or rejecting damage and loss claims. Because Fertel describes AI integration as part of a broader efficiency effort, an automated first-line support structure could add layers between a customer and whoever has authority to approve an exception or nonstandard payout. We couldn't find a publicly documented escalation path that guarantees access to an in-house Poplin employee with authority to make a final call on a claim. That doesn't mean one never gets involved behind the scenes, only that we couldn't verify it.

One case we're tracking involved more than 30 separate support contacts over six days before reaching a resolution. The records don't establish why so many contacts were necessary. Champ AI's own case study says its system is built to retrieve prior order and ticket history rather than start fresh each time, so a simple "no context between contacts" explanation is less likely than we originally suggested. What actually caused this specific case to take 30 contacts isn't something we can determine from the outside.

What this means if you're dealing with Poplin support

• Don't expect the first automated response to be the final word. If it doesn't resolve your issue, ask explicitly to escalate to a live agent, and again to a supervisor if needed.
• Keep your own written record. Screenshot every chat, and save every email ticket number. Keeping your own record means you're not dependent on each new support contact having access to, or actually reviewing, the full history of your case.
• Don't assume persistence alone is broken if it takes many contacts. It may reflect the structure of the support system itself, not a mistake on your part.

Have your own Poplin support experience to share, good or bad? Submit your story.

Sources: Champ AI case study, "Complex Workflows, Fully Automated: How Poplin Reached 55%+ Support Automation" (champ.ai/case-studies/poplin); public LinkedIn profile and posts of Poplin co-founder Nachshon Fertel (Head of AI Operations, 2023-2026; Board Member, 2026-present); Poplin.co Contact page; PoplinReviews' own testing of Poplin's in-app support flow; customer reports collected via BBB, Trustpilot, and SmartCustomer.

---

FAQ

Is Poplin's customer service outsourced? Multiple customer reports describe Poplin's live agent and email support as outsourced rather than staffed in-house. Poplin's own co-founder has publicly described integrating AI across customer service as part of a broader cost-cutting effort. We couldn't independently confirm who actually employs Poplin's human support staff.

Does Poplin use an AI chatbot for support? In our own testing, yes: the in-app support experience began with an automated chatbot before any live agent became available. Other customer accounts describe the same sequence.

Is there a phone number for Poplin customer support? Poplin doesn't publish a general customer-support phone number. General requests go through its Help Center and ticketing system. It does publish a phone line specifically for U.S. customers age 65 and older.

Why does resolving a Poplin issue sometimes take dozens of contacts? We don't know for certain in any specific case. Champ AI's own case study says its automation is built to retrieve prior order and ticket history rather than start fresh each time, so a simple "no context between contacts" explanation is less likely than it might seem. What actually causes cases like this to take dozens of contacts isn't something we can determine from the outside.

Does Poplin use a third-party company for AI support automation? Yes. A detailed case study published directly by Champ AI, an AI support-automation vendor, names Poplin as a customer and says Champ automates more than 55% of Poplin's support volume, up from about 20% with a prior vendor. According to Champ, Poplin would like to push automation to 70-80% going forward. The case study says the system can pull order records, review message and ticket history, and resolve or close cases without a person handling them.

Do customers always realize when Poplin support is AI? Apparently not. Champ AI's case study quotes Poplin co-founder Nachshon Fertel saying, "Most of our customers don't even realize it is an AI."