On September 9, Klaviyo traditionally held its annual K:BOS 2026 conference, during which new features of the Klaviyo platform were unveiled. As Master Elite Partners, we were there and heard all about the latest updates firsthand. At the conference, we gathered insider information and all the details about the new capabilities.
From an AI-powered marketing agent that creates ad campaigns while you sleep to a customer support system that independently resolves 86% of inquiries, read on to learn about all these features and more so you can take full advantage of the new functions included in your subscription.
Klaviyo as an Autonomous B2C CRM: Direction of Updates
This year, Klaviyo is actively expanding beyond the scope of email marketing. Recently, the platform has significantly changed its positioning and is now officially classified as a standalone CRM for B2C. The new features unveiled at K:BOS 2026 are specifically focused on this newly established positioning.
New features aren’t just minor interface tweaks. The Klaviyo updates released this year address every aspect of how a DTC brand conducts customer retention marketing. They focus on how campaigns are planned, how artificial intelligence executes them, how your support team or AI agent handles post-purchase communications, and how customer data flows between systems and channels.
- According to Omnisend statistics, 37% of all email revenue is already generated by automated workflows, which on average account for only 2% of all email campaigns.
- Furthermore, according to Gartner’s estimates, 60% of brands are expected to use agent-based AI for personalized interactions by 2028, indicating the continued rapid growth of this trend.
These areas are part of Klaviyo’s current direction in developing and scaling its platform. Next, as experts in retention marketing, we’ll discuss in detail the 5 most important platform updates that are significant for eCommerce marketing leaders.
New Klaviyo AI Features
Klaviyo’s focus on developing autonomy, which is currently the platform’s priority, is also aimed at advancing embedded AI agents and models. More on these below.
‼️Despite the high level of autonomy, all campaigns and agent responses must be manually approved to protect the brand’s reputation.
1. Key Release of the Year: Klaviyo Composer
As we just mentioned, Klaviyo keeps up with trends and is evolving in the field of artificial intelligence. Composer is its latest solution in this area.
Composer is a proactive marketing agent powered by artificial intelligence. It’s not just some chatbot assistant, but an internal system that constantly monitors your account, identifies opportunities, and provides ready-to-use results for your review.
Moreover, Klaviyo announced that it is currently continuing to develop a mobile app that will be based around Composer, giving teams access to conversation analytics and a campaign builder from anywhere. These features, without a doubt, place Klaviyo Composer amongst the most significant advancements of 2026. It was launched in public beta back in June, and has been significantly expanded since then, leading up to the presentation. Here are the main new features related to AI agents that Klaviyo unveiled at K:BOS 2026:
- Analysis with links. Thanks to the new SQL engine in the Klaviyo Data Platform, every Composer response now includes the exact query on which it was based. This means you can now provide your CFO or CEO with analytics that follow consistent logic.
- Proactive tasks. Composer can now automatically perform regular audits. It will generate monthly performance reports for you so you don’t have to set them up from scratch. Most importantly, it will alert you to any drops in customer engagement in the flow before you even notice them yourself. As an added bonus, there’s another proactive feature: a notification about audience overlap. All of this now works even if no one opens Klaviyo.
- Customizable Composer skills. Klaviyo’s AI agent can now be trained on brand specifics permanently. No more repetitive prompts with explanations, Composer will learn your standards as guidelines for brand style, email template standards, and content hierarchy rules, consistently applying them.
- Clarifying questions during campaign creation. It often happens that when creating a prompt for the AI, you forget a specific detail, and the AI produces a low-quality result. Klaviyo has solved this problem in its AI, and now Composer asks additional relevant questions when key details are missing from the prompt. This ensures a more accurate result, rather than forcing you to repeatedly work on a rough draft.
Why This Matters for Your Brand
We’ve already mentioned Gartner’s predictions about brands rapidly transitioning to agent-based AI. Composer gives Klaviyo brands the opportunity to stay ahead of these trends. Starting today, you can automate your analysis and planning processes and focus your efforts on strategy, creative direction, and channel development.
Potential drawbacks:
➖ The risk of AI malfunctions and accidental deviations from the brand’s tone or style if left unsupervised.
➖ The need for the team to spend additional time manually reviewing campaigns generated by the automated system.
➖ Potential limits on the number of AI requests on basic pricing plans.
2. Development of Klaviyo Customer Agent’s Autonomy
Klaviyo Customer Agent is another influential product on the platform among AI-powered agents, and like the others, it also received significant enhancements in 2026.
Since its launch, the Customer Agent has exchanged over 1.3 million messages with real customers. Further proof of its impact comes from metrics shared by the brand Naked Wardrobe, which uses it to autonomously resolve 86% of customer inquiries. In 2026, the agent received three significant updates:
- Expanded international localization. Klaviyo is expanding Customer Agent’s language capabilities, and it now supports over 100 languages with automatic locale detection. Detection occurs through integration with Shopify Markets and includes the appropriate product catalog, prices, and currency for each market.
- AI-powered call handling (coming soon). Klaviyo has announced the upcoming release of full voice processing for incoming calls via Customer Agent. It won’t just answer questions, but will also know the customer’s complete interaction history upon receiving the call. No more long waits in line to speak with an operator or numerous transfers.
- Personalized learning from conversations. Personalization is now significantly more powerful, as every communication now becomes profile data available to Composer and enhances the level of personalization. Additionally, agents can learn not only from conversations with customers but also from conversations with you. Simply describe the improvements you need, and it will adapt through self-learning and simulation.
Why This Is Important for Your Brand
Support costs are often perceived as a drain on eCommerce profitability. If AI can handle calls independently and even use them to enhance personalization, it will transform from a cost center into a valuable retention asset.
Potential drawbacks:
➖ Errors in complex or non-standard customer inquiries, which may lower satisfaction rates.
➖ The complexity of initially configuring the voice agent to accommodate the specific characteristics and accents of regional markets.
➖ The risk of providing incorrect information about product availability or discounts in the event of synchronization failures.
3. Klaviyo’s New Personalization Through AI Specialization
Klaviyo has rebranded and completely redesigned Marketing Analytics. It is now called Klaviyo Personalization, reflecting the changes to the system. The results of customer analysis by an AI model are no longer generated by a single universal model, but by a specialized model for each decision.
The system processes over 4 billion signals per day for more than 9 billion profiles, and expanding the number of models not only helps avoid the generalization errors of a universal model but also provides significantly more accurate predictions for specific tasks. Now, Klaviyo will display recommendations that have been analyzed by a specialized AI model like that:
- Who to send to. A specialized model will optimize the audience based on current RFM data (recency, frequency, monetary value).
- What to show them. Based on the target customer, a separate model will determine what is best to send and show, based on their interaction history.
- When to send. Now, with customizable control groups and aggregated analytics, a dedicated model will determine the most effective personalized time to send the email.
Why This Is Important for Your Brand
The profitability of personalization is hard to deny. Its effectiveness has long been confirmed by a McKinsey study showing that companies that succeed at personalization generate 40% more revenue than their average competitors. Shifting from a universal segmentation approach based on purchase history to specialized AI models for each 1:1 decision will have a significant impact on revenue. The biggest difference will be seen in repeat purchase rates and LTV, especially for brands offering consumer goods or products that require replenishment.
Potential drawbacks:
➖ Low prediction accuracy for new brands or websites with limited historical data.
➖ It is difficult for marketers to manually verify and justify the logic behind individual AI recommendations.
➖ The risk of narrowing audience reach due to excessive hyper-segmentation.
Klaviyo System Upgrade: MCP & SQL
Having optimized the autonomy of its platform through innovations in embedded artificial intelligence, Klaviyo has also implemented updates beyond AI. Read on for more details.
4. Klaviyo Model Context Protocol: A New Addition to Integrations
The second most important feature of 2026 is a context-passing protocol for LLMs. This feature provides access to over 260 MCP (Model Context Protocol) tools that can be called directly from AI systems such as Claude and ChatGPT, without even opening the Klaviyo interface.
Model Context Protocol is an open standard that allows AI models to connect to external data sources in a structured and secure way. Klaviyo’s MCP server was introduced back in mid-2025, but its scale greatly expanded in 2026. Klaviyo is now one of the first marketing automation platforms to have an MCP server of this scale, which is provided by the manufacturer and developed in direct partnership with Anthropic, the company behind Claude.
More than 260 tools cover campaigns, flows, lists, segments, profiles, catalogs, templates, and metrics. In practice, this means:
- Your customer retention manager can ask Claude to pull 90 days of flow performance data, identify underperforming sequences, and even generate a brief report without logging into Klaviyo.
- A brand developer can create a real-time reporting dashboard in tools like Cursor or Lovable that will be updated with live data from Klaviyo, without having to set up an API.
- The marketing team can automate a weekly performance analysis, with the results posted straight to Slack.
- A brand can ask ChatGPT which product is the most popular this week, have it draft a campaign, and send it to Klaviyo ready for sending.
Why This Matters for Your Brand
Brands that integrate Klaviyo into AI-powered workflows via MCP can reduce the time between gaining insights and acting on them without increasing their technical staff. For DTC brands with small marketing teams looking to scale, this speed in moving from audit to action in minutes rather than hours quickly adds up over the course of a customer retention calendar.
Potential drawbacks:
➖ The need for technical specialists to create custom dashboards and integrate with external services.
➖ Risks of personal data leaks if third-party AI platforms are configured incorrectly.
➖ Dependence on the stability of APIs from third-party developers (Anthropic, OpenAI).
5. Reliable Data-Driven Responses from the Platform Thanks to SQL
Another significant update was the addition of SQL (Structured Query Language) support to the Klaviyo Data Platform. SQL in KDP allows you to query profiles, events, segments, lists, campaigns, and catalogs. Additional development is planned for the near future, which will allow SQL to also query forms, flows, templates, and custom objects. Here’s an example of how SQL works in real-life situations:
“A marketer urgently needed to determine which customer segment had the highest churn risk ahead of the fourth quarter. Previously, to get this answer, they had to wait for additional help from a technical specialist or estimate it approximately using numerous manual filters. The updated SQL allows this issue to be resolved with a single query. Klaviyo will formulate a query to the system and return a response based on the brand’s own data, along with the exact query that generated it.”
The precise query generation is the key feature that greatly differentiates the SQL function from a simple AI assistant. Each of its responses is accompanied by a verifiable link, which enhances reliability. To work effectively with SQL, the team will need to understand the logic of relational databases.
Why This Matters for Your Brand
Data-driven marketing decisions require clear justification, which is what distinguishes SQL from an AI agent. SQL in KDP bridges the gap between simply having data and fully processing it without extra effort. For brands with revenue ranging from $5 million to $50 million, this is often the bottleneck between knowing something and acting on it.
Potential drawbacks:
➖ Requires marketers to have a basic understanding of analytics and database structures.
➖ Risk of creating incorrect queries, which could lead to erroneous business conclusions.
➖ The advanced KDP features are primarily available only on higher-tier or premium plans.
Final Thoughts on K:BOS 2026 Updates
The most important announcement for CMOs at K:BOS 2026 is not only the update and enhancement of functionality but also a significant shift in the direction of Klaviyo as a platform. The transition from a traditional email/SMS marketing and automation platform to a standalone B2C CRM system is a development that reveals details about its features. Advanced AI agents and systems such as SQL are precisely the features that give the platform significant autonomy.
To better understand the platform’s latest capabilities, we recommend our Klaviyo Audit by Flowium. After conducting the audit, our specialists will identify areas that can be significantly improved with the new features and provide guidance tailored to your specific requests.
Frequently Asked Questions
Are the new features included in an existing Klaviyo subscription, or do they require an additional fee?
Composer has limited free capacity and paid usage; Customer Agent has separate pricing for AI conversations. MCP Server is available for free through Klaviyo’s developer tools. It is important to note that full access to KDP analytics and increased AI request limits may require switching to enterprise pricing plans.
Is it safe to connect Klaviyo to Claude or ChatGPT via MCP, and will customer data be used to train the AI?
Klaviyo MCP integration operates through secure API protocols. Under their standard API terms of use, Claude and ChatGPT do not use users’ corporate data to train public models.
‼️When transferring customers’ personal data to third-party AI models, it is important to strictly comply with GDPR requirements and verify the API’s privacy settings.
Do I need a technical specialist to connect Klaviyo MCP to Claude or ChatGPT?
A basic connection via Klaviyo’s official MCP server requires minimal technical knowledge and takes just a few minutes to set up. However, building custom dashboards in Cursor or Lovable will require a developer.
