How Outlier Group Lifted Trustpilot From 3.5 to 4.0 With an AI Support Agent

Outlier Group manufactures and distributes direct-to-consumer cosmetics and perfume products in the UK. One of its ecommerce websites, The Scent Reserve, operates at high volume: roughly 1,200 to 1,500 orders per day.

Customer service sat at the center of that operation. The team was small, knowledgeable, and trusted with more than basic ticket replies. They understood the products, the compliance requirements, and the operational edge cases behind each customer conversation.

But as order volume grew, support volume grew with it.

“We found the challenge of volume of tickets increasing to a point where we can't add as much value as we want to the conversation.”

“Currently it's around about 90% of the inquiries are tracking related.”

Most customer inquiries came from the same place: customers asking where their order was after it left the warehouse and entered the courier journey.

The information already existed in courier systems, but customers still wanted a clear answer from support. That created a queue full of repetitive tracking tickets, even though the team wanted to spend more time on product range, compliance, regulatory questions, ingredients, and senior-level investigation.

Outlier Group tried the obvious fixes first.

They increased staffing levels. They used email templates. They looked at the AI tools offered by their support provider.

None solved the problem cleanly. Staffing was hard to predict because demand moved with order volume. Templates still required a human to read, click, select, and send. The built-in AI option would have required too much implementation work from a small team that already had to protect live service levels.

“There's no way it would have been in place, even answering one single ticket, in the length of time your solution came in for 90% of tickets.”

“The AI agent obviously works 24/7. It copes with demand very, very easily.”

Outlier Group implemented an Adelante AI agent inside its existing support operation.

The agent connected the context that mattered for ecommerce support: policies, warehouse management systems, ticket handling, courier logic, order tracking, product information, and escalation rules.

The rollout was controlled from the start. The team began with around 10% of tickets, monitored the replies, checked the information, and verified the responses. Then they moved to 20%, 30%, and kept increasing volume as confidence grew.

That mattered because the team had two clear concerns before launch: data safety and customer perception.

Their customer base skews slightly older, and James, Head of Group Operations at Outlier Group, worried that customers might react badly to an AI agent replacing a UK-based support agent. He was especially concerned about Trustpilot.

“We were concerned that the AI may have a negative impact on particularly things like our Trustpilot score.”

“I think currently it's 80 plus percent of all tickets are solved end to end by the AI agent.”

The AI agent now resolves more than 80% of tickets end to end.

That result gave Outlier Group more than operational efficiency. It created a visible customer experience win.

Since launching Adelante, Outlier Group's Trustpilot rating for The Scent Reserve climbed from 3.5 to 4.0. James tied that lift directly to the AI support experience because customers were specifically mentioning the AI agent, the responses it provided, and the actions it took.

“The amount of comments specifically around the AI agent and the responses provided in the actions it's taken has been fantastic.”

The public reviews back that up without needing a wall of screenshots. One recent Trustpilot review names Stacey, the AI support agent, and describes a support issue resolved quickly with zero stress.

The Trustpilot lift also created a commercial effect. As the score started climbing again, Outlier Group saw sales increase.

“We've even seen an increase in sales because our Trustpilot score has started to climb again.”

“This is a tool to help you with your workload. It's not designed to replace.”

Outlier Group positioned the AI agent as a workload tool, not a replacement plan.

The support team could see the impact in Trustpilot, ticket volume, and orders going out the door. That helped the rollout land positively internally.

Instead of reducing the team, Outlier Group kept the knowledge inside the business and gave the team more time to work on higher-value customer problems.

“What we're actually doing is keeping and retaining that talent.”

As the business grows, the team does not need to add headcount in the same linear way.

“Don't be concerned about the time it takes, the consumer reception, or your response from the team.”

Outlier Group did not use AI to make customer service feel smaller. It used AI to make customer service faster, more consistent, and more useful.

For a high-volume cosmetics and fragrance ecommerce website, the result was clear: 80%+ of tickets resolved end to end, faster answers for customers, more time for the support team, and a Trustpilot score that climbed from 3.5 to 4.0.

Customers did not punish the brand for using AI. They mentioned the agent positively in reviews. That makes this case study stronger than a standard automation story.