Why we raised $17 million to build the Agent of Record for consumer brands

We raised a $17 million Series A led by York IE, with participation from Joyance Ventures and our existing investors Aglaé Ventures, Sierra Ventures and Super Angel Fund. We're using it to build the Agent of Record for consumer brands: a platform where agents and the people running the brand work from the same understanding of the customer, and add to it with every interaction.

We started with an empathic customer service agent. The job was simple to describe and hard to do well: understand what a customer needs and do the work to help them. Today we work with hundreds of brands, including SPANX, FIGS, HexClad and Comfrt, and have automated hundreds of millions of customer interactions.

Somewhere along the way, our customers started asking for things that had nothing to do with support tickets. Could Siena plug into their social media tool? Could it help with QA? Could it tell them why returns were up in one category? They were using Siena for support, but they could tell it understood their business, and they wanted that understanding in more places.

We also noticed something about the tools they were asking us to integrate with. Most of them were built for a team manually working a queue. So we asked ourselves a question: why integrate with a tool designed around manual work, when we could build the agent-first version of it on top of everything Siena already knows about the customer? That question led to our Social Agent. Then the Shopping Agent. Then voice.

Here's what became obvious over hundreds of deployments. Someone tells a support agent why a product didn't work for them. That should matter when a shopping agent makes the next recommendation. It should also reach the product team if other customers are saying the same thing. Too often, the conversation gets resolved and the understanding stays locked in the system where it happened.

Most software touching customers was built around the team managing the interaction. One tool for support, one for social, one for marketing, each holding its own slice of the customer. The customer is not particularly interested in your org chart.

So we're rebuilding the customer-facing commerce stack around agents that carry that understanding through the entire relationship.

Every brand needs an Agent of Record

A consumer brand already runs on systems of record. The ERP or OMS is the record of orders, invoices and returns. Shopify is the record of the storefront. There's a system of record for loyalty points, another for subscriptions, another for reviews. Each one is good at what it does, and each one records a transaction.

None of them records the relationship. What the customer said. Why she bought the sweater, why she returned the last one, what she asked about in a DM last Tuesday, what the brand did about it and whether it helped. Today that lives in support transcripts nobody reads twice, in social comments that get archived, in a note someone meant to write down.

That's the gap Siena fills. We tap into the systems of record a brand already has and combine them with what customers are actually saying. Siena Intelligence brings conversations together with orders, subscriptions, reviews, loyalty data and the brand's own knowledge, and turns it into shared customer context. As agents serve customers, they add what was asked, what they did and whether it worked. Siena Memory carries that forward, so a customer can move from support to shopping to social to a phone call without having to introduce herself again.

Take a return. An agent learns that a customer liked the product but the fit was wrong. The next time she's shopping, the recommendation accounts for that. Across thousands of customers, the same feedback shows a sizing issue the product team should look at before the next production run. One interaction improves how the brand serves a person and what it knows about its business.

That understanding isn't locked inside Siena, either. Your team can ask Siena directly, in our workspace or in Slack, and get an answer grounded in real customer conversations. Through MCP, other tools and agents can draw on the same context. The customer context belongs to the brand, and it should be useful wherever the brand needs it.

If you're a consumer brand that wants to run on agents, you have two options today. You can stitch together a handful of tools and build your own scaffolding to keep context moving between them. Or you can work with a platform that does the heavy lifting behind the scenes, so every new agent you launch starts with everything the others have learned. That's the Agent of Record.

What brands are already doing with Siena

Support, shopping and retention often happen inside the same conversation. Someone asking about fit may be ready to buy. Someone asking to cancel a subscription may just need a different delivery schedule. An agent needs enough context to offer the right option, follow the brand's policies and respect the customer's decision.

Our Shopping Agent answers product questions, recommends products and helps customers complete purchases without leaving the conversation. Our Social Agent handles moderation, community engagement and customer support across comments and DMs, including the questions that show up under paid campaigns. A marketer can spend all day earning someone's attention. It helps if the brand is there when that person finally asks a question.

At Eskiin, Social Agent took weekly comment engagement from roughly 1,600 to 2,400 and gave their social lead back at least 15 hours a week. Weekend replies that used to take 48 hours now arrive in under a minute, while the customer is still deciding.

The same understanding reaches the people making decisions inside the brand. At SPANX, Siena spotted a pattern in shipping complaints tied to one region. The team dug in and found a carrier issue affecting nearly every delivery there. Instead of treating each complaint as a one-off, they had a specific operational problem to fix.

At a consumer packaged goods brand, a proactive Siena agent picked up early signs of a manufacturing defect in customer feedback. Product development and operations acted within days. Customers were already describing the problem. Siena helped the team see it.

At Coterie, Ask Siena cut customer analysis that used to take several hours down to ten minutes. Insights teams, analysts, QA teams and executives use it to understand what customers need and where operations are falling short. Marketers use the same feedback to shape campaigns. The information gathered while taking care of customers becomes useful across the whole company.

Building for the way consumer brands actually operate

An apparel brand's exchange policies have little in common with a subscription brand's retention journeys. Inside a larger brand, requirements vary across storefronts, regions, technologies and policies. We chose to focus on consumer brands because those differences change what an agent actually has to do.

Our agent architecture was built through hundreds of consumer-brand deployments. Each one taught us more about the context agents need, the systems they act in and the controls that keep their work on brand and within policy. Those lessons are built into the infrastructure brands use to deploy Siena today.

Anyone can build a support agent prototype over a weekend. Putting one in production, in front of real customers, with a support team, complex business rules and a dozen systems to act in, is a different problem entirely. Then you have to keep it running, keep it secure and keep it on script when someone tries to pull it off. We've spent three years on exactly that so our customers don't have to.

Shared context gives these agents a common foundation. They still need different ways of working. Voice needs quick responses and natural turn-taking. A public social reply carries different consequences from a private support exchange. Getting approved by Meta and TikTok to connect social conversations to customer context took us six months on its own. A brand can bring everything it has already invested in customer understanding into those channels while deploying each agent for the job it needs to do.

As brands add more agents, managing them becomes part of the job. Teams need to keep policies current, understand how agents are performing and carry improvements across all of them. Siena's shared agent infrastructure gives them one operating environment for that, alongside the customer context their agents run on. Autopilot helps build and improve agents, while the Quality Agent evaluates their work. Operators can A/B test policies, journeys and discounts on a chosen share of customer traffic, compare responses and outcomes, and roll out what works.

We also don't charge per seat. You pay for outcomes and for intelligence. The help desk is included, so as your agents take on more of the work, more of your team can access the same workspace and the same customer context without the bill going up per person.

What's next

The Series A lets us build more autonomous agents and expand the shared infrastructure behind them. Once agents work from a brand's accumulated customer context, there's a lot more they can do: a recommendation that accounts for what someone owns and why they returned something else, a message that addresses a concern they raised before buying, a subscription reminder that reflects the delivery schedule that actually works for them. One-to-one experiences become useful when they carry the history of the relationship.

The same understanding that helps an agent resolve a conversation can help a founder decide what to build next. Our ambition is for agents to manage more of the customer journey, with the context, tools and oversight to do it well.

There's plenty left to build, and we're building it alongside the brands who got us here. If you want to help build the Agent of Record for consumer brands, we're hiring.