Meridian — The Outcomes Engine: Intelligence Layer
Context Graphs, BrandTwins, and M-Agents — Genuine Memory at Scale
From data to memory. From segments to markets of one. From operators to autonomous systems.
- Meridian is NeoMarketing’s answer to the incentive failure. Its domain is the Best — the 20% of a brand’s customer base generating the majority of revenue. These customers are already acquired. The challenge is not to find them again — they are present, active, and spending. The challenge is to deepen the relationship, expand LTV, and ensure they never quietly slip into the drifting middle. Best customers rarely announce their departure. There is no dramatic churn event, no complaint filed, no cancellation form submitted. They simply buy less. Their engagement frequency drops. Their basket size shrinks. And by the time the metrics reflect it, the drift and exit is already well underway.
- The intelligence foundation of Meridian is the Context Graph. This is a decisive move beyond static data and static segments. The Customer Context Graph captures evolving customer state — preferences, engagement patterns, purchase momentum, fatigue signals, category affinity, and brand sentiment. The Product Context Graph captures the world of offers, substitutes, complements, inventory positions, and margin economics. The Decision Trace Graph captures what marketing actions were taken with this customer, why they were taken, and what outcomes followed. Together, they give the marketing system genuine memory — not storage, but usable, decision-relevant intelligence about each individual Best customer. A segment tells you where a customer sits. A Context Graph tells you who they are, how they move, and what they are likely to do next.
- Built on Context Graphs, BrandTwins are AI-powered individual customer advocates — N=1 personalisation at scale, one per Best customer, each with its own live ledger tracking actions and outcomes. Where current CEE treats customers as members of segments, BrandTwins treat each customer as a market of one. One Best customer may be momentum-rich but fatigue-sensitive: send too often and they disengage. Another may respond powerfully to discovery and novelty but not to discounting. Another may be at quiet drift risk despite apparently healthy aggregate metrics. The BrandTwin holds that evolving, individual understanding — and acts continuously to preserve relevance, deepen the relationship, and steer the customer away from the drift pattern before it becomes visible in the data.
- The orchestration layer is M-Agents: single-purpose marketing agents — Insights, Audience, Content, Decisioning — coordinated by a Co-Marketer orchestrator into a multi-agent system. This matters because one of martech’s most persistent problems has not been lack of features but lack of usable operational intelligence and purposeful execution. CEE platforms became extraordinarily capable — and extraordinarily underutilised. Teams were handed powerful instruments and not enough and capable drivers. M-Agents change that. The Co-Marketer orchestrates the single agents into a system that works toward defined outcomes. The platform no longer waits for human configuration of every branch, every workflow, every trigger rule. It becomes an autonomous system executing toward agreed goals — with the brand’s marketing team shifting from operators to strategists.
- The TWIN data framework underpins all of this. T — Transactional data. W — World data: context, macro signals, category dynamics, external environment. I — Individual data: preferences, behavioural history, engagement patterns. N — Nano data: moment-level signals, real-time intent, micro-context. Together, TWIN gives BrandTwins and M-Agents the data richness they need to move from pattern-matching to genuine individual prediction — and from prediction to action. The system knows not just what each Best customer has done, but what they are likely to do next, and what intervention is most likely to compound rather than disrupt the relationship.
Key Takeaway: A segment tells you where a customer sits. A Context Graph tells you who they are, how they move, and what they are likely to do next. That is the difference between data and memory.
