No 2026 marketing plan had AI visibility on it. The work showed up anyway, an audit by spring, a workstream by summer, and for a lot of teams it reshaped the year. Responding was the right call. What it displaced was never decided, and that difference ran the year.
The budget and the plan used to be one December exercise, and only the budget survived it. Finance requires a number, so the number still gets set every year. Nothing requires the plan, and once the things it used to schedule stopped following a calendar, it quietly stopped being written.
Why planning went away
Media went first. Annual planning existed partly because media was sold annually, with rates negotiated in spring and inventory reserved before the fiscal year opened, and that requirement is simply gone. Streaming platforms now let buyers cancel up to 100% of upfront commitments 14 days before a campaign runs, and the money committed in the upfront has declined four years straight. When you can buy on a Tuesday, nothing forces a December decision.
Launches stopped holding a calendar around the same time. Products that used to need three quarters of runway come together in weeks, which means the August launch doesn’t exist yet in December and a launch calendar written at year end is wrong by February. Content went the same direction, the editorial calendar assumed a quarter to produce anything worth shipping and production now takes days. The tools turn over in months while the martech line renews annually, the team flexes by project and by quarter, and the measurement that would have graded any of this stopped cooperating, with the research moved out of view and AI search changing who gets found in the middle of the year.
Each change made one more piece of the plan feel pointless, and the plan went away one reasonable surrender at a time. What replaced it is a budget and a year of responses.
What a year of responding costs
Responding is the job now. Something new arrives every quarter, and a team that ignores the arrivals falls behind in ways that compound. The costs come from responding without having decided anything in advance.
The money comes from somewhere. When AI visibility arrived, the budget that covered it came out of work already underway, and usually out of the long work, because the long work is the only line without a vendor contract attached to defend it.
Nothing runs long enough to prove anything, either. A channel needs quarters of consistent spend before its numbers mean much, and a year of reallocation produces a dashboard where every line was touched and none of it is evidence. So marketing arrives at budget season able to list its activity but not what the activity was building toward, which reads as a cost center, and cost centers get cut. The year of reacting shrinks next year’s budget.
Both things are true
You can’t plan the way you used to, nothing holds a calendar and the biggest workstream of next year doesn’t have a name yet. You also can’t run a year of pure response, because regulated buyers read constant direction changes as instability and finance reads unprovable spending as waste. Marketing has to respond fast and still show in December that the year added up to something, and the old plan format answers none of that.
What the plan is now
The plan that fits this speed is loose on activities and fixed on what connects them. It doesn’t schedule the year’s work, because it can’t. It decides what the year’s work will have in common.
Messaging pillars come first, the three or four things the company keeps saying no matter what arrives. When something new shows up, and it will, the response gets built from the same pillars as everything else, so the AI answer, the tradeshow booth, and the campaign that didn’t exist in December all sound like one company. Teams with pillars absorb unplanned work into the story they were already telling. Teams without them end the year with 12 months of disconnected output.
Then one theme carrying the year, the same idea across the tradeshows, the same activations working venue to venue, the webinar series and the content building on each other instead of starting over each quarter. A year can be half unplanned and still read as one body of work if the half you did plan was the connective layer.
The mechanics underneath are tablestakes. Commit the year’s outcomes, protect the work that compounds, hold a reserve for the arrival nobody has named yet, set quarterly review points, and write down who makes the mid-year call. All of it is necessary. The plan is the pillars and the theme, the only decisions that make a flexible year coherent.
Planning is still possible. What changed is what gets planned.
A budget is how much. A plan is what for.
