This year I arrived at a phase readout with more material than fit in the hour. There was a main deck, three supplemental decks, and an interactive summary. I asked the client to review the rest on their own and to schedule deeper dives where they wanted them. A few weeks earlier, on a different engagement, the client's product leader asked us to change the plan. His team had received a lot from us and had not been able to sit with most of it. He wanted a full working day together, even though it would disrupt the schedule. He wanted it before the executive readout, because without it there was not enough to read out.
On both engagements I had planned the work the way I always had. Then I produced at a rate the plan did not account for. This is written for other consultants working with the same tools, because the problem is ours to solve.
Our work ends where production starts
Most of what is written about AI moving the bottleneck is about building software. There, the output is code, and someone other than the author can review it. Our work happens before that, in the decisions about what to build and why. The output is a decision. A decision is made when the person who has to defend it understands it well enough to do so, and nobody else can do that part for them. The path from a decision to working software is shorter than it has ever been, so a wrong decision reaches customers faster too. The upstream work carries more weight now.
This was true before AI
Clients hire an outside partner because they need expertise or capacity they do not have, and usually both. Either way, the work still needs their time. We do not take a brief and come back two months later with an answer. The client knows what will work in their organization and what the problem actually is. None of that is available to us without them. For years we have known that some clients do not have the time to get help. The clients who hire us for capacity usually have the least time to give.
What changed
Every meeting, interview, and test on an engagement is now transcribed into a shared record the whole team can search. A synthesis across a round of stakeholder interviews is drafted in minutes. The people who ran the interviews then spend an hour checking it against what they heard. A few years ago that was a day of the whole team at a wall of sticky notes.
Producing a draft is fast now, and editing it takes as long as it did before. Pascal apologized for a long letter by saying he had not had time to make it shorter, a line often credited to Twain, and we are in the same position. On one engagement we ran roughly twice the interviews we would have a few years ago, in the same two weeks. That meant twice the people who had told us what they thought and expected to see it reflected back. On another, the client asked for breadth before depth, and we generated close to two hundred concepts at low fidelity. Every one of them needed someone to judge it. The people who had to judge them had no more hours than before.
It is well understood by now that production is no longer the slow part. What I have not seen discussed is what that does to the plan. Our plans were written as if the client's time and the client's budget moved together.
What an afternoon together found
The in-person day the product leader asked for narrowed those two hundred concepts to under forty in an afternoon. It also showed a gap in the set that nobody had seen in weeks of reviewing it asynchronously. The concepts covered one part of the product well and barely touched the rest. The breadth had been worth having, because the gap was only visible against it. A week later he asked for a set of principles his team could use to judge what we had made.
Two things we control
The client's hours are fixed, and more budget does not create more of them. We used to scope an engagement to the budget and assume the client would find the time. For years that worked. Now the same budget buys more output than the client has hours to judge. The production plan has to be sized to their time.
The second thing we control is how the judging gets done. Every SOW we write states the problem and what success looks like. Judging the work against that definition has always had its place, in the hour or two we have with the client each week. That time has not grown, and the amount to judge has. So judging needs more of the calendar than it used to. The criteria deserve the same iteration with the client that we would give any concept, because a concept can be judged quickly against criteria the client has argued over. On the concept engagement we spent more time making concepts than judging them. The judging is the part the client has to be in.
Where our hours go now
Nothing in this section is new to us. We have always synthesized for the client's audiences, run working sessions meant to end in a decision, and tested with users. What has changed is the ratio. When production was slow, most of a team's hours went to producing, and the rest went to helping the client judge. Production now takes a fraction of the hours it did, so the ratio should move the other way. The client's scarce resource is the hours they can spend judging. Our hours are most valuable where they make the client's hours worth more.
In practice that means more of our time on synthesis for the client's own audiences, so that what they take upward can be defended by them. It means more time designing working sessions that end in a decision, and more rounds iterating the criteria with the client. It means testing concepts with users before the client has to choose among them, so that what they judge is evidence. And it means holding fidelity low until an option has passed their judgment, because a finished-looking artifact costs more of their time to evaluate than a rough one. The total hours do not change. The share that goes to producing gets smaller.
The plan on file
The tools stay as fast as they are, and the work gets sized to the people who have to use it. The plan most of us have on file starts from budget and team size. It treats the client's time as something that will be found. It is worth rewriting so that it starts from the client's time.
If you are rewriting yours, message me on LinkedIn.