Martech AI guides: MCP servers, AI agents and AI-ready data
These guides are for marketing ops, RevOps and data people who want AI assistants to answer questions and do work on top of their real stack. They cover five areas. MCP servers: what they are, which official servers exist for HubSpot, Salesforce, GA4, Google Ads, Meta, Braze, Snowflake, BigQuery and the product analytics tools, and how to connect them with the right permissions. AI on your marketing data: connecting ChatGPT or Claude to your numbers, and the context files, semantic layer and data checks that keep the answers consistent. The AI already built into tools you pay for, such as HubSpot Breeze, Agentforce, Snowflake Cortex and Braze, and when it is enough. Agents, skills and slash commands for routine marketing ops work, with approval, evaluation and monitoring. Agentic marketing operations: agents that watch tracking and pipelines, triage data alerts, test new tracking and keep documentation current on a schedule, with the authority levels that keep a person on each fix. If you are new to the topic, start with what an MCP server is and the list of servers for marketing. If you want this set up and governed for your team, see our AI-ready data service.
Start here
- What is an MCP server? A guide for marketing and RevOps teams
- MCP servers for marketing: 11 official options and how to choose your first connection
- Why AI answers on marketing data go wrong, and the seven fixes
- Context engineering for marketing and RevOps: the six files your agent needs
- Build or Buy a Marketing AI Agent? Compare HubSpot, Agentforce and Custom Workflows
MCP servers for the marketing and data stack
Start with the guide
MCP servers for marketing: 11 official options and how to choose your first connection
Compare 11 official marketing MCP options by workflow, access and setup. Use a practical selection checklist and known-answer tests for your first connection.
Last updated
What is an MCP server? A guide for marketing and RevOps teams
MCP connects AI applications to business tools and data. Learn how it works, where it helps marketing, and how to test permissions and useful output.
Last updated
MCP vs API vs Zapier: How Marketing Ops Should Choose
Choose APIs, MCP, Zapier or scripts for marketing ops. Compare workflow fit, task costs, retries, duplicate handling and the judgment each job actually needs.
Last updated
MCP security checklist for marketing and RevOps teams
Review an MCP connection before sharing business data. Check identity, permissions, prompt injection, data exposure, write approvals and recovery.
Last updated
Internal MCP server for RevOps: what to build and how to scope it
Build an internal MCP server around specific business jobs. Learn how to scope tools, enforce permissions and test customer-data workflows end to end.
Last updated
HubSpot MCP server: remote setup for Claude, ChatGPT and Claude Code
Connect HubSpot's remote MCP server to Claude, ChatGPT or Claude Code, see which CRM objects it can read and update, and test access before any write.
Last updated
Salesforce MCP server: hosted setup, Claudeforce and a first CRM workflow
Set up Salesforce hosted MCP servers in Claude, start with read-only SObject access, plan for Flex Credit billing and move off the retiring DX MCP server.
Last updated
Mixpanel MCP server: connect Claude and get product answers you can check
Connect Mixpanel MCP to Claude, choose the right region, check permissions and build product analyses with clear definitions, verified counts and useful decisions.
Last updated
Amplitude MCP server: setup, permissions and useful AI workflows
Connect Amplitude MCP to Claude, check read and write access, use progressive discovery and build reliable analytics, tracking-plan and dashboard workflows.
Last updated
PostHog MCP for marketers: funnels, UTMs and campaign questions
Connect PostHog MCP, scope analysis access and investigate funnels, UTM gaps and campaign costs with practical prompts, attribution checks and clear AI billing.
Last updated
Google Analytics MCP server: setup and reporting checks for GA4
Use Google's Analytics MCP server for repeatable GA4 reporting. Follow the setup, validate metrics and turn vague prompts into evidence-backed investigations.
Last updated
Google Ads MCP server: setup, access changes and useful reporting workflows
Set up Google's read-only Ads MCP server using current Cloud project access. Includes reporting prompts, a GAQL example and campaign validation checks.
Last updated
Meta Ads MCP server: connect your account and keep campaign changes under control
Connect Meta Ads through its official MCP server. Understand access, reporting limits and the checks needed before an AI workflow changes live campaigns.
Last updated
Google Tag Manager MCP: Audit Containers Without Guessing What Fires
Use Google Tag Manager MCP to inspect tags, identify duplicate conversion paths and verify fixes with runtime evidence before publishing a container.
Last updated
Braze MCP server: connect Claude or ChatGPT, check access and review campaigns safely
Connect the Braze MCP server to Claude, ChatGPT or Cursor with OAuth, see which tools read, write or need a beta, and run a campaign review you can verify.
Last updated
Klaviyo MCP Server: Connect Safely and Get Reports That Reconcile
Configure Klaviyo MCP access, confirm the right account and reconcile flow reporting with the dashboard before acting on AI recommendations.
Last updated
Customer.io and Intercom MCP: Choose the Workflow Before the Connector
Choose between lifecycle MCP, support-data MCP and Fin connectors. Design a useful cross-system workflow with correct identity, evidence and message rules.
Last updated
Snowflake MCP server: give marketing answers without handing over the warehouse
Snowflake's managed MCP server lets Claude, ChatGPT and Cursor query your account through tools you govern. The read-only setup, step by step.
Last updated
BigQuery MCP server: Google's remote service, MCP Toolbox and the controls you need
Give an AI assistant controlled BigQuery access. Compare connection routes, define trusted datasets and test query correctness, permissions and costs.
Last updated
MCP for GTM Engineering: Clay, Apollo and Instantly in Claude
Use MCP in GTM workflows with clear access, identity and review controls. Compare current tools, test a bounded workflow and decide when an API fits better.
Last updated
AI on marketing data: ChatGPT connections, context and semantic layers
Start with the guide
ChatGPT for Marketing Analytics: Connect Your Data and Verify the Answer
Connect HubSpot, GA4, Google Ads and warehouse data to ChatGPT. Use practical prompts and a worked campaign example to verify metrics, coverage and conclusions.
Last updated
Why AI answers on marketing data go wrong, and the seven fixes
Claude and ChatGPT answer from whatever your stack gives them. The seven checks that make AI answers on marketing data right, in the order to fix them.
Last updated
Context engineering for marketing and RevOps: the six files your agent needs
Give marketing AI the definitions, evidence and working rules it needs. Includes a context template, ownership model and tests for trustworthy answers.
Last updated
Customer Context Layer: Making Golden Context Useful for a Small SaaS
Make Golden Context useful with clear customer facts, company rules, ownership and freshness. Build one reliable AI workflow before buying another platform.
Last updated
Semantic Layers for AI: dbt, Snowflake Semantic Views and LookML
Compare dbt, Snowflake semantic views and LookML for AI. Define metric grain, time rules and access, then test whether agents use the intended calculation.
Last updated
AI inside the tools you already pay for
Start with the guide
Build or Buy a Marketing AI Agent? Compare HubSpot, Agentforce and Custom Workflows
Compare HubSpot Agent Builder, Agentforce and custom marketing agents using one acceptance brief, realistic ownership costs and a practical handover test.
Last updated
HubSpot AI in 2026: Breeze, Agent Hub, credits and what to implement first
Understand HubSpot Agent Hub, Breeze, AI agents and credit costs. Use worked budgets and acceptance tests to choose a useful first implementation.
Last updated
Salesforce Agentforce for Small Teams: Scope, Cost and Ownership
Decide whether included Salesforce AI or a custom Agentforce workflow fits your team. Scope the job, estimate usage and test the ownership and access model.
Last updated
Snowflake AI explained: Cortex, CoWork, CoCo and the data work underneath
Understand Snowflake CoWork, CoCo and Cortex services. See how trusted models, semantic definitions and access controls support useful marketing AI.
Last updated
BigQuery AI: Gemini, Data Agents and AI Functions for Marketing Data
Choose BigQuery AI tools for marketing data. Understand Gemini costs, data-agent setup, AI functions, permissions and how to verify useful outputs.
Last updated
Snowflake CoWork, BigQuery or Looker: Which AI Analyst Fits Your Team?
Choose an AI analyst around trusted models, permissions and total cost. Compare Snowflake CoWork, BigQuery and Looker with a practical evaluation plan.
Last updated
GA4 AI: Ask Advisor, AI Assistant Traffic and Answers You Can Check
Use GA4 Ask Advisor, understand the AI Assistant channel, and verify AI analysis with clear metrics, practical prompts and a worked demo-request example.
Last updated
Braze AI: Choose a Specific Lifecycle Decision to Improve
Choose the right Braze AI feature for timing, churn or message selection. Define inputs, controls and business outcomes before expanding your lifecycle pilot.
Last updated
AI Decisioning: Compare Hightouch, Braze and RudderStack by Workflow
Compare where Hightouch, Braze and RudderStack fit in a decisioning workflow. Evaluate context, execution, controls, freshness and complete operating costs.
Last updated
AI agents, skills and slash commands for marketing ops
Start with the guide
AI agents for marketing ops: what to automate, what to review and where to start
Choose marketing AI agents by the work they can reliably finish. Compare workflow types, estimate review costs and design a pilot with measurable outcomes.
Last updated
AI for RevOps: Useful Workflows, Reliable Data and a Pilot That Proves Value
Choose useful AI workflows for RevOps, compare native tools with custom agents, and test evidence quality, missed issues, forecast accuracy and real effort.
Last updated
Claude Code for marketing: a practical first workflow for non-engineers
Use Claude Code for practical marketing operations work. Start with bounded files, clear instructions and reviewable outputs before connecting live systems.
Last updated
Claude skills for marketing: how they work and 20 workflows to copy
Turn recurring marketing work into reusable Claude skills. Includes skill structure, example instructions, testing criteria and safe update practices.
Last updated
15 Claude Code Slash Commands for Marketing Ops: Setup and Recipes
Build useful Claude Code slash commands with 15 marketing ops recipes, a complete pipeline-review skill, worked examples and clear permission boundaries.
Last updated
AI Agent Approval Workflows: Control Business Actions Clearly
Design AI agent approvals around exact actions, current record state, shared limits and reliable retries. Includes an approval record and practical tests.
Last updated
How to Evaluate a Marketing AI Agent Before Release
Build realistic marketing AI agent evaluations with reference cases, critical constraints, a baseline and review-cost measures. Includes worked examples.
Last updated
AI Agent Observability: Trace Answers, Actions and Failures
Trace AI agent requests through evidence, tools and verified outcomes. Diagnose failures, retries, review costs and sampling bias with a practical event contract.
Last updated
Want a second pair of eyes on your stack?
Bring your current setup and the decision you need to make. We can help you scope the next useful improvement.
Book a Free Audit
A free audit of your stack and what it needs. No pitch.
Thirty minutes, and you leave with a plan
We look at what you have, what it is costing you and what to do about it first.
- 30 minutes
- Video call, link on booking
- No obligation
What we'll cover
- Your current stack and where it hurts
- Quick wins we can ship first
- An integration roadmap
- Timeline, scope and next steps
Calendar not loading? Open it in a new tab or email hello@martechdevs.com.