How Many AI Launches Can One Team Absorb in a Single Quarter?
How many AI launches a team can absorb in a single quarter depends on available person-weeks rather than an arbitrary count, evaluated using Dr. Jonah Tebaa's Landing-Load Table. Under this framework, cumulative absorption demands are compared against a default ceiling of 25 percent of a team's quarterly person-weeks. In a composite finance case, this ceiling drops to 15 percent during peak periods like year-end close, where absorbing two concurrent launches would exceed capacity unless leadership slips dates, splits rollouts, or adds temporary cover.
Picture a finance team of eight people in the quarter that holds the year-end close. Two AI launches have been approved for that same quarter, and both will change how the team works every day. One touches supplier payments. The other touches month-end reporting. The board has seen the business cases, the budgets and the dates. It has not seen what the two launches will ask of those eight people while they are closing the books.
This is an illustrative composite, not a client engagement. Dr. Jonah Tebaa builds scenes like it to make a point about how boards approve AI roadmaps: they check cost and value carefully, and they rarely check whether the people on the receiving end have the time.
A second budget, paid in weeks
Dr. Tebaa argues that every funded AI roadmap carries a second budget that never appears in the pack. It is the working time of the teams that must absorb each launch. In his work he describes what fills that time: learning the new way, running old and new methods side by side, dealing with the cases the tool does not handle, redoing what came out wrong, and repairing the process around the tool.
His reasoning for why it goes unseen is plain. Those who estimate a launch tend to count the build, because the build is what they own. The absorbing is owned by someone else. So the roadmap passes on value and on dates, and the load surfaces later, as a team that is quietly overstretched.
The instrument: a Landing-Load Table
His remedy is one table that a board can request in a single meeting. He calls it the Landing-Load Table. Receiving teams run down the side and quarters across the top. Each cell shows the person-weeks of that team's time the launches will consume. He compresses the method to five moves: name the team whose daily work changes, ask that team's lead for the weeks, compare each team's quarterly total with a ceiling, resolve whatever is over, and re-run the table when dates or headcount move.
The default ceiling in his method is a quarter of a team's total person-weeks. That total is simply the number of people multiplied by the thirteen weeks in a quarter.
Why the landing team, not the sponsor
A detail Dr. Tebaa insists on is whose name goes in the row. It is the team that lives with the change, which is often not the executive who sponsors the project or the manager who owns it. A sponsor can be enthusiastic and fully available while the team two floors down has no room. A table organised by sponsor would show a healthy spread of ownership and hide the pile-up. Organised by landing team, the same plan reveals where several launches converge. One initiative may also land on more than one team, and each of them gets a row.
The peak quarter changes the number
Back in the finance scene. Eight people over thirteen weeks give 104 person-weeks. At the default ceiling of 25 percent, the roadmap may take 26. The two launches ask for 9 and 11, which is 20 in total, comfortably inside that ceiling by 6.
Dr. Tebaa’s point is that the default is the wrong test for this quarter. A year-end close is the team’s peak, when its normal work is heaviest. He lowers the ceiling there, and in the composite he sets it at 15 percent, about 16 person-weeks. Against that figure, the same 20 is over by 4. A plan that looked safe on the default is not safe in the quarter that matters most, and only a table that tracks each team’s calendar shows it.
Three ways to resolve a cell that is over
When a cell exceeds its ceiling, Dr. Tebaa offers a board three choices. The first is to slip a launch by a quarter, provided nothing else in the plan depends on it. The second is to split it, so that a smaller first landing fits the weeks available and the rest follows later. The third is to add capacity, for example temporary cover, and to state its cost openly.
In the finance composite, any of the three would bring the quarter within its peak ceiling. Moving one launch to the next quarter would cut the load to 9 or 11 person-weeks, both inside the 16. He is also clear that each choice has a price. A slip delays a benefit. A split leaves part of the work undone. Cover costs money. The board should see the chosen option and its price, and not only the tidy result.
Who should ask first, and what the table cannot do
Dr. Tebaa suggests the request is most natural from the people who sign off the dates: chief executives, chief operating officers, finance chiefs and non-executive directors, and the heads of strategy or transformation who prepare the pack. The question is easy to put and needs no technical knowledge: which team absorbs each launch, and in which quarter?
He is candid about the limits. Person-weeks are estimates, not measurements, and the ceiling is a judgement. The table does not rank which launch is worth more, and a team can sit below its ceiling and still lack a skill the new work needs. For him the table is a way to make one hidden constraint visible before the dates are approved.
His summary line carries the whole argument: a roadmap is not funded until the people who must absorb it have the weeks.
Related evidence: The ILO's assessment of global occupational exposure to generative AI finds that its main effect is likely to augment rather than automate jobs. (ILO working paper on generative AI and jobs)
RAND's study of AI project failures finds that misunderstandings and miscommunications about the intent and purpose of a project are the most common reasons AI projects fail. (RAND research on why AI projects fail)