02 Categories, not tools
AI tools change every few months. The leader in any category today may not be the leader in twelve months. Names shift, features rearrange, new tools arrive, and tools you rely on get retired or absorbed into something else.
The categories of work AI does in comms are different. They change much more slowly. The five categories below have been stable for some years and are likely to be stable for some years more.
This course teaches the categories first, the tools second. The reason is practical. If you understand the categories, you can swap tools in and out of your shortlist when something changes, without rebuilding your practice each time. If you only know specific tools, every change forces you to start over.
The framework underneath: the nine constructs of the AI-ICM
The five tool categories tell you what AI tools exist. They are useful, but they only get you halfway. The other half is what those tools are actually doing for you: generating content, tailoring it for specific audiences, and distributing it across channels. That work is the centre of the AI-ICM.
The AI-ICM, the Artificial Intelligence Impact Communications Model, places AI content generation, tailoring, and distribution at the centre, and identifies nine constructs that surround it. Each construct names something the surrounding work needs to address. Other courses teach tools. Others teach board-level principles. This one sits in the operational middle: it teaches tools alongside the nine constructs that make the use of those tools credible, accountable, and effective.
The AI-ICM: AI generation, tailoring, and distribution at the centre, nine constructs around it, the operating conditions framing all of it. Together they form a governed service capability: a repeatable function with checks, accountabilities, and evidence it operates safely.
Reading the diagram. At the centre is the work AI is actually doing in your communications: generating content, tailoring it for audiences, and distributing it across channels. Around the centre sit the nine constructs. Each names something that work must address to be trustworthy, effective, and accountable. The constructs are not a checklist; they are the dimensions along which your use of AI in comms is either credible or not. A practitioner who attends to all nine produces AI-enabled comms that holds together. A practitioner who attends to two or three produces comms with gaps a colleague, a regulator, or an audience can find.
How the constructs relate to the five tool categories
The categories and the constructs are complementary, not competing. The categories describe what AI tools do. The constructs describe what your communication needs to address. Most practitioners use tools from several categories to deliver each construct. Trust and transparency, for example, is delivered partly through conversational tools that cite sources, partly through workflow tools that keep an audit trail, and partly through monitoring tools that flag drift.
In Week 1 you map your tools. In Weeks 2 to 6 you build the practices that address each construct. By Week 6 your AI Playbook covers all nine through a tool stack that is yours.
The nine constructs came from three sources: a structured review of how AI is used in high-stakes communication today, co-design workshops with communicators and their audiences, and a synthesis of behavioural, communication, and AI governance research. They are not OpenAI's, Google's, or Microsoft's framework. They are what emerged from one question: what does AI-enabled communication need to be in order to be credible and accountable in real services? Hold the constructs alongside your work, not as a list to memorise. Each is a question worth asking at every stage of planning, generating, and distributing. Some weigh more heavily on some work than others.