ian decisionning landing

AI Decisioning: a simple yet fundamental definition

AI Decisioning is not some futuristic promise. It is not a new layer of chatbot, nor an “AI” feature added to an existing tool. It is an operational software component – already very much a reality – that is fundamentally transforming the way brands make their marketing decisions.

Its role is precise: to determine, for each individual, which action to initiate – or not to initiate – on which channel, at what time, and with what content. And to do so whilst taking into account the customer’s actual situation at that precise moment: their behavioural context, their stated preferences, their recent interactions, the level of marketing pressure they have already been subjected to, and the prevailing business constraints.

The distinction from traditional automation is fundamental and often misunderstood. An automation tool executes what it has been told to do. It runs fixed scenarios – “if A then B” – without ever questioning the relevance of the action. It knows how to send out multiple messages. It does not know how to choose the right one.

AI Decisioning, on the other hand, decides what needs to be done

In an environment where customer behaviours no longer follow linear sequences – they intersect, interrupt, contradict and accelerate – this nuance makes all the difference. It observes the customer’s actual state, weighs up all possible options, and decides: should we follow up? Wait? Suggest something else? Or refrain entirely — because that silence is the best relational decision at that moment?

It is this shift that introduces a new concept into marketing: contextual relevance. No longer personalising the content of a message, but personalising the decision itself. No longer managing campaigns, but managing choices — for each individual, in real time, across millions of contacts.

And this is where CRM ceases to be a message factory and becomes what it should always have been: a system of discernment.

Definition

AI Decisioning = deciding, not just executing:

Your automation tool knows how to do what it’s told. It doesn’t know if it’s the right time. It doesn’t know if your customer is already overwhelmed. It doesn’t know that it would be better to stay silent.

AI Decisioning, on the other hand, chooses. For each individual, at every moment: what action to take – or not to take – on which channel, and with what intensity. Taking into account the customer’s actual context, not a rule written six months ago.

The difference comes down to one line: automation executes what it has been told to do. AI Decisioning decides what needs to be done.

Automation

Predictive AI

AI Decisioning

LOGIC

If A then B

Probability

Contextual decision-making

PERSONNALISATION

Segment

Predicted individual

Real individual

DECISION NOT TO ACT

Yes

CONTINUOUS LEARNING

Partial

Yes

The 3 intelligences

Three AIs. Three roles. A single system.

01 – Generative AI creates. It writes, adapts and personalises content at scale: tone, angle, format. It brings the creative variability that humans cannot produce in real time. But without a framework, it produces volume – not consistency.

02 – Analytical AI predicts. It calculates probabilities: open rates, purchases, churn, optimal timing. It transforms past behaviour into actionable signals. It doesn’t decide – it informs.

03 – Decision-making AI arbitrates. It is the conductor. It cross-references the data, incorporates business and ethical constraints, and makes the call. Who to contact? When? On which channel? With what level of urgency? Or should we hold back?

These three forms of intelligence are not interchangeable – they feed off one another. It is their interplay that drives performance.

Cognitive CRM & Trinity Engine

The Trinity Engine – The brain of ian®

ian® by Notify is based on a continuous three-layer architecture, inspired by the functioning of the brain:

PERCEPTION – Capturing the real situation Behavioural signals, external contexts, stated preferences, level of consent. The system sees the individual in their current state — not in the segment to which they belong.

DECISION – Weighing up all options Balancing business, relational and ethical priorities. Choosing the right channel, the right level of engagement, the right timing. Or deciding that the best course of action is to take no action.

EXECUTION – Act within your existing stack The chosen action is sent via your existing tools: email, SMS, push notifications, app, call centre. The result immediately feeds into the models for the next decision.

A marketing cortex that coordinates without dictating, streamlines without imposing, and learns from every interaction.

The algorithmic pool

Six engines. One goal: the right decision.

ian® is not a single model. It is a set of specialised engines that work together – each an expert in its own field, all governed by the orchestration engine.

Perfect Timing – Identifies, for each individual, the moment when they are actually available. Not the statistical best time to send: their own specific moment.

Behavioural scoring – Assesses in real time the probability of engagement or conversion, based on history and recent signals.

Marketing pressure – Regulates the frequency of communications on a contact-by-contact basis. It knows when to stop before the customer unsubscribes.

Channel selection – Selects the best channel based on actual preferences, observed deliverability and recent performance per individual.

Prediction – Anticipates churn, inactivity and reactivation. Take action before the problem becomes visible in your metrics.

Affinity – Models implicit preferences for content, category and offers – even in the absence of explicit signals.