Meiro AI · Agentic marketing on your CDP
Meiro AI agents that build the CDP
Piper, the AI agent in Meiro, drafts sources, audiences, journeys and campaigns from a plain-language request. Nothing is applied until someone on your team approves it.
AI agents for data teams
The agent that sets up the data layer
Sources, event types, identity rules and attributes described in natural language and drafted as a proposal.
Setup without the implementation project
Ask Piper how tracking, identity and attributes work, before any setup starts.
Cleaner data foundations
Piper surfaces what the data holds and proposes new customer attributes for richer profiles.
Plain language into SQL
Ask a business question in plain language and Piper proposes the attribute logic for review.
AI agents for marketing teams
The agent that builds audiences, journeys and campaigns
Piper drafts content and recommends use cases your data is ready for.
Audience segmentation, no SQL
Describe the customers you want; Piper proposes the audience and exclusions.
Journey orchestration
Piper drafts journey content such as emails and banners for each segment.
Use-case recommendations
Ask what to run; Piper checks your data and recommends use cases with readiness and size.
AI Decision in journeys
Choose a journey branch per customer from selected attributes and events, with a confidence threshold and a fallback.
One agent across CDP and marketing automation
The data engineer’s agent and the marketer’s agent
Piper works from your configuration and drafts the work. What it drafts depends on where you point it.
Meiro Pipes · CDI
The data engineer’s agent
Collection, identity and routing.
- Proposes sources, destinations, pipes and transforms
- Proposes warehouse and profile syncs and reverse ETL
- Every change is a proposal you approve
Meiro Engage · CDP
The marketer’s agent
Segmentation, content and activation.
- Proposes audiences and lookalikes
- Drafts journeys, campaigns, Liquid and HTML content
- Nothing launches or sends until you approve and activate it
Piper and MCP prepare proposals. Changes are applied only after someone approves them.
Why Meiro AI
What makes Meiro AI different
Local AI CDP
For deployments that need the model inside your own environment, see Local AI CDP.
Local AI, explained →
Bring your own model or key
Use Google Gemini, OpenAI or Anthropic, including Anthropic through Google Cloud Vertex AI with an EU, US or global region, or an OpenAI-compatible endpoint.
Model Context Protocol (MCP)
Use Meiro from MCP-capable AI tools to read data, open profile reports and prepare proposals that someone approves in Meiro.
Quality Data Foundations
An agent is only as good as the profiles underneath it, and Meiro's ETL and identity resolution produce them.
Customer stories
Customer success stories
Meiro AI, answered
Frequently Asked Questions
How is sensitive data protected when the AI uses a cloud-based LLM?
There is a bring-your-own-key or bring-your-own-model option, allowing you to use your own subscription to LLMs and maintain full control over your data. We currently support OpenAI (all current models available over API), Gemini (2.5 Pro and Flash), Anthropic (Sonnet 4 and Opus 4). All other models matching or exceeding capabilities of mentioned models that are available on providers like AWS Bedrock, Azure AI Foundry, OpenRouter.
It is also possible to connect to local models running on vLLM or ollama over OpenAI compatible API.
Additionally, when we use our models (like Gemini), we have settings in place to ensure that your data is not logged, used for training, or retained by the LLM provider after processing. This ensures confidentiality and compliance.
What happens to the data in an on-premise deployment?
We fit neatly into whichever setup best protects your data and suits your workflows. If you already rely on a trusted cloud AI provider like OpenAI, Google, or Microsoft, we connect to it through your own subscription. If you need complete data residency, we can instead deploy and manage a fine-tuned, open-source language model entirely inside your infrastructure, giving you modern AI power without any data crossing your firewall.
What if the AI gives inconsistent or inaccurate answers?
Piper proposes, a person approves. Every configuration change, audience definition and campaign arrives as a proposal that can be read in full and rejected before anything runs, so an inaccurate answer costs a review rather than a production change.
When a proposal is wrong, Piper can explain the reasoning behind it, and it can be instructed differently for the next one. Some capabilities, such as SMS and mobile push, need an administrator to enable them in instance settings before Piper can act on them at all.
What makes Piper useful inside Meiro?
Piper works on your own customer data and on the Business context you set in Settings, such as objectives, tone of voice and average order value. You choose the model provider, and every change is a proposal you approve.
