How Dossier Scaled International Growth with AI Agents and Zero Black Friday Panic

Dossier is the perfume house for the next generation that's been scaling fast both in the US and internationally. With a CX team led by Oana Sale, who's been with the company for 6 years, they've gone from a domestic operation to an international brand in a relatively short time. That kind of growth is exciting, but it also means the support volume grows exponentially across multiple markets, time zones, and customer expectations.

The challenge

When you're growing fast across multiple markets and channels, the support volume doesn't just increase—it explodes. Dossier's team was covering everything: email, TikTok, Trustpilot, social media, chargebacks, QA, and negative reviews. Every review on every platform gets a response. That's a lot of ground to cover.

The real problem? When Oana's team slept, no one was replying to customers. And customers really appreciate fast responses.

"We noticed that customers really appreciate when you reply to them very fast. And even if we did that during the night, when my team sleeps, of course no one replied."

— Oana Sale, CX Lead, Dossier

They'd tried other AI tools before Siena. They worked okay at first, but as Dossier grew, they quickly outgrew those tools. "We had another AI tool. It was okay for us since we were at the beginning. But then, as we noticed with other tools, since we grew so much and so quickly, these tools were not adequate for us anymore. Basically, we outgrew them."

They needed something that could actually keep up and improve as fast as the AI world moves. In AI, everything changes from one month to the next. You need a partner that can stand out and stay competitive.

Why Siena

Oana's criteria was simple: she wanted an AI-native platform, not a ticketing system that also does AI on the side. She needed a team that was dedicated to improving AI every single day, not treating it as a side feature.

"We needed someone competitive that was really dedicated all the time for improving this tool. And since in the AI world, everything happened so fast and everything changes from one month to the other, you need someone that could really stand out. And we noticed that Siena could do so much compared to the other two, and that is why we chose you."

Her goal from the beginning was clear:

"If you want to try an AI, just go with one that is dedicated to that. Not one that's like a ticketing platform that also does AI or some other app that also does AI. My goal was from the very beginning to have an AI that is dedicated to that. So you make sure that every day, hours at work are dedicated to improving the AI app." — Oana

After testing Siena against other options, the difference was clear. Siena could do much more.

Implementation

Dossier kicked off their Siena implementation in September 2024. They started with Siena's certification program, which gave the team a strong foundation. "We did that AI course that you sent to us, and it was really helpful. And after that, we had the integration part, the implementation part, and it seemed already familiar to us because we did that certification program."

Oana put 3 people through the program initially. Her thinking: "I was thinking, okay, if someone is not present or goes on vacation, we need someone else to be there." She recommends at least 2 people, ideally 3 at the beginning. "The more the merrier" for the implementation phase, though they realized later they didn't need as many people actively involved once things were running.

The key recommendation? Get people from customer care involved. "At least someone from customer care. It's my team, but it always seems the team that knows best about customers. Their feedback, how they feel, what they want, what their pain points are."

Because Dossier had already used another AI tool before, the team was comfortable with the concept and the platform. The process was smooth overall—regular calls, sharing topics and use cases. "Everything went smooth. Sometimes it's not that smooth, but we made it happen. The team from Siena was really helpful and always on top of things."

Results

The team started seeing results fast. Really fast.

43% automation rate across all conversations

Within the first week or two of going live, the signs were clear.

"We instantly noticed, while we added more procedures to be automated, we noticed how much time my team had for themselves to improve and work on other tasks. And we already noticed that since I think one week, two." — Oana

Black Friday without the backlog

Usually peak season means struggling to keep first response times decent. Teams scramble, hire temporary help, pay for overtime, and still end up with a backlog that carries into January.

Not this time.

"I think the most important part was noticed not only by me, but by everyone. Even my manager was during Black Friday, how the first response time really, so it didn't go to the roof. And that's, I think, the most important thing, because while other companies during Black Friday struggle, we also struggled when we didn't have AI, struggled to meet a decent first response time. We were very happy to see that the first response time was the same as the one in the normal days."

Dossier doesn't hire additional help during peak anymore. They pay their team extra to do extra hours, but Siena handled the volume surge.

"Siena was a big, big help. Because if not, I don't know, maybe we would be in a backlog even today." — Oana

That's right—without Siena, they might still be clearing the Black Friday backlog in January.

Team capacity unlocked for high-value work

When AI handles the repetitive, automatable work, your team gets time back for tasks that actually need humans. Dossier's team now focuses on:

• QA work to maintain quality standards

• Handling chargebacks

• Managing negative reviews (Siena currently covers the positive reviews, team handles negative ones)

• Growing professionally with reporting and presentations

• Social media management

• TikTok (which is its own platform entirely)

• Tasks that require human judgment and empathy

"It's important for them to have time and do tasks that do not require that we need real humans to do them. Because if an AI app can be taught to do things that can help us and automate them, it's so much better so that my team can focus on improving, I don't know, growing even professionally and knowing more about reporting, presentations, etc."

The real ROI

Cost savings are real—no more paying for extra help during Black Friday and peak season. But Oana sees the ROI differently.

"People can, and not only the members of my team, the ones in my team, but also the team as a whole, can be happy with and focus on other things. And I think that's the most important thing."

Team happiness. Professional growth. Time to focus on work that matters. That's ROI you can't always capture in a spreadsheet, but it shows up in retention, quality, and morale.

Beyond support automation

Dossier started with the support agent as their foundation in September. Once they had the core email procedures in place and automated, they expanded to other agents.

Reviews agent: game changer for a small team

Dossier responds to every single customer review. On the website. On TikTok. On Trustpilot. Every platform, every review. They used to do it all manually.

"We want to reply to each and every single customer that leaves us a review. And this takes a lot of time. And we used to do it manually, basically."

Once they launched Siena Reviews Agent, everything changed for their small reviews team.

"All of a sudden, the team doing reviews, which is a very small team, could focus again on other stuff and improving the way they require another platform, etc. I think that it was a game changer in terms of reviews." — Oana

Currently, Siena covers the positive reviews while the team handles negative ones manually. That division of labor makes sense—let AI handle the high-volume thank-yous and positive feedback, let humans handle the sensitive stuff.

Shopping agent: like chatting with a concierge

The shopping agent just launched, so Dossier is at the beginning of tracking metrics like conversion rates. But the quality of the experience is immediately clear.

"We noticed that it's so good. The replies are really good. And since the shopping agent is like a chat that people can come and ask all sorts of things—not only returns, exchanges, but also recommendations and advice—we noticed that it was like really talking almost to a human, basically in chat."

It acts like a concierge. Customers can ask for help with returns and exchanges, but also get product recommendations and advice. "It's like chatting with someone real on the other side."

Do they disclose it's AI? "If someone asks, yeah. Why not?" The quality of the conversation speaks for itself, even when customers know they're talking to a chatbot.

The future of CX teams

Oana sees AI changing roles across every department, but especially in customer care since they're using it the most.

"AI will change people's roles in every department, I think, but mostly in the customer care department. At least for us, since we were the ones using it the most. People can improve in other tasks. Can improve themselves. Can focus on other important requests that they receive. And we receive so many since we are a continuously growing company."

Even hiring is starting to change as the role of CX evolves from reactive ticket-solving to more strategic, high-value work.

Advice for other CX leaders

"If you want to try an AI, just go with one that is dedicated to this. Not one that's like a ticketing platform that also does AI or some other app that also does AI. No. My goal was from the very beginning to have an AI that is dedicated to that. So you make sure that every day, hours at work are dedicated to improving the AI app." — Oana

Find a platform that's AI-native, dedicated to your vertical (like ecommerce), working on the same problems every day. That's how you make sure the tool keeps improving at the speed you need.

Get your customer care team involved early. They know your customers best. And don't be afraid to put multiple people through training initially—you can always scale back once you're running smoothly.

Most importantly, don't wait until you're drowning to implement AI. Dossier made the move as they were scaling internationally, not after they were already in crisis mode. That timing made all the difference.