Shouldn’t customer service AI remember you? Siena to expand 'Agent of Record' combining purchase history with support and social

If you've spent any time seeking help from a customer service chatbot online, you've probably been frustrated by how slow, limited and "dumb" they can be — requiring you to rephrase your basic problem multiple times before you finally give up and demand a human.

And forget about the agent knowing already what you purchased or anything about your prior history with the company — you often waste precious minutes re-explaining information the vendor theoretically already has. Nor is it likely to know if you complained already to the brand on social media or had a discussion in direct messages with a company support bot there.

Well, the AI customer support startup Siena is out to change all that, offering a better, more unified experience that unifies a customer's data so it can be accessed by support chatbots and used for future, more tailored product and service recommendations.

The New York company calls its approach an “Agent of Record.” Its pitch is that conversations should contribute to a persistent understanding of each customer, helping both software and people decide what to do next. Individual customer profiles accumulate information from purchases and conversations, giving the brand’s agents a common understanding of each shopper. Existing commerce and business systems continue to maintain their own records, which Siena draws on to inform its responses and actions.

Siena already serves hundreds of consumer brands, including FIGS, SPANX, Comfrt and HexClad, according to materials shared with VentureBeat. A $17 million Series A announced today and led by York IE, with Joyance Ventures, Aglaé Ventures and Sierra Ventures participating, will finance further development. Siena says the round brings its total funding to $29.7 million.

There is an important availability distinction. Siena says newer shopping, social, voice and customer intelligence capabilities are in early access, with some features available or rolling out now. A native help desk and additional proactive agents are on its development roadmap. The October 6 announcement does not establish general availability for the entire suite.

Customer memory that reaches beyond support

Siena began shipping its support product around the end of 2022 and beginning of 2023, cofounder and CEO Andrei Negrau told VentureBeat in a recent video call interview. Its founders previously worked in e-commerce, including as Shopify merchants.

That background shapes the company's target market: midsize and enterprise consumer brands, rather than every industry needing a chatbot.

“We're not building agents for everyone,” Negrau said.

The underlying idea is to combine transaction records with information customers reveal in conversation. Siena Intelligence brings together purchases, subscriptions, reviews, loyalty information, brand knowledge and interactions, according to the company. A return explanation can inform a future recommendation; repeated complaints can reveal a broader product problem.

Siena's Memory product describes capturing preferences and feedback for reuse in subsequent exchanges. In the interview, cofounder and president Lisa Popovici described consolidation around a shared knowledge graph.

That does not mean Siena replaces every database a retailer already operates. Negrau said the platform draws information from commerce, order management and enterprise resource planning systems, which continue to hold records for their respective business functions. Siena aggregates customer context and makes it available to its agents as needed.

For technical buyers, the distinction matters: adopting Siena still requires connecting the systems that hold authoritative information about orders, refunds and subscriptions. Shared memory cannot substitute for accurate operational data.

Negrau said employees can query the accumulated information using Ask Siena, including through a Slack integration. He also described access through Model Context Protocol, or MCP, connections, which let AI tools connect to external information and services. The supplied materials do not specify connector coverage, access rules or availability for every such connection.

From answering questions to completing commerce workflows

Siena's website describes agents that can update delivery addresses, create return labels, process refunds and arrange replacements. Subscription workflows include skipping deliveries, pausing subscriptions and switching products.

Its shopping agent handles product questions and recommendations. Its social offering covers comment moderation, community interaction and customer assistance in direct messages, including questions beneath advertisements. Voice agents use the same customer context, according to the company.

The aim is to preserve information as the customer moves between those encounters. A preference disclosed during support should remain useful when the person shops again, rather than forcing another agent to start from scratch.

Siena also describes a shared inbox for employees and AI, performance assessment through a Quality Suite, live A/B testing and an Autopilot tool to help configure agents.

The next stage is more proactive. In a follow-up email, Siena's representative said planned agents would respond to signals such as a likely cancellation, an incomplete purchase or a suitable recommendation opportunity, rather than waiting for an inbound request.

Those plans should not be confused with capabilities available across all deployments today. Negrau also said the company is building its own help desk, intended to be included in the Siena offering rather than sold through additional per-seat charges.

What customers report so far

Siena says its deployments automate as much as 80% of customer interactions. That is an upper-bound claim, not a published average across its customers.

More specific examples suggest how brands use the product. A SPANX testimonial on Siena's site attributes automation of half its conversations, customer satisfaction above 90% and a halving of handling time to the platform. Siena also says its intelligence tools helped SPANX trace a regional increase in delivery complaints to a particular shipping carrier.

The company reports that Eskiin increased weekly comment engagement by 50% using its social agent and saved its social lead 15 hours per week. At Coterie, it says a feedback analysis exercise that previously required weeks became a ten-minute query through Ask Siena.

The potential benefit extends beyond reducing support workloads. In a testimonial published by Siena, True Sea Moss marketing leader Luka K. describes using customer analysis to identify demand for smaller flavor samples, informing a new sampler product.

These are company-supplied results and hosted customer testimonials. The materials do not provide an independent audit, consistent measurement periods or a controlled comparison with rival systems. They illustrate reported uses, rather than establishing a performance ranking.

How Siena compares with Gorgias, Fin, Sierra and Netomi

Commerce specialization gives Siena a defined market, but neither shared context nor agents that take actions are exclusive to it.

Gorgias documentation describes an ecommerce agent combining store knowledge, configurable behavior and actions in connected applications. Its Shopping Assistant uses product information to guide purchases. That makes it a particularly direct comparison for brands seeking both presale assistance and postpurchase automation.

Intercom's Fin offers another route: use the agent with Intercom's help desk or connect it to an existing customer service platform. Sierra advertises deployment across chat, SMS, WhatsApp, email and voice, alongside memory connecting conversations and business systems, proactive engagement and conversation analysis.

Netomi also competes for enterprise customer service deployments, including retail. Its platform supports autonomous and employee-assisted interactions, product recommendations and proactive service, with controls such as confidence scoring, fallback behavior and audit trails. It says customers can move between communication channels without losing context. Its Kustomer integration illustrates its approach of connecting AI to existing support and business systems.

The distinction is one of focus: Siena emphasizes consumer-brand workflows and consolidating support, shopping, social operations and customer intelligence. Netomi markets a broader enterprise platform across industries including travel, insurance, banking and retail. That makes Netomi relevant for organizations prioritizing complex integrations and governance, while Siena's commerce specialization is a proposition to test against a retailer's specific needs. The available evidence does not establish that either delivers better automation or lower total costs.

Comparison reflects vendor documentation, not an independent test. Sources: Siena pricing, Gorgias pricing, Gorgias billing documentation, Intercom pricing Sierra and Netomi’s platform listing.

The billing units deserve scrutiny. Siena's advertised automated ticket is not automatically equivalent to a Fin outcome. Intercom counts confirmed resolutions, certain conversations where customers do not request further assistance, and completed workflows that can include handoffs. A lower headline rate therefore does not establish a lower cost for equivalent work.

Siena's more defensible differentiation is its proposed combination of commerce-specific implementation, social operations and customer intelligence within one platform. Netomi and Sierra's context and proactive-service offerings make shared intelligence a competitive requirement rather than proof of a unique advantage. Whether Siena's approach reduces integration work or improves outcomes requires testing against each brand's actual workflows.

The enterprise buying decision

Siena's pricing page lists a $750 monthly platform charge and $0.90 per automated ticket, alongside onboarding and support.

At those advertised rates, 10,000 automated tickets would imply $9,750 per month before any other charges or negotiated terms. The page does not establish that every newer agent is covered by that calculation.

It also lists SAML single sign-on, granular permissions, audit logs, human escalation, an uptime service agreement and SOC 2 Type II compliance. Buyers should verify which protections and products their contract includes.

Important technical details remain unspecified in the supplied materials: underlying model choices, retention periods, data residency, cross-channel identity matching and how conflicting or outdated customer memories are corrected. These details affect whether a shared profile remains useful as interactions accumulate.

Negrau's argument against building internally centers on operating costs: business rules, connected tools, security, platform approvals and ongoing maintenance. He said Siena spent six months navigating approval processes for Meta and TikTok integrations. That is the company's experience, not a universal implementation timeline.

For prospective customers, the practical test is whether information collected in one interaction reliably improves the next—and whether the agent can complete the resulting task under the brand's rules. Siena's existing deployments offer a starting point for that evaluation; its broader roadmap still needs to prove itself in production.