Today, Treasure Data becomes Treasure AI.
After fourteen years of building the most trusted enterprise customer data platform — over 400 enterprise customers, billions of unified profiles — we're launching an Agentic Experience Platform: where AI agents run the Customer Intelligence Loop continuously, while humans set the direction, define the guardrails, and bring the creativity.
We're introducing Treasure AI Studio — your command center on Web, Desktop, Mobile, CLI, and Voice — and AI Super Agents that give every marketer the execution capacity of an entire team and every analyst instant access to customer intelligence. This post is the why, the what, and the promise behind it.
Why now
We founded Treasure Data on a mission: build the data foundation to improve human life — by turning customer data into intelligence that serves every person as an individual.
That mission hasn't changed. What changed is the world around it.
For over a decade, a customer data platform was a tool humans queried. Marketers logged into dashboards, built segments, and scheduled campaigns. The data sat in a system, waiting for someone to ask it a question. That model created enormous value. It also hit a ceiling.
In 2026, the primary user of customer data is no longer just a human. It's also an AI agent — querying unified profiles at millisecond speed, autonomously deciding the right message, the right channel, the right experience for every single customer. Not once a week. Continuously, 24/7.
The market doesn't need more software. It needs outcomes. Our customers didn't buy Treasure Data to have a database. They bought it to increase revenue, reduce churn, and personalize at scale. The distance between data and outcome should be as close to zero as possible. AI agents collapse that distance — guided by the humans who know their customers best.
And AI won't replace people. But people who use AI will replace those who don't. AI is not here to take anyone's job. It's here to take the parts of your job you shouldn't be doing — the manual segmentation, the A/B test babysitting, the spreadsheet wrangling at 11pm on a Sunday. What AI can't do is set the vision, choose which customers matter most, or know when a number on a dashboard represents a real person having a bad day. That's human work. That will always be human work.
That's why we rebuilt. Not a feature. The foundation.
The Customer Intelligence Loop
Every CDP (Customer Data Platform) exists to close what we call “Customer Intelligence Loop”: Collect → Unify → Understand → Decide → Engage — and feed the result back to Collect.
The speed this loop turns is the single best measure of a CDP's power. For most of the industry's history, it turned slowly — at the speed of a marketer clicking through tabs, waiting for queries, scheduling sends.
AI agents change the clock speed. They run the loop continuously — in minutes, not weeks. They discover segments humans would miss. They generate creative at the speed of thought. They optimize campaigns in real time, not in next month's review meeting.
The question is no longer "do you have a CDP?" It's: how fast can you close the Customer Intelligence Loop?
The Agentic Experience Platform
An Agentic Experience Platform is a unified foundation where AI agents run the Customer Intelligence Loop continuously, while humans set the direction, define the guardrails, and bring the warmth and creativity that make every customer experience feel human.
Treasure AI is the control plane for the Customer Intelligence Loop — orchestrating CDP (Customer Data Platform), Omni-channel Messaging (Email, SMS, Mobile Push, In-App Message, LINE), Personalization, and AI in one platform.
Your data can be stayed where it lives. Snowflake, Databricks, BigQuery, your own warehouse — our AI connects to your data plane, wherever it is.
One control plane, one loop, machine speed.
Treasure AI Studio: Your command center
Today, alongside our rebrand, we're also launching Treasure AI Studio: the single interface to your AI agents and customer data — from any device, any environment. Web, Desktop, Mobile, CLI and Voice. You can interact with the platform anytime, anywhere.
Open the Web app at your desk and AI is already working — segments discovered overnight, creative drafts waiting for your review, campaign performance updated in real time.
Open the Desktop app and go deeper — build workflows, fine-tune agent guardrails, run complex analyses side by side. The full power of Studio, native on your machine.
Open the Mobile app on your morning commute, and last night's campaign results are already summarized — what worked, what didn't, and a fix ready to deploy before you reach the office.
 |
 |
| Desktop | Mobile |
Open your terminal, and Treasure Code is ready. Treasure Code is a CLI for engineers that ships with deep platform knowledge — it comes with pre-built agentic ‘skills’: query optimization patterns, workflow syntax, CDP data models, engine-specific conventions. Describe what you want; it produces production-ready code you can read, review, version-control, and roll back. Works in Claude Code, Cursor, Windsurf, and ChatGPT — so engineers stay in their editor.
Clip on Treasure AI Voice — our AI Note Taker device powered by Plaud.ai — and every meeting, every brainstorm, every hallway conversation becomes data your AI can act on. Walk out of a customer call, and the transcript is already in AI Studio — key insights extracted, action items flagged, ready for your agents to turn into the next campaign or analysis.
AI Studio is where your team meets AI. It's conversational, intuitive, and available everywhere — on screen, at the command line, and in the room with you.
Under the hood, Studio orchestrates an expanding ecosystem of AI super agents — purpose-built AI agents, each specialized for a domain of the Customer Intelligence Loop.
AI super agents for marketers
Your marketing team's force multiplier. For marketers, Treasure AI Studio leverages super agents that act as a force multiplier to solve for marketing use cases. These don't replace your marketers — it gives each of them the execution capacity of an entire team:
- Discovers high-value audience segments that humans would miss across billions of profiles
- Generates creative — ad copy, email subject lines, landing page variants — tailored to each segment using your brand voice and CDP data
- Orchestrates campaigns across email, ads, push, and SMS — based on real-time customer behavior, not last month's report
- Personalizes every touchpoint using the full depth of your unified CDP data
- Optimizes continuously — reallocating budget, adjusting creative, shifting channel mix — autonomously, within the guardrails your team sets
In early access, one enterprise retail customer (50M+ profiles, 12 data sources) compressed their segment-to-launch cycle from 5 business days to under 8 hours — discovering a 340K-person high-intent segment that manual analysis had missed for months, driving a 23% lift in campaign conversion within two weeks.
This super agent layer is agentic marketing made real. It runs at machine speed — with your team in the driver's seat.
AI super agents for data analysts
The end of "can you pull me a report?" For data analysts, Treasure Code leverages super agents that act as a force multiplier to solve for data and analytics use cases. It gives every team instant, conversational access to your customer data:
- Answers complex business questions in seconds — "Which accounts showed the highest engagement increase in Q1?" "What's our churn risk by segment?" — without writing SQL or waiting for the data team
- Generates visualizations, trend analyses, and cohort comparisons on demand
- Connects to your CDP, data warehouse, and BI tools with enterprise-grade security and permissioned access
In early access, a financial services team went from waiting 3 days for analyst reports to getting answers in under 60 seconds — surfacing a retention risk across a $12M account segment that had gone undetected in quarterly reviews.
This democratizes customer intelligence. It operates at the speed of thought — harnessed by human curiosity and judgment.
Governance and trust — built in, not bolted on
Agentic AI without governance is a liability, not an innovation. We built Treasure AI for the industries where trust is non-negotiable.
Your controls, enforced at every layer. AI agents inherit your existing access controls and permissions. Every agent operates under the principle of least privilege — scoped to only the data and actions its task requires.
Every action auditable. Every agent action is logged in a tamper-evident audit trail — what data was accessed, what decision was made, and why. Designed for SOC 2, GDPR, CCPA, and financial regulatory requirements.
Human approval where it matters most. High-impact actions — campaign sends, large-scale data exports, external activations — require explicit human approval by default. Your team controls the threshold.
Your data stays yours. Your data is never used to train AI models per our AI terms. Customer data processed during inference is encrypted and not used for model training. Our AI Terms make this commitment explicit and legally binding.
Built on enterprise-grade infrastructure. Our AI systems run on AWS Bedrock — with SOC 2, ISO, GDPR, and HIPAA-eligible compliance, VPC isolation, encryption at rest and in transit, and built-in Guardrails for hallucination prevention. Foundation models from Anthropic Claude and OpenAI are accessed under strict data processing agreements — no customer data is ever used for model training by any provider.
Backward compatibility commitment: All existing APIs, SDKs, and workflow definitions remain fully supported. Again, we maintain the backward compatibility for all API endpoints. No breaking changes. No forced migrations.
Our design principle: Harnessed by humans
Every AI we build at Treasure AI follows one rule: it is harnessed by human warmth and creativity.
Not "supervised." Not "monitored." Harnessed — the way a rider harnesses the power of something far faster and stronger, directing it with intention, judgment, and care.
AI can process a million profiles in seconds. It can't feel what a customer is going through. AI can optimize a campaign to statistical perfection. It can't decide that this particular customer deserves a handwritten note instead. AI can flag the accounts most likely to leave. It can't choose to call someone just to say thank you.
The warmth is human. The creativity is human. The empathy is human.
AI is muscle. Humans are the mind and heart.
This isn't a philosophy we'll revisit when it's convenient. It's a design constraint we build into every product. An AI agent doesn't just accept guardrails — it's architected around them. Your team sets the budget ceiling — the agent stops before it's reached. Your team approves the campaign brief — no message goes out without it. Your team defines the brand voice — the agent writes within it, not around it.
Every AI that carries the Treasure AI name will be built this way. That's our promise.
What's changing — and what's not
What's changing:
- Our brand name: Treasure Data → Treasure AI
- Our domain: treasuredata.com → treasure.ai
- Our product: Treasure AI Studio launches on Web, Desktop, Mobile, CLI and Voice — with AI agents
- Our positioning: from a Customer Data Platform to an Agentic Experience Platform
What's not changing:
- Our mission: build the data foundation to improve human life
- Our platform and API contract: everything you rely on today continues to work
- Our people: the same team that built the most trusted enterprise CDP
- Our design principle: every AI is harnessed by human warmth and creativity
- Our commitment to data governance, privacy, and trust — now stronger than ever
To our team, customers, and the industry
To our team
The platform we're announcing today would not be possible without the data infrastructure you've spent years perfecting. That foundation is not being replaced — it is being unleashed.
Every integration, every identity resolution algorithm, every late-night deployment that made a customer's campaign work — that's the foundation AI super agents stand on. You didn't just build a customer data platform. You built the only enterprise customer data foundation trusted enough for AI agents to act on — and human enough to remember that every data point is a person.
The name is new. The ambition is the same, just bigger.
To our customers
First of all, everything you rely on today continues to work. Every API endpoint, every SDK, every integration, every workflow. All existing API endpoints will remain fully operational with backward compatibility guaranteed. Your segments, workflows, and connections are exactly where you left them.
What you gain are new teammates: AI super agents that execute at machine speed, Treasure AI Studio that puts intelligence at your fingertips on any device, and Treasure Code that gives your engineers a faster way to build — all harnessed by your team's warmth and creativity.
Treasure AI Studio is included for all existing customers. No additional cost to get started.
We are also increasing the conversations you get for your current plan: same price, more value. As of today, we are increasing the number of Conversations you can use with 1 AI Credit from 100 to 600. We're giving you more Credits specifically so you have the headroom to explore what AI-native data work looks like — without worrying about burning through your allowance. If you haven't had a chance to explore Treasure AI Studio yet, now's a great time — you have more runway than before.
To the industry
The CDP category is being reborn. The question is no longer "do you have a CDP?" It's: "How fast can you close the Customer Intelligence Loop? Can it power AI agents that act on customer data at enterprise scale, with enterprise governance? And are those agents harnessed by humans?"
We believe the answer defines the next decade of customer experience.
Treasure AI is what Treasure Data was always becoming. Treasure AI Studio is the proof.
Available to all customers from day one in the US. No waitlist. No phased rollout. Log in and meet your new teammates. Available in other regions next month.
Let's build. — Kazuki Ohta (Co-Founder & CEO, Treasure AI)
Frequently Asked Questions
Why is Treasure Data rebranding to Treasure AI?
In 2026, the primary user of customer data is no longer just a human — it's also an AI agent. We rebuilt our foundation to become an Agentic Experience Platform where AI agents run the Customer Intelligence Loop continuously, while humans set the direction, define the guardrails, and bring the creativity. The name reflects what the platform has become: an AI-native system built on the most trusted enterprise data foundation.
What is Treasure AI Studio?
Treasure AI Studio is the single interface to your AI agents and customer data — available on Web, Desktop, Mobile, CLI, and Voice. It orchestrates AI super agents that discover audience segments, generate creative, optimize campaigns, and answer complex business questions in seconds. Think of it as your command center: open it at your desk, on your phone, or in your terminal, and AI is already working.
What is the Customer Intelligence Loop?
The Customer Intelligence Loop is the cycle every CDP exists to close: Collect → Unify → Understand → Decide → Engage — and feed the result back to Collect. Historically, this loop turned at the speed of a marketer clicking through tabs and scheduling sends. AI agents change the clock speed, running the loop continuously in minutes instead of weeks. The speed this loop turns is the single best measure of a CDP's power.
How does Treasure AI handle data governance and privacy?
Governance is built in, not bolted on. AI agents inherit your existing access controls and operate under the principle of least privilege — scoped to only the data and actions each task requires. Every agent action is logged in a tamper-evident audit trail designed for SOC 2, GDPR, CCPA, and financial regulatory requirements. High-impact actions — campaign sends, large-scale exports, external activations — require explicit human approval by default. Your data is never used to train AI models, and this commitment is explicit and legally binding in our AI Terms.
