Platform integration & agentic orchestration
Run Salesforce, HubSpot, Snowflake or Google Ads? See where mAI fits.
Find your platforms, see what changes, and keep your team in charge of every material decision.
Keep the systems your teams already use. mAI is designed to read approved signals, recommend next actions and draft work across your stack. Any execution depends on the access and approval boundaries you define.
How to read this page. Platform labels show integration status. Every platform here is currently “Designed to integrate”: an API, export or connector path we validate with you in discovery, not a certified connector or a partnership. Workflows are proposed designs; your team sets the access and approval limits.
Start with your role and the systems you run
Choose your role or stack to see an operating pattern, the decisions your team keeps and how to measure value.
1 · I am a…
2 · My stack looks like…
Choose your role and stack to focus the page.
CMO / Head of Marketing
Start with one business outcome and a documented baseline. Review where decisions need your approval and how operating effort translates into marketing value.
Marketing Ops / Martech Architect
Inspect your systems, available APIs or exports, data flow and permissions. Scope read access separately from any write action so integration work is explicit.
Demand Gen / Paid Media Lead
See how a performance signal becomes a proposed change, a human decision and a measured test. Attribution limits and budget boundaries stay visible.
Data / Analytics Lead
Begin with approved aggregate views, data quality and a measurement contract. Every interpretation should state confidence and what the evidence cannot establish.
CIO / Security / Privacy / Compliance
Review the proposed deployment environment, access scope, retention, model-provider terms and approval records before connecting systems.
Agency / Consultancy Lead
Keep client context and approvals scoped to each engagement. Define who operates mAI, who reviews outputs and how repeatable methods support your team.
If your stack looks like this
Your systems remain the systems of record and execution. Start with one workflow, agreed data access and a documented baseline.
CRM-centric B2B
Core systems
Often alongside
Designed to integrate
How mAI works in this stack
- Reads approved pipeline stages, campaign membership and lead outcomes, aggregated first.
- Drafts segment and routing recommendations; flags stalled stages.
- Prepares campaign briefs and reporting narratives tied to pipeline quality.
Example workflow: Lead quality and routing →
What stays with your team: Audience definitions, scoring changes and outreach sends.
Measured by: Pipeline quality · CPL/CPA.
Proposed change: move from reviewing separate lead reports to reviewing one evidence-backed routing recommendation.
HubSpot-centric growth team
Core systems
Often alongside
Designed to integrate
How mAI works in this stack
- Reads campaign reports, attribution and content signals within agreed access boundaries.
- Drafts content briefs, landing-page tests and nurture logic.
- Prepares weekly optimization memos with evidence and uncertainty.
Example workflow: Content to citation →
What stays with your team: Publishing and new messaging pillars.
Measured by: Conversion rate · Organic and AI-citation visibility.
Proposed change: connect campaign and content findings in a weekly memo your team can review.
Data-first enterprise
Core systems
Often alongside
Designed to integrate
How mAI works in this stack
- Reads modeled, aggregated marketing data from approved warehouse views.
- Interprets what changed, likely causes, confidence and attribution limits.
- Drafts recommendations as documents, tickets or tables; writing requires explicit permission.
Example workflow: Measurement narrative →
What stays with your team: Data access scope and model governance.
Measured by: Reporting latency · Share of recommendations with documented evidence and confidence.
Proposed change: turn approved aggregate warehouse views into an analyst-reviewed decision narrative.
Adobe / enterprise suite
Core systems
Often alongside
Designed to integrate
How mAI works in this stack
- Reads approved analytics and campaign performance reports.
- Drafts briefs and asset variants using approved brand context.
- Routes content through the agreed editorial and approval process.
Example workflow: Content to citation →
What stays with your team: Brand-critical creative and regulated claims.
Measured by: Time-to-launch · Operating hours per approved asset.
Proposed change: route source-backed briefs and variants through your existing editorial process.
Ecommerce and lifecycle
Core systems
Often alongside
Designed to integrate
How mAI works in this stack
- Reads approved campaign, creative and revenue signals.
- Recommends test designs with holdouts where feasible.
- Drafts lifecycle variants and creative briefs for review.
Example workflow: Paid media optimization →
What stays with your team: Budgets, discount rules, offers and audiences.
Measured by: ROAS · Conversion rate.
Proposed change: review a test proposal alongside revenue signals, offer rules and attribution limits.
Regulated organization
Core systems
Often alongside
Designed to integrate
How mAI works in this stack
- Scopes Enterprise mAI (Hosted) within the environment your organization approves.
- Starts read-only and drafts against approved knowledge and claims.
- Routes outputs to named reviewers with an approval record before activation.
Example workflow: Governed content approval →
What stays with your team: All claims, audiences, sends and access decisions.
Measured by: Time-to-launch · Unapproved claims released (target: zero).
Proposed change: review a draft alongside its claim sources and named approval path.
Find your platforms in the integration map
Find the systems in your stack and the workflow each category can support.
Designed to integrate applies to every platform below. Listing does not establish certification or partnership.
CRM and sales
Reads approved pipeline and campaign outcomes; recommends segments and routing.
- Salesforce
- Designed to integrate
- HubSpot
- Designed to integrate
- Microsoft Dynamics 365
- Designed to integrate
Data warehouse, lakehouse and CDP
Reads approved aggregate data; interprets marketing measurement.
- Snowflake
- Designed to integrate
- Google BigQuery
- Designed to integrate
- Databricks
- Designed to integrate
- Segment
- Designed to integrate
- mParticle
- Designed to integrate
Marketing automation and email
Drafts journeys, nurture logic and variants for review.
- Marketo Engage
- Designed to integrate
- Adobe Experience Platform
- Designed to integrate
- Braze
- Designed to integrate
- Klaviyo
- Designed to integrate
- HubSpot
- Designed to integrate
- Account Engagement
- Designed to integrate
Paid media
Flags anomalies and drafts change requests inside agreed boundaries.
- Google Ads
- Designed to integrate
- Meta Ads
- Designed to integrate
- LinkedIn Ads
- Designed to integrate
- Google DV360
- Designed to integrate
- The Trade Desk
- Designed to integrate
Analytics and tagging
Interprets results; proposes measurement plans and test designs.
- GA4
- Designed to integrate
- Google Tag Manager
- Designed to integrate
- Adobe Analytics
- Designed to integrate
Content, CMS and SEO/GEO
Produces briefs and drafts; interprets organic and AI-citation visibility signals where available.
- WordPress
- Designed to integrate
- Adobe Experience Manager
- Designed to integrate
- Contentful
- Designed to integrate
Work management and collaboration
Routes approval requests, summaries and tasks through agreed delivery paths.
- Slack
- Designed to integrate
- Notion
- Designed to integrate
- ClickUp
- Designed to integrate
- Asana
- Designed to integrate
- Google Workspace
- Designed to integrate
- Workfront
- Designed to integrate
AI model layer
Scopes model-provider options against task, hosting and data requirements.
- OpenAI
- Designed to integrate
- Anthropic
- Designed to integrate
- Google Gemini
- Designed to integrate
Governance and identity
Capabilities to scope: consent rules, identity and access controls, and approval records. These are requirements for your environment.
- Consent management
- SSO / identity
- Audit logging
Platform names and any logos shown are trademarks of their respective owners and are used for identification only. Listing does not imply endorsement, partnership or certification.
Don’t see your platform? We can evaluate an API, export or connector path against your requirements. Tell us your stack →
From a signal to a reviewed decision
Agentic orchestration coordinates analysis, drafts, review and measurement across an agreed workflow. The AI Marketing OS runs Plan → Execute → Measure → Optimize; the AI Marketing Brain interprets context and signals, under Human Command.
Lead quality and routing
- TriggerA weekly review or a new or stalled lead signal.
- Agent actionAnalyze aggregate lead outcomes and propose segment or routing changes, with evidence and reasoning.
- Human gateMarketing and sales leads approve audiences and scoring changes. Approval required
- MeasurementCompare pipeline quality and segment conversion with the documented baseline.
- OptimizationPrepare the next change request with the result, uncertainty and risk.
Paid media optimization
- TriggerAn anomaly in CPA, ROAS or spend pacing.
- Agent actionDiagnose likely causes and draft a bid, budget, creative or audience change request.
- Human gateAct only within client-approved limits; route audience and brand-critical decisions to a person.
- MeasurementUse a test or holdout where feasible; last-click attribution alone does not justify a major shift.
- OptimizationFeed the documented result and attribution limits into the next plan.
Budget thresholds and approval rules
Client-approved limits govern every action. Proposed defaults: cumulative monthly budget changes up to 5% stay within the pre-approved range and are recorded; above 5% through 10% require a named reviewer; above 10% require named approval. Audience and brand-critical decisions always retain their human gate.
Content to citation
- TriggerA calendar item or a gap in search and AI-answer visibility.
- Agent actionUse approved knowledge in the AI Marketing Brain to build a brief and first draft; check claims against their sources.
- Human gateAn editor reviews the draft. A named person approves claims and brand-critical copy. Review required
- MeasurementTrack organic visibility, AI citations, engagement and qualified conversions.
- OptimizationRefresh or expand based on documented visibility and conversion signals; citation is never guaranteed.
Measurement narrative
- TriggerThe agreed reporting cycle.
- Agent actionInterpret aggregated results: what changed, by how much, likely causes, confidence and what not to conclude.
- Human gateAn analyst reviews the interpretation before circulation. Review required
- MeasurementTrack reporting latency and time from a reviewed finding to a documented decision.
- OptimizationRecommend the next test and update measurement assumptions.
Governed content approval
- TriggerA customer-facing asset requiring review.
- Agent actionDraft from approved sources, include AI-assisted disclosure where required, and classify claims as factual, performance, comparative, regulatory, forecast, testimonial or opinion.
- Human gateRoute to marketing, then compliance and the named approver. Retain the approval record. Compliance check
- MeasurementTrack time-to-launch, revision count and unapproved claims released, with a target of zero.
- OptimizationUpdate the approved claims library and review rules from documented decisions.
Choose who operates mAI and how far it can act
Deployment follows your operating constraints. Read access, model choices and write permissions are scoped independently for each workflow and system.
Managed mAI
marktgAI coordinates agreed AI Marketing OS and AI Marketing Brain workflows with your team. Your business context, review responsibilities and performance baselines shape the operation.
Enterprise mAI (Hosted)
Enterprise mAI (Hosted) scopes a Custom AI Marketing Model around approved knowledge, permissions and your chosen environment. Provider, hosting and security requirements are validated before deployment.
Access model
Read-only first. Write access is off by default and granted per system and action type. Prefer aggregate information and define the minimum access needed for the task.
One context, one operating cycle
Approved company context and performance signals inform the AI Marketing Brain. The AI Marketing OS coordinates the agreed workflow across your systems. Measured outcomes feed back into context and future recommendations. Human Command governs the full cycle.
- 1 · Read
- 2 · Recommend
- 3 · Draft
- 4 · Act with approval
You choose how far each workflow goes. Hybrid delivery can combine operating support with client-defined controls.
Make Human Command explicit
Human-led intelligence. AI-powered precision. Define permissions and review responsibilities before activation so each material decision has an accountable owner.
- People approve strategy, audiences, material budgets, new channels, regulated claims, brand-critical creative and material confidential-data use.
- Use review tags where applicable: Approval required, Review required and Compliance check.
- Classify claims and retain sources, uncertainty, review decisions and approval records.
- Review consent, suppression, retention and AI-assisted disclosure requirements for the agreed workflow.
- Your privacy, security and compliance teams assess the deployment against applicable requirements before access is granted.
Integration status does not establish security certification or regulatory compliance. Review the evidence and governance methodology →
Measure Productivity + Precision (P²)
Agree on a baseline, a measurement window and a decision rule before connecting a workflow. Evaluate the value delivered after integration, operating and review costs.
- Productivity: Measure operating hours per approved output, reporting latency and time-to-launch. Count time saved only when it is removed or reassigned to useful work.
- Precision: Compare the relevant business KPI, such as pipeline quality, conversion rate or ROAS, while recording confidence, attribution limits and approval quality.
- ROI: Compare verified capacity savings and incremental business value with total workflow costs. Use a test or holdout where feasible; treat low-volume findings as directional.
Use the reviewed result to refine the workflow, adjust its access boundaries or decide whether to expand it.
Review the evidence in its operating context
Explore measured results from real client deployments, names withheld. Review the operating context and limits; these cases do not verify individual connectors in this map.
Review real client case studies →
For product integration, licensing or co-development, explore mAI for Platforms & Product Teams →
Questions about integration, access and control
Scope platform fit, access and measurement before activation.
Which platforms does mAI integrate with?
mAI is designed to integrate with the systems listed in the map through a scoped API, export or connector path. Discovery checks your product edition, permissions, data needs and required actions before confirming fit.
Does mAI need access to our customer data?
Many reporting and recommendation tasks can start with aggregate data rather than individual customer records. The required access depends on the workflow. Your team defines approved fields, sources, retention and purposes before connection. Any need for personal or confidential information is reviewed separately, with access limited to the agreed task and environment.
Can mAI make changes in our systems on its own?
Write access is off by default in the proposed access model. A workflow can stop at reading, recommending or drafting. Execution requires explicit permission for the system and action type, plus agreed approval rules. Strategy, audiences, material budgets, regulated claims and brand-critical decisions remain under Human Command, even when routine actions have pre-approved operating limits.
What is the difference between Managed and Hosted?
Managed mAI is operated with marktgAI and your team around agreed workflows and controls. Hosted mAI Custom AI Marketing Models follow a client-controlled deployment approach, presented on this site through Enterprise mAI. Discovery defines the actual environment, responsibilities, model-provider constraints and data boundaries. Hybrid arrangements can combine operating support with the control your organization requires.
How do approvals and budget thresholds work?
Your organization approves the rules before activation. The paid-media workflow shows proposed budget defaults, measured cumulatively across a month; they are not blanket permission to change spend. Named reviewers, approvers, records and escalation paths belong in the workflow. Audience changes, regulated claims and brand-critical decisions retain their own approval requirements regardless of the budget percentage.
Which AI models do mAI models run on?
The mAI Framework is model-agnostic by design. OpenAI, Anthropic and Google Gemini are provider options to evaluate against the task and approved environment. We confirm the model, region, retention terms and hosting for each deployment.
How long does an integration take?
Timing depends on the workflow, available API or export path, access approvals, data quality and validation requirements. Start with one system and one measurable operating problem. A stack-fit conversation identifies dependencies and produces a scoped plan before any timeline commitment. Regulated review or write access can add steps that need to be scheduled explicitly.
How do you keep our data and brand safe?
Start by defining the controls: approved sources, minimum access, model-provider constraints, retention, consent and suppression rules, named review gates and approval records. Brand and claim rules guide drafts before activation. Your security, privacy and compliance reviewers assess the deployment against applicable requirements. This page describes the governance approach; it does not claim a certification or guarantee compliance.
Does this work alongside our agency or internal team?
Yes, the operating model is designed around existing teams and systems. Your team or agency can retain strategy, creative judgment, platform ownership and approval authority while mAI supports agreed analysis, drafts and reporting. Responsibilities are defined per workflow and client context. Any white-label, licensing or partner arrangement requires a separate scope rather than being implied by this page.
How do you measure whether it is working?
Define a baseline, target, measurement window and decision rule before activation. Productivity measures can include operating hours and reporting latency; Precision uses the relevant business KPI, such as conversion rate or pipeline quality. Compare the value of changes with integration and operating costs. Use tests or holdouts where feasible, state attribution limits and treat small samples as directional.
Tell us your stack
We’ll map where mAI fits and where a human stays in charge. Bring one operating problem, your systems, access constraints and the outcome you want to measure.
Marketing. Run as a System. · Human Command. AI-Powered Precision.
