Verity over MCP
Your marketing data platform, run from Claude.
Ask questions, build dashboards and reports, create audiences and draft journeys and onsite campaigns, all from the assistant you already use, through 100+ MCP tools. Your data stays in a warehouse you own, and you approve what reaches your customers.
Analytics · Dashboards · Customer data platform · Data management
Verity connected
Example conversation. The tools are real; the numbers are illustrative.
What your agent can do
Ask in plain language. The agent picks the tools, works on your real customer data and shows you what it did.
Analyse
“What was our blended ROAS last month, per channel?”
“Which campaigns brought new customers, not returning ones?”
“Did anything unusual happen to revenue this week?”
Dashboards and reports
“Build a weekly performance dashboard for the board.”
“Add a revenue per segment table to this dashboard.”
“Set up the monthly report and send it to me for review.”
Audiences and activation
“Create a segment of first time buyers from paid social.”
“Draft a win back journey for lapsed customers, with a holdout.”
“Send this audience to Meta Ads and Google Ads.”
Data and models
“Connect our BigQuery dataset and propose the semantic layer.”
“Which dashboards would break if we change this dbt model?”
“Set our targets and budgets for next quarter.”
The agent drafts, you decide
Agents that act, without acting alone
Every tool acts as the person who connected it and never more. It sees the workspace you see and can do what your role allows. Reading is instant. Changes inside Verity, such as a draft segment or a dashboard, need a read and write connection.
Anything with an effect outside Verity works differently. Syncing an audience, publishing a journey, switching an onsite campaign on or setting a schedule comes back as a review link. A person opens it, checks the change and approves it in Verity. That is how you let an agent do the work and still know exactly what reaches your customers.
66
read tools
Questions, dashboards, segments, journeys, results. Nothing changes.
45
write tools
Drafts and dashboards inside Verity, on a read and write connection.
12
tools that hand over to you
Syncs, journeys, campaigns and schedules wait for a person.
One server, the whole platform
Verity does not expose a query window over MCP. It exposes the whole platform: analysis, dashboards and reports, segments and profiles, activation, and data management. The full reference, with every tool and its access level, is public.
Open the tool reference11 tools
17 tools
23 tools
49 tools
18 tools
5 tools
Agent ready
Built for how agents actually work
A long list of tools is not enough. Verity gives the agent what it needs at the moment it needs it, and shows you the result where you are working.
Guidance on demand
The agent starts with a short index of rules that always apply, and fetches a detailed chapter only when the job needs it. Nothing gets cut off, and an agent changing one number does not read the whole manual.
The help centre, when it needs it
The agent searches the product documentation and reads the one article that answers its question, instead of guessing how a feature works.
Charts and reviews inside Claude
Answers come with an interactive Verity chart you can add to a dashboard in one click. A proposed data model opens as a review screen with the personal data columns first.
Never out of date
Every answer carries the version of the tools it came from, so the agent notices when Verity changed during a long conversation and tells you to start a fresh one.
Connect in a minute
In ChatGPT, claude.ai or Claude Desktop, add a custom connector in the settings and paste the server address you find in Verity under Settings, MCP Connections. Sign in with Google, pick the account and workspace, and choose read only or read and write.
Works with ChatGPT, claude.ai, Claude Desktop, Claude Code, Codex on your own computer and Cursor. MCP and write access are switched on per account by Verity.
Questions
What can an AI agent do in Verity over MCP?
Almost everything a person can do in Verity. Its 100+ tools cover analysis on your semantic layer, dashboards and reports, segments and activation, and data sources and models. Every tool and its access level is public at ask-verity.com/docs/mcp, so you can check what an agent can and cannot do before you connect it.
How does an agent find its way through 100+ tools?
Verity hands over its knowledge in layers. The agent reads a short index of rules and fetches a detailed chapter, or a help article, only when the job needs it. That keeps the context small and the guidance complete, also in assistants that cut long instructions short.
Can Claude build dashboards in Verity?
Yes. An agent can create a dashboard, lay it out from a sentence, add and fill blocks, set filters and format numbers, and submit it for review. Publishing a dashboard to the team is done by a reviewer in Verity.
Can an AI agent create and activate customer segments in Verity?
Yes. An agent can look up the fields, count who matches, and create or change a segment as a draft. Sending that audience to a destination such as Meta Ads, Google Ads, Brevo, Klaviyo, ActiveCampaign or your Shopify store is proposed by the agent and switched on by a person in Verity, through a review link.
Does Verity work with Claude and ChatGPT?
Yes. Verity works with ChatGPT, claude.ai, Claude Desktop, Claude Code, Codex on your own computer and Cursor. You add it as a connector, sign in with Google and choose what the assistant may do.
Can an agent publish journeys and website personalisation?
Yes, up to the last click. An agent can build a segment, draft a journey, set up a storefront campaign per audience, design an onsite message and propose the audience sync or the change to recommendations. Going live is one approval by a person in Verity, through the review link the agent hands you. An agent can also pause a live journey on its own.
Is my customer data safe when an agent uses it?
The agent acts as the person who connected it and never more: same workspace, same rights. You choose read only or read and write when you connect. Segment and journey numbers come back as counts, never as a list of people. Every call is logged under Settings, MCP Connections, where you can revoke a connection.
Where does the data live?
In your own BigQuery, in your own Google Cloud project. Bring your own, or Verity provisions one in your name. Segments are defined on the customer profiles in that warehouse, so there is no separate copy of your customers to keep in sync, and you keep everything if you leave.
See your agent run Verity
Book a demo and watch Claude answer, build a dashboard and activate an audience on your own data.