What enterprise brands actually need from a CDP in 2026
The customer data platform (CDP) market has fractured. Some platforms are going wide, racing to consolidate marketing execution, email, creative, and orchestration under one roof. Others are going deep, investing in the data foundation that every downstream system depends on: identity accuracy, profile trust, and AI readiness.
For enterprise B2C brands evaluating an Amperity vs Treasure AI comparison, this is the first question worth answering. Not "which platform has more features?" but "where does each platform concentrate its depth, and which concentration matches our priorities?"
Note: Treasure Data rebranded to Treasure AI on April 20, 2026, repositioning from an "Intelligent CDP" to an "agentic experience platform." This post uses the new name throughout. Product references (Diamond Record, Treasure Boxes, Treasure Code, Marketing Super Agent) remain unchanged.
Both platforms are established, analyst-recognized, and used by large global brands. Both are investing aggressively in AI. But they arrive at the CDP problem from fundamentally different starting positions, and those origins shape everything from how they resolve customer identities to how they handle data architecture and team access.
The stakes for getting this decision right have increased. According to Gartner, at least 50% of generative AI (GenAI) projects were abandoned after proof of concept due to poor data quality, inadequate risk controls, escalating costs, or unclear business value. That finding reframes CDP selection as an AI infrastructure decision, not just a marketing technology purchase. The platform you choose to unify customer data will directly determine whether your AI initiatives produce trustworthy results or expensive mistakes.
How Amperity and Treasure AI compare
The platforms diverge most sharply on identity resolution methodology, data architecture, and the sequence of their AI investments. Treasure AI concentrates on breadth of marketing execution and real-time activation. Amperity's Customer Data Cloud concentrates on depth of identity accuracy, lakehouse-native architecture, and a data foundation designed for AI readiness.
Identity resolution
This is the single largest point of divergence between the two platforms, and the one most likely to determine long-term satisfaction.
Treasure AI's identity framework centers on what they call the Diamond Record, a single customer view constructed through their matching logic. The built-in matching capability is deterministic and rules-based: your team (or Treasure AI's professional services team) configures the rules that govern how records are linked. Probabilistic matching is available but requires "Treasure Boxes," which are essentially custom code packages that a services team builds and maintains. The distinction matters because it means the system doesn't autonomously learn from new data patterns. When a new data source comes online, someone has to write and validate new rules to incorporate those identifiers.
A second consideration is transparency. Gartner Peer Insights reviewers have noted limited backend visibility, and Treasure AI's own product documentation for the Diamond Record does not describe in-product tools for inspecting why specific records were merged or kept separate. For data teams that need to audit matching decisions, that gap can create friction with compliance and analytics stakeholders. Treasure AI's own documentation also notes that profile unification features will "likely need a developer on your team to implement."
Amperity's Identity Resolution takes a different architectural approach. The system combines deterministic matching, machine learning (ML)-based probabilistic matching, and transitive logic across dozens of purpose-built models trained specifically on the messiness of enterprise consumer data: nickname variations, name changes, address standardization, household-versus-individual detection. Match and merge are separated into distinct processes, which means how records are associated and how profiles are constructed can be tuned independently.
Amperity also introduces contextual identity: the ability to run multiple identity graphs concurrently on the same customer data, each optimized for a specific business use case. A marketing team might need a broader graph that maximizes reach. A loyalty operations team might need conservative, account-level matching. A compliance team might need the most restrictive, traceable graph possible. Rather than forcing every use case through a single identity framework to produce one 360-degree view of the customer, contextual identity lets each team operate with the matching tolerance appropriate to their function.
On stability, Amperity's adaptive approach keeps IDs consistent in day-to-day operations. When new data reveals a connection the system didn't previously have, it adapts and tracks the change transparently rather than silently reassigning identifiers. For brands with downstream systems that depend on persistent customer IDs (loyalty platforms, CRM, analytics), that consistency reduces operational risk.
AI and agentic capabilities
Treasure AI is making big AI bets in the CDP market. Over the past year, they have reoriented their entire platform narrative from "Intelligent CDP" toward an "AI Marketing Cloud" positioning, culminating in the April 2026 rebrand from Treasure Data to Treasure AI and the launch of Treasure AI Studio, a conversational workspace for marketing and data teams. In January 2026, they launched Marketing Super Agent, a multi-agent orchestration system designed to handle audience intelligence, strategy, creative generation, activation, and real-time optimization within a single governed workspace. In February 2026, they released Treasure Code, an AI-native command-line interface (CLI) that lets technical teams operate the entire platform as code, augmented with natural language input. According to Treasure AI, roughly a quarter of their customer base adopted Treasure Code within days of launch, which signals genuine product-market fit among technical users.
Treasure AI is also running an aggressive displacement play with their Trade-Up program, offering enterprises up to 24 months free if they replace their existing email service provider (ESP) or customer engagement platform (CEP) with Treasure AI's Engagement AI Suite.
Amperity's AI investment follows a different thesis. Rather than expanding horizontally across the marketing execution stack, Amperity concentrates AI on the data foundation and the gap between insight and action. The Customer Data Agent allows marketers to build segments and customer journeys through natural language conversation. Describe a campaign goal, and the agent proposes the segments, journey logic, and channel assignments needed to execute it. Critically, nothing executes until the marketer reviews and approves the plan, and every created asset can be undone in a single step.
On the data engineering side, Amperity's Identity Resolution Agent uses machine learning to dynamically match and merge fragmented customer records, compressing what historically required months of data engineering into a significantly shorter timeline. Chuck Data, an AI assistant that runs in the terminal, lets data engineers build customer tables, resolve identities, and tag personally identifiable information (PII) using natural language prompts.
The strategic distinction comes down to this: Treasure AI is layering AI across the broadest possible surface area of marketing operations. Amperity is grounding AI in the data foundation first, with the position that every AI agent, model, and decisioning system is only as reliable as the identity and profile data underneath it. Both approaches have merit. The question for buyers is which sequence, breadth-first or foundation-first, matches their current maturity and risk tolerance.
Data architecture and lakehouse integration
Treasure AI now offers two deployment modes: Complete Mode positions Treasure AI as the source of truth for customer data and Composable Mode positions the customer's own data warehouse as the source of truth, with Treasure AI reading from and writing to it. They have also introduced a "No Compute" pricing model that charges based on unified profiles rather than processing volume, which is designed to make costs more predictable.
Treasure AI offers zero-copy integrations with Snowflake and Databricks, but their architecture is read-oriented: their Snowflake integration uses federated queries, and their Databricks integration consumes Delta Sharing but, per their own documentation, stores the data in Treasure AI's memory for identity unification, parent segment refresh, and activations. The warehouse feeds Treasure AI, but Treasure AI remains the processing environment.
Amperity's architecture is lakehouse-native by design. Amperity Bridge provides bidirectional, zero-copy data sharing with Snowflake, Databricks, and BigQuery. Identity outputs, enriched profiles, and predictive attributes write back directly to the customer's warehouse without data leaving their environment. The customer owns the resolved data in their own infrastructure, which simplifies data sovereignty, clean room interoperability, and vendor portability. If you decide to move away from Amperity, the identity graph and enriched profiles remain in your warehouse.
Activation and real-time personalization
Treasure AI has genuine strength in real-time activation. Their web and app SDKs, real-time personalization APIs, and broad connector library are well-regarded by practitioners. Their journey canvas supports multi-step campaigns with batch and real-time triggers, capping, and exclusions. For brands whose primary use case is real-time web and app personalization, this is a legitimate differentiator.
Amperity's activation capabilities have expanded significantly. The Profile API delivers sub-second lookups for real-time personalization use cases, streaming ingest captures events with sub-second precision, and always-on activation pipelines sync audiences to paid media, email, SMS, and other channels. Predictive segmentation, powered by out-of-the-box models for customer lifetime value (CLV), churn propensity, and product affinity, lets teams build audiences based on forward-looking signals rather than historical behavior alone.
Both platforms can push audiences to channels quickly. The differentiation is in what those audiences are built from: whether the underlying profiles are accurate, complete, and correctly resolved at the person level.
Usability and time-to-value
Treasure AI's interface demos well, particularly for marketing use cases. But multiple sources, including Gartner Peer Insights reviewers, former customer success professionals, and system integrator (SI) partners, note a gap between the demo experience and production reality. Documentation is oriented toward technical users. The services team is built for technical engagement. Treasure Boxes, pitched as out-of-the-box modules, typically require SIs to write custom code for ETLs, joins, and data modeling. Standard SI-led implementations run 16 weeks or longer. Post-purchase costs can grow through seat licenses for reporting tools, profile limits, and incremental services hours.
Amperity is designed for cross-team usability from the data layer up. Marketers, analysts, and data engineers each have purpose-built interfaces for their workflows, all operating on the same trusted data foundation. The Customer Data Agent gives non-technical users a conversational path to segment creation and journey building without SQL. Implementation follows a 90-day pilot model.
Which platform fits your priorities
There's no single best CDP platform for every enterprise. The right choice depends on where your organization needs the most depth, and on how confident you are that a vendor's expanding ambition will be matched by its execution. Treasure AI's April 2026 rebrand signals a significant strategic expansion from CDP to agentic experience platform, which means buyers should evaluate whether the roadmap, pricing, and support model match the scope of what's being promised.
Consider Treasure AI if:
You need real-time web and app personalization as your primary, highest-priority use case. Your organization has deep in-house data engineering resources to configure and maintain identity rules, Treasure Boxes, and custom data models. You are looking to consolidate downstream execution tools (ESP, CEP, journey orchestration) into a single vendor. Or you prioritize breadth of marketing execution capabilities over depth of identity accuracy.
Consider Amperity if:
Identity resolution accuracy is your highest-stakes requirement, and you need matching that adapts autonomously as data sources change. You need multiple identity graphs for different business units, use cases, or risk tolerances (contextual identity). Your architecture is lakehouse-native and you want true zero-copy data sharing with Snowflake, Databricks, or BigQuery. You need non-technical teams to self-serve trusted customer data without engineering dependency. Or your AI initiatives depend on a governed, accurate data foundation that you own and control.
Talk to Amperity about building a trusted data foundation for your enterprise.