Client briefs arrive written for search and social: objectives, budget, audience adjectives, a competitor list. The planning job is to convert that into ad groups defined by situations. Doing it in a fixed sequence keeps the work to a couple of hours rather than a fortnight of debate.
Step one is extraction. Pull five things out of the brief verbatim: the objective in the client's words, the named competitors, the stated constraints, the proof points the client believes in, and anything the client says the product is not for. That last item is usually buried in a caveat and is the most valuable line in the document.
Step two is situation mapping. For each named audience, write the two or three circumstances in which that audience would actually go looking. Push for triggers — a renewal, an audit, a launch, a headcount change. A situation without a trigger will not produce realistic prompts.
Step three is evidence matching. Against each situation, list what the engine would need to verify to recommend the client. Mark each item as available, partial or missing. The missing column becomes a workstream, and showing it to the client in the kickoff is how you get feed work funded before it becomes an excuse later.
Step four is the plan itself: one ad group per situation, with a hint, a core promise, the landing path, and the prompts you will monitor. Sequence the groups by evidence readiness, not by audience size — a large situation you cannot yet evidence should be scheduled after the feed work, not launched hopefully.
Step five is the commercial frame. Attach an expected role to each group: defend, expand or explore. Defend groups protect prompts you already win. Expand groups push into adjacent situations with existing evidence. Explore groups test situations where you would need new proof. A plan that is all explore looks exciting in a pitch and fails in month three.