Treasure AI

Treasure AI is the new name of Treasure Data, the enterprise customer data platform (CDP) company. The About page says "Before we were Treasure AI, we were Treasure Data. Since 2011, we've built the most trusted enterprise customer data platform (CDP), serving over 400 enterprise customers with billions of unified profiles." It now sells itself as an Agentic Experience Platform: a CDP, a customer engagement platform and AI agents in one system, used to run personalized campaigns across email, web, paid media, mobile apps, SMS and LINE. It is not a newsletter tool or a self-serve email service. It is for large brands that want email to be one output of a unified customer data layer.

How email fits in

The product page lists email marketing, mobile app marketing, SMS marketing and LINE marketing among its channels and says teams "Use Treasure AI Suites to launch and manage personalized experiences across email, web, paid media, and every other channel that matters to your business." The Engagement AI Suite is the part aimed at email teams: the site describes it as powering "cohesive omnichannel experiences" and says it is "Built for lifecycle marketers and campaign managers".

The other suites support it. The Personalization AI Suite delivers real-time, one-to-one digital experiences for digital marketers and product managers. The Creative AI Suite helps creative and brand marketers produce one-to-one personalized marketing assets. A paid media suite serves growth and ABM marketers running ad campaigns, and the Service AI Suite brings the same profiles to support reps, clienteling and B2B sales, with an AI chatbot connected to the Diamond Record, on-site status updates and surveys.

The product page gives example programs a business user can launch: "abandoned cart recovery, optimize paid media spend, personalize product recommendations, or re-engage high-value customers". A case study on the same page describes Brother Sales in Japan moving marketing automation natively into its Intelligent CDP, with targeted, real-time emails that reduce support costs and improve loyalty. The product FAQ says the platform is built for business users and does not require data science support.

AI agents with a human in the loop

The core pitch is agents that do the work while people set direction. The site says "Agents continuously analyze behavioral signals, purchase history, and real-time context against your defined business objectives." The About page puts the division of labor this way: AI agents accelerate marketing "while humans set the direction, define the guardrails, and bring the creativity." The Intelligent CDP already includes AI features such as the Audience Agent, and the AI Agent Foundry is where customers build custom agents.

Treasure AI Studio is the conversational workspace on top. Users describe a segment, a multi-stage customer journey or a scheduled workflow in plain language, and Studio writes the queries and configuration for review before anything runs; its FAQ says "Nothing runs without your approval." It comes with 50+ pre-built skills, runs on web, mobile (iOS) and desktop (Mac), and keeps chat history and context across sessions. The Studio page gives one example: it "built a complete waterfall report with segment creation in minutes. Last year, a similar report took 2 months and still couldn't build segments from the results." For an email team, the practical use is turning a campaign brief into a live audience segment, or a re-engagement strategy document into a journey with decision points, condition waits and activation steps. The FAQ says the AI works with the CDP's metadata and structure, not the customer PII itself.

Data and integrations

The CDP unifies data into profiles that the suites activate. It can run as a Complete CDP on Treasure AI's managed infrastructure with no warehouse needed, or as a Composable CDP on the customer's own Snowflake, Databricks or BigQuery warehouse. The integrations page lists the supported formats: real-time streaming, batch file processing, blob or object storage and (S)FTP, APIs, webhooks, and native SDKs for JavaScript and the server side. Where no connector exists, the page says the company's "dedicated core engineering team for connectors can build a production-ready version in 6–8 weeks".

Pricing

The pricing page publishes the model but no amounts; quotes come on request. Treasure AI prices "by customer profiles or behavioral events", not by query volume or compute. There are three tiers:

• Intelligent CDP: an annual subscription with tiered pricing by profiles and behaviors, covering data ingestion and profile unification, the AI Agent Foundry, and security, privacy and governance controls.
• Intelligent CDP plus AI Suites: a fixed annual license per suite covering SaaS installation, activation and standard support, plus consumption-based AI Suite Credits for real-time features, tracked monthly and usable across any suite.
• Treasure AI Studio: included with the plan as a free workspace with no license or seat fee; usage draws on the same credit pool. The Studio page says one AI Credit now covers 600 conversations, up from 100.

A two-week proof of concept lets buyers try before they buy. The pricing FAQ says a typical implementation takes 8 to 12 weeks, which AI features can shorten to days. Standard support is included for all customers, with upgrade packages, a Customer Success Manager and a Technical Account Manager. A Trade-Up Program offers incentives to replace an existing CDP, ESP or customer engagement platform.

Security, compliance and data location

The security page is detailed. Treasure AI undergoes annual SOC 2 Type 2 and SOC 3 audits, ISO 27001:2022, ISO/IEC 27701, 27017 and 27018 certification audits, an annual HIPAA audit, and PrivacyMark audits in Japan; it also cites CSA Star Level 1 and a TRUSTe Responsible AI certification. A DPA covering GDPR and CCPA is available, and a BAA is offered to HIPAA-covered entities on request. The page says affected customers are notified within 72 hours of a confirmed breach.

Data runs on AWS. The security page names 5 AWS regions: US East (N. Virginia), US West (Oregon), Europe (Frankfurt), AP Tokyo and AP Seoul, with separate encryption keys per region, and says region selection is contractually binding with no cross-region replication without explicit consent. It reports 99.99% uptime over the past 90 days. Access controls include SAML 2.0 SSO, MFA enforcement and permission policies across 6 layers, and column-level dynamic masking lets marketers use PII columns in campaigns without seeing the values. On AI, the named LLM providers are Anthropic and OpenAI via Amazon Bedrock, agents run in a server-side sandbox with every external call logged, and customer content is contractually excluded from model training.

Size and recognition

The About page lists 550+ employees, 80+ Forbes Global 2000 customers, 75+ countries supported, 2B+ profiles managed and ">50%" of the global connected population served. It cites The Forrester Wave for Customer Data Platforms for B2C, Q3 2026, IDC MarketScape reports, and 4.5/5 stars on Gartner Peer Insights. The home page lists customer stories including a leading automaker that boosted ad efficiency by 14.5X, a top brewer unifying 2,000+ data sources and an entertainment giant that saved $1M.

Who it suits

Treasure AI suits large consumer brands that want email, SMS, app and ad campaigns driven from one governed customer data layer, with AI agents doing the segmentation and journey work under human approval, and with security documentation that satisfies an enterprise review. It is a poor fit for a small team that simply wants to send a newsletter: the product is a CDP first, priced on annual licenses by profiles and behaviors, and email is one channel among many.