At a glance
One master framework. Method, evidence rules and governance, documented once.
One company baseline. Your ICP, claims, KPIs, roles and risk limits.
Tested before release. Safety behaviors are tested repeatedly, then retested when the framework changes.
People approve what matters. Strategy, audiences, budgets and claims stay with your team.
The build route: baseline one company → build from the master framework → test before release → govern in operation → measure against your baseline → learn under human command.
01 · One framework, one company
Custom where it should be, consistent everywhere else
mAI is built in two parts. A master framework holds the operating method, the evidence rules and the governance controls. A company baseline holds what is specific to you. Every field is labeled, so it is always clear what is fixed, what is a default you can change, and what belongs to your company.
02 · From company URL to a governed model
Part automated, part approved by your people
Onboarding starts from your public footprint and ends inside your approval workflow. Automation does the first draft. Your team decides what the model is allowed to treat as fact.
What you provide
Company context, approved claims, KPIs and baselines, named approvers, and the risk limits you want enforced.
What marktgAI does
Researches, drafts, baselines and tests the model, and documents every decision so it can be inspected later.
What stays with your people
Strategy, audiences, budgets above threshold, brand-critical creative and regulated claims. Always.
03 · Two layers, one loop, human command
Every model runs the same operating loop
Whatever the company, the model plans, executes, measures and optimizes in that order, and never separates strategy from execution or measurement.
AI Marketing OS
How the work runs: Plan → Execute → Measure → Optimize, with explicit inputs, outputs and handoffs.
AI Marketing Brain
What the model knows: approved company knowledge, and how confident it is in each piece of it.
Human Command
Who approves what, before anything material happens. People lead. The model accelerates.
04 · Rules the model cannot outrank
A fixed order of precedence, and four plain-language flags
When instructions conflict, higher outranks lower. A lower rule can never override a higher one, and the task at hand sits last.
Approval required
The action cannot proceed without a named person’s approval.
Review required
A person checks the work before it is used.
Compliance check required
A compliance or legal review applies. The model flags it and pauses.
Unverified
A fact the model could not support. It says so instead of filling the gap.
05 · Governed decisions
Recommendations come with evidence, risk and an approver
Each material recommendation follows the same record: context, signal, hypothesis, action, confidence and risk, then measurement and learning. Risk decides what the model may do alone.
Low risk: proceed. Internal summaries, first drafts, ideation, alerts.
Medium risk: recommend, then a person reviews. New campaign variants, audience ideas, moderate budget shifts.
High risk: flag and pause. Strategy shifts, regulated claims, major budget moves, sensitive segments.
Illustrative decision record
Constructed to show the mechanism. Not a client record. Thresholds are examples.
Context: a cost-per-lead target is set. Budget moves up to 5% proceed, 5 to 10% are reviewed, above 10% need explicit approval.
Signal: cost per lead has drifted above baseline for two weeks. Attribution is platform-reported.
Hypothesis: shifting 8% of budget to high-intent terms lowers cost per lead. Test against a holdout for 14 days.
Confidence and risk: platform-reported evidence, moderate confidence. 8% exceeds the automatic threshold, so a named approver confirms in writing.
Learning: the result is compared with the holdout and logged. The playbook changes only if the result repeats.
06 · Tested before release
Safety behaviors are tested repeatedly, not once
The framework ships with a structured test matrix across every module. Tests are classed by what is at stake, and the classes carry different bars.
Safety-class tests
Approvals, claims, privacy, budget limits and evidence handling. These are repeated across several runs before a model is released.
Quality-class tests
Structure, clarity, use of brand terminology and completeness of deliverables. These are reviewed and improved continuously.
When the framework changes, the affected tests run again. After launch, live outputs are sampled for human review. Test logs are shared during onboarding or a formal review, not published.
07 · Portable by design
Model-agnostic, so your knowledge stays yours
mAI is an intelligence and operating layer. It orchestrates and governs leading third-party AI models. marktgAI does not develop or train foundation models.
What stays with you when the model changes:
- Approved company context and decisions
- Operating logic and governance controls
- Measurement methods and decision memory
08 · Measured on Productivity + Precision (P²)
A baseline first, then a direction of travel
Every model reports against a baseline agreed before launch. P² is a direction, not a guarantee. We do not promise a specific result, and any estimate is labeled as a hypothesis with what would confirm or disprove it.
Productivity
The human time needed to plan, produce, analyze and revise. Measured through time-to-launch, hours saved, reporting latency and automation coverage.
Precision
The quality of decisions and outcomes. Measured through engagement, conversion quality, CPL or CPA, ROAS and pipeline quality, with confidence stated for each.
Questions buyers ask
Does marktgAI train its own AI model?
No. mAI orchestrates and governs third-party frontier models. The value is the marketing-specific context, decision logic and governance around them.
Who approves what?
You name the approvers during onboarding. Strategy, audiences, brand-critical creative, regulated claims and budget moves above threshold always need a named person.
Can we change the thresholds?
Yes. Budget thresholds and risk tiers are defaults until your owner confirms them. The fixed rules, such as the order of precedence, are not adjustable.
How long does onboarding take?
We measure time-to-launch for every engagement and report it against your baseline. We do not publish a generic timeline.
Is mAI certified for compliance?
No certification is claimed. mAI is built to support GDPR, CCPA and PIPEDA obligations, and regulated clients add a compliance review step before launch.
Can we see the test results?
Test logs and decision records are shared during onboarding or a formal review.
