A hypothetical form of AI capable of performing any intellectual task a human can, across domains, without task-specific training for each one. Agencies encounter this term primarily in strategic planning conversations and vendor positioning, where it is often invoked to suggest capabilities far beyond current tools, rather than anything technically precise.
Also known as AGI, strong AI, general AI
Artificial general intelligence (AGI) refers to a machine capable of performing any cognitive task that a human can perform, with comparable or greater competence, and the ability to transfer knowledge between domains without being explicitly retrained. Today’s AI systems, including the most capable large language models, are narrow: they are trained for a class of tasks and perform poorly when applied outside that class. AGI would have no such constraint.
There is no consensus on whether AGI is achievable, when it might arrive, or what it would require. Estimates from researchers range from decades to never. The goalposts also shift: tasks once considered milestones of general intelligence (playing chess, generating coherent text, recognizing faces) are now routine, which makes it harder to define what general intelligence actually means once machines can do specific things well.
The term has significant rhetorical weight in the AI industry and attracts large amounts of funding and media coverage. Agencies should be able to distinguish it from current AI capabilities when clients or vendors invoke it casually.
Agencies are in the business of anticipating change and advising clients on what it means for their brands. AGI sits at the outer edge of that mandate, but it is relevant precisely because of how often it gets misused in strategy conversations. A CMO who believes AGI is imminent will make different budget decisions than one who understands where the technology actually is.
Vendor credibility. AI tool vendors frequently gesture toward AGI-level capabilities to justify pricing and adoption urgency. An agency that can evaluate those claims critically is a better partner than one that passes vendor hype along to clients unchanged. The question to ask is always: what specific task does this tool do better, and what does it still get wrong?
Long-range planning. Clients increasingly ask agencies to help them plan for AI-driven disruption to their industries. A coherent answer requires separating current narrow AI capabilities from speculative general intelligence. Conflating them leads to either complacency (“AGI is far off, so nothing matters yet”) or overcorrection (“AI will soon replace all human judgment”).
Staff and culture conversations. The question of whether AGI would replace creative professionals comes up in agency team meetings. Having an accurate, grounded answer builds trust with staff and prevents distorted thinking about where the real near-term exposure lies.
A technology client asks the strategy team to assess how soon AGI will affect their consumer base. The strategist’s job is not to predict timelines. It is to build a scenario framework that separates what current AI can demonstrably do from what would require capabilities that do not yet exist. That framework becomes the basis for near-term decisions (what to automate now, what to protect) and contingency plans (what to monitor as signal).
The output is useful regardless of when or whether AGI arrives, because it forces a granular look at which tasks in the client’s workflows depend on human judgment versus which are already automatable with narrow tools. That analysis is valuable on its own terms, independent of the speculative question.
The generative AI foundations module of the workshop covers how today’s models work, what they can and can’t do, and how to choose between them.