AIM
Back to AI Hub

Documentation

Architecture & API overview

How AIM turns disconnected marketing platforms into a single intelligence system that reasons, recommends and acts.

ArchitectureMarketing Intelligence LayerIntegration ProviderBusiness ObjectsAction FrameworkMarketing HealthAI AgentsFuture Integrations

Architecture

AIM is built as a layered pipeline. Marketing platforms connect through the Integration Provider. Their data flows into the Marketing Intelligence Layer, which normalises everything into unified business objects. Specialist AI agents analyse those objects in parallel, and the AI Chief Marketing Officer synthesises their findings into prioritised recommendations and approvable actions.

Integrations → Intelligence Layer → Business Objects → Specialist Agents → AI CMO → Recommendations → Approved Actions

The critical design principle: the AI never changes when you add a platform. New integrations map into the same business objects, so agents and the CMO keep working unchanged.

Marketing Intelligence Layer

Every platform exposes data differently. The Marketing Intelligence Layer ingests those raw API responses and converts them into a single, consistent schema. This is what lets AIM compare a Shopify store and a lead-gen site with the same reasoning.

{
  "revenue":     { "mrr": 48250, "changePct": 12.4, "churnRate": 3.1 },
  "traffic":     { "visitors": 82140, "bounceRate": 41.2, "changePct": 8.0 },
  "campaigns":   [{ "name": "Brand — Search", "spend": 4200, "roas": 3.8, "status": "active" }],
  "keywords":    [{ "query": "ai marketing platform", "position": 4.2, "ctr": 3.9 }],
  "conversions": { "conversionRate": 2.4, "trackingHealthy": true },
  "reviews":     { "averageRating": 4.7, "sentiment": 88 }
}

Integration Provider

A single managed layer handles OAuth, API keys, token refresh, rate limiting and secure action execution across every platform. Adding a platform means adding one registry entry — its data categories and available actions — not rebuilding the AI.

Managed OAuth & API-key auth
Automatic token refresh
Unified data syncing
Secure, audited action execution

Business Objects

The unified vocabulary the AI reasons about. Regardless of which platforms are connected, AIM works with the same objects:

Revenue

MRR, ARR, churn and lifetime value unified across payment and store platforms.

Traffic

Visitors, sessions, sources and bounce rate from analytics.

Campaigns

Every paid campaign with spend, ROAS, CTR and status.

Customers

Orders, AOV and cohorts across ecommerce and CRM.

Keywords

Organic positions, impressions and click-through rate.

Ads

Blended paid performance: spend, conversions, CPA and CTR.

Conversions

Goal completions, conversion rate and tracking health.

Funnels

Stage-by-stage drop-off from first touch to purchase.

Reviews

Ratings, review volume and sentiment for reputation.

Marketing Health

The composite score summarising overall marketing performance.

Action Framework

AIM does more than advise. When it recommends a change, it can propose a concrete, typed action. Nothing executes without explicit human approval, and every action is logged with its rationale and expected impact.

{
  "type": "pause_campaign",
  "provider": "google_ads",
  "label": "Pause 3 underperforming keywords",
  "rationale": "CPC up 28% with ROAS below 1.5x over 14 days.",
  "expectedImpact": { "monthlySavings": 1240, "currency": "GBP" },
  "status": "awaiting_approval"
}

Actions move through states: proposed → awaiting_approval → approved → executed (or rejected). Execution is routed back through the Integration Provider.

Marketing Health

AIM scores each core area from 0–100 and combines them into an overall Marketing Health Score, giving the CMO a fast, comparable read on where to focus.

Website0–100
SEO0–100
Advertising0–100
Email0–100
Conversion0–100
Revenue0–100
Social0–100
Reputation0–100
Overall Health0–100

AI Agents

Specialist agents each own a marketing discipline. They run against the business objects and return structured findings (good, watch or action) that the CMO ranks by revenue impact.

SEO Agent

Tracks keyword rankings, impressions and CTR, and finds the pages one push away from the top of the results.

Paid Media Agent

Watches blended spend, CPA, CTR and ROAS across every ad platform, flags wasted spend and recommends budget changes or pausing failing campaigns.

Analytics Agent

Reads sessions, sources and bounce rate to explain where traffic comes from and where it leaks.

Email Agent

Analyses open, click and revenue-per-recipient across campaigns and flows to grow owned-audience value.

Conversion Agent

Studies funnel drop-off, tracking health and page friction to lift conversion rate at each step.

Brand Agent

Tracks reviews, ratings and sentiment to protect reputation and sharpen market positioning.

Social Agent

Measures followers, reach and engagement across social channels and finds the content worth amplifying.

Future Integrations

Because everything maps into the same intelligence layer, the roadmap is simply more platforms. AIM currently supports 18+ marketing integrations, with more added continuously across analytics, advertising, ecommerce, CRM and social.

Google Analytics 4Google Search ConsoleGoogle AdsMeta AdsMicrosoft ClarityGoogle Business ProfileStripeShopifyWooCommerceMailchimpKlaviyoHubSpotSalesforceLinkedIn Company PagesYouTubeTikTok AdsGoogle Tag ManagerWordPress

Machine-readable resources: /llms.txt · /ai/schema