Key takeaways
The most cited reasons buyers replace an existing CDP were data readiness and quality, inability to demonstrate ROI, scalability or real-time performance issues, and AI features that did not work in practice.
Amperity has been named a Leader in the IDC MarketScape: Worldwide AI-Enabled Customer Data Platforms for B2C Users 2026 Vendor Assessment.
The CDP category has moved
For most of the last decade, a customer data platform (CDP) had one job: unify disparate customer data into a profile and send it to a marketing tool. Vendors competed on connectors, schemas, and segment builders.
The goalposts have moved. The job now is delivering trusted customer context that humans and AI can act on in the moment. Audiences still matter. They are no longer enough.
This is not just our read of the market. IDC MarketScape named Amperity a Leader in the 2026 IDC MarketScape: Worldwide AI-Enabled Customer Data Platforms for B2C Users Vendor Assessment.*
That bar has a name. Trusted customer context, not raw customer data, is the operational layer for every AI-powered customer interaction, and it's what Amperity has been building toward from the start.
Trusted customer context follows a clear order. Identity Resolution establishes who the customer actually is across every system. From that resolved identity, a single governed view is built, historically and in the moment. Intelligence, the predictions and decisions a brand acts on, is what becomes possible once that context can be trusted. Skip the foundation, and everything above it inherits the error.
What AI actually demands from customer data
"AI-enabled" is doing a lot of work in CDP marketing right now, and most of it's decorative. A chatbot bolted onto a segment builder is not an AI-enabled CDP. Neither is a copilot that writes SQL faster.
The change underneath is structural. AI is now a consumer of customer data, not only the marketing team. Models decide which message to send, which offer to surface, which next action to take, at a volume no human queue can match. That breaks the assumptions the CDP category was built on, and three new demands replace them.
Identity Resolution has to be load-bearing
When a marketer worked from a fragmented profile, the cost was a slightly worse audience. When a model works from one, the cost is wrong decisions at scale. The same customer, split across three records, gets three different offers, three different journeys, three different recommendations, all confidently wrong.
Identity Resolution is not a feature. It is the non-negotiable foundation every AI decision in the business stands on or falls from. It is also where most platforms quietly fall short. A single identity graph, or one stitched to a third-party spine, cannot resolve the same customer consistently across loyalty, service, web, and store. Without trusted identity, an AI-enabled CDP is an AI-enabled liability.
Context has to be where the decision happens
Trusted customer context is only useful where decisions get made. That used to mean a marketing cloud and a few endpoints. It now means every AI tool, model, and workflow the business runs, from service to commerce to loyalty to the store floor. A CDP that locks context inside its own platform forces the business to copy data into every model that needs it, governed differently in each place, drifting apart on every refresh. Locking context inside one vendor's cloud creates governance debt.
Decisioning has to be grounded
A model working on a hunch is the real risk, because nothing bounds how many decisions it makes. Grounded decisioning means the AI queries trusted customer context, not a guess it constructs on the fly. The decisive question is whether the foundation under the decision can be trusted and traced, on your data, in your environment, under your governance. That is exactly what closed, copy-everything architectures cannot promise.
The new buying criteria
The evaluation criteria have changed. Profile count, connector total, and audience velocity still matter, but they no longer tell you which platform is ready for AI. Four criteria do.
Trusted identity foundation. Trusted customer context begins with resolved identity: first-party, fully auditable, consistent across every touchpoint a customer has with the brand. If the identity is not trusted, nothing built on top of it's either.
Governance and accountability. Grounded decisioning is about what goes into a decision. Governance is about what comes out of it, and whether the result can be explained. When an AI makes a customer decision, someone has to answer how, with lineage, controls, and a defensible record of which data shaped which output.
AI readiness. The question is not "does this CDP have AI." It is "what is the AI standing on."
Measurable business outcomes. Profiles unified is a platform metric. Revenue lifted, customers retained, and lapsed buyers reactivated are business metrics. Buying criteria should match.
These criteria don't describe a future state. They describe what enterprise brands are demanding now, and they map directly to what Amperity has built.
Why Amperity is built for where the market is going
Each criterion maps to something Amperity built deliberately, not tacked onto a legacy platform.
Trusted customer context starts with identity Amperity built itself
The foundation is Contextual Identity, with the Stitch engine and its ten patents underneath. Multiple purpose-built graphs from first-party data, no third-party spine, fully auditable. That is the resolved identity every downstream decision depends on.
Context that lives where your data lives
Amperity is lakehouse-native, with BYOC and BYOS, so trusted customer context lives in your environment, on your storage and compute, under your controls. Amperity Bridge shares it zero-copy into the systems that need it. No rip-and-replace, no copying customer data into a vendor's walls to make it usable.
Decisioning grounded in trusted context
AmpAI is the reasoning layer over fully resolved data, with the Identity Resolution Assistant and the Customer Data Assistant. The MCP Server brings that governed context into any LLM or workflow the team uses, so decisions are grounded in trusted customer context wherever they happen.
Outcomes the business can measure
The proof is in outcomes, not architecture diagrams. By unifying 3.4M profiles on Amperity, New Look identified 24% more high-value customers and lifted ROAS 50%. The identity foundation is what made both numbers possible. Resolve the customer first, and the measurable outcomes follow.
The future belongs to trusted customer context, not more customer data
A lot of the current AI conversation is still about acquiring more signals. All useful. None of it fixes the underlying problem. Pour more data onto a fragmented identity foundation and you do not enrich a profile. You make the fragmentation louder, and the model running over it does not get smarter. It just gets more confident about the wrong thing.
See more on Amperity's recognition in the 2026 IDC MarketScape: Worldwide AI-Enabled Customer Data Platforms for B2C Users 2026 Vendor Assessment excerpt.
*IDC MarketScape: Worldwide AI-Enabled Customer Data Platforms for B2C Users 2026 Vendor Assessment, #US53952526, June 2026.