Skip to content

Platform · V7AV

AI marketing operations you can audit.

Specialist AI agents read your Google Ads, GA4 and Search Console data and propose changes. An 11-critic Council reviews every proposal, hard limits in code decide what may run on its own, and everything else waits for a human. Nothing ships silently.

Illustrative exampleV7AV

Proposal

  1. 01Marketing Agent
  2. 02Pause keyword
  3. 03CPA 1,380 DKK vs target 450 DKK
  4. 0414 days, 0 conversions

Council

  • 01MoneyApprove
  • 02TechApprove
  • 03DataApprove
  • 04LawApprove
  • 05SecurityApprove
  • 06EthicsAbstain
  • 07HumanApprove
  • 08BrandApprove
  • 09OperationsApprove
  • 10FutureConcern
  • 11DevilConcern

Decision

Tier 0 · within limits · logged

  • 11-critic Council
  • Hard limits in code
  • Fireman kill switch
  • Operated by V7.dk

Short answer

What is V7AV?

V7AV is the operating system behind V7.dk's AI marketing work. Agents do the analysis and write proposals. A Council of eleven specialist critics challenges each one. An orchestrator, A.B. BOB, routes it by risk: small, reversible changes inside hard limits can run on their own; everything else reaches a person first.

Agents
Marketing, SEO and Client Success, coordinated by A.B. BOB
Council
11 critics, three with a veto: Law, Security, Ethics
Safety
Magnitude limits in code and the Fireman circuit breaker
Data
Google Ads, GA4, Search Console, Morningscore
Status
In pilot on V7.dk's own accounts

The problem

AI can act fast. Unchecked, that is the risk.

Most AI marketing tools optimise one number and change accounts without asking. Agencies and in-house teams then find out afterwards.

  1. 01

    An automated rule pauses the keyword that brings your best customers, because it only looked at cost.

  2. 02

    Tracking breaks, the AI keeps optimising on bad data, and nobody notices for weeks.

  3. 03

    Ad copy drifts off brand, or makes a claim your legal team would never approve.

  4. 04

    No one can explain afterwards why a change was made, who approved it, or what it did.

How it works

From signal to decision, with a paper trail.

Four steps, every time. The same flow runs whether the change is a 5% bid adjustment or a new campaign.

  1. 01Hourly

    Data flows in

    Google Ads, GA4, Search Console and Morningscore sync into one database. Failed syncs raise an alert instead of failing silently.

    You getOne current picture of every account

  2. 02Daily and weekly

    Agents propose

    Specialist agents look for waste, drift and opportunity, and write a concrete proposal with the data behind it.

    You getSpecific changes, never vague advice

  3. 03Every proposal

    The Council reviews

    Eleven critics score the proposal from their own angle. Law, Security and Ethics can block it outright.

    You getA recorded vote with reasons

  4. 04By risk tier

    BOB routes it

    Routine changes inside hard limits run and are logged. Tactical and strategic changes wait in your approval queue.

    You getControl where it matters, speed where it does not

The Council

Eleven critics. Three with a veto.

  1. 01

    Money

    Is the spend proportional to the expected return, and are there hidden costs?

  2. 02

    Tech

    Will it work, can it be rolled back, does it strain APIs or quotas?

  3. 03

    Data

    Does tracking and attribution stay intact, and is the agent reading fresh data?

  4. 04

    LawVeto

    GDPR, ad platform policies, consumer and advertising law.

  5. 05

    SecurityVeto

    Secrets, attack surface, blast radius if something goes wrong.

  6. 06

    EthicsVeto

    No dark patterns, no manipulation, no vulnerable audiences.

  7. 07

    Human

    How will a real person experience this ad, page or message?

  8. 08

    Brand

    Would the client's founder be proud of it or embarrassed?

  9. 09

    Operations

    Can the team actually run and maintain what this creates?

  10. 10

    Future

    Where does this leave the account in 90 days?

  11. 11

    Devil

    The strongest argument against doing it at all.

Safety

Hard limits in code. Not promises in a prompt.

The AI never gets the final say on risk. Every action type belongs to a tier, and the limits are enforced by code before anything runs.

Tier 0 Routine

01
  • Small budget and bid changes, maximum 10%
  • Pausing clear low performers
  • Runs on its own, logged with before and after

Tier 1 Tactical

02
  • Negative keywords, ad copy tests, meta descriptions
  • Always waits for approval
  • Low-urgency queue

Tier 2 Strategic

03
  • New campaigns, large budget moves, tracking changes
  • Waits for approval, flagged as urgent
  • Full Council reasoning attached

Tier 3 Platform

04
  • Prompt, policy and permission changes
  • Always stops for the platform owner
  • Never automatic

Fireman, the circuit breaker, watches spend, tracking and data syncs around the clock and can pause any scope. Only a person can resume.

Why V7AV

Why agencies trust V7AV.

  1. 01

    Nothing ships silently

    Every proposal, vote and decision is stored with its reasoning.

  2. 02

    Limits you set

    Budgets, change sizes and tiers are rules in code, not suggestions to a model.

  3. 03

    Veto where it matters

    Law, Security and Ethics can stop any action on their own.

  4. 04

    A circuit breaker

    Fireman pauses first and asks later when something looks wrong.

  5. 05

    Isolated data

    Each agency and client is separated at database level.

  6. 06

    Used on ourselves first

    V7.dk runs its own marketing on V7AV before any client does.

Fit

Is V7AV right for you?

Good fit

  • You spend on Google Ads and want changes explained, not just made
  • You want AI speed without losing sign-off on strategy
  • You care about GDPR, brand and ad policy as much as CPA
  • You want reporting that shows what changed and why

Not a fit

  • You want a fully hands-off black box
  • You have no tracking and no plan to fix it
  • You need a tool to self-install today - V7AV is currently run by V7.dk as a managed service

FAQ

Questions about V7AV.

01

Does the AI change my ad account without asking?

Only Tier 0 actions inside hard limits, such as a bid change of at most 10%, may run on their own, and only when the Council raises no block. Everything else waits for a person, and every change is logged with its before and after values. During the pilot, changes to live ad accounts are made with a person in the loop.

02

What is the Council?

Eleven AI critics, each with one lens: Money, Tech, Data, Law, Security, Ethics, Human, Brand, Operations, Future and Devil. They vote approve, concern, block or abstain. Law, Security and Ethics can block on their own.

03

Which data does V7AV use?

Google Ads, Google Analytics 4, Search Console and Morningscore. Access is read-only for analytics, and each client's data is isolated at database level.

04

Can I buy V7AV as software?

Not yet. V7AV is currently operated by V7.dk as part of its managed services and is in pilot on V7.dk's own accounts. Book a walkthrough to see it on a real account.

05

What happens if tracking breaks or a sync fails?

Fireman, the circuit breaker, checks spend, conversions and data syncs every few minutes. It raises an alert and can pause the affected scope. Only a person can resume.

06

Who is behind V7AV?

V7.dk, a Copenhagen-area performance marketing agency working with Google Ads, SEO, tracking, web and AI agents.

Next step

See the Council work on a real account.

A 30-minute walkthrough on V7.dk's own live setup: a real proposal, the Council's vote, the routing decision and the audit trail.

No commitment. We show you the real thing, including what it does not do yet.