Every few weeks, another AI assistant promises to take a task off your plate. The promises are tempting: it can draft the email, sort the inbox and plan the week. The question is no longer whether these tools can do the work. It’s whether they are lightening the load or quietly adding to it. Morissa Schwartz, an entrepreneur, author and AI strategist, recently discovered just how easily an attempt to improve productivity can turn into another job. “I believe in AI. But my AI assistant gave me more work than it could finish. I opened my daily to-do list and found 156 items,” she wrote in
Fortune.
Despite having AI assistants to manage her workload, she found her daily to-do list flooded with 156 items. Instead of making her work easier, her AI assistant had created a growing list of tasks. The reminders and follow-ups threatened to overwhelm the very system designed to keep her organised.
Her to-do list hit 156 items
Morissa Schwartz explained how asking AI to help with a project often leads to a plan, which is then followed by a checklist and eventually a series of additional tasks. Though each may appear useful individually, the growing list can turn productivity management into another task.
She isn’t against technology. Her complaint is that an assistant can produce work faster than anyone can finish it. Her fix was to cap her daily list at roughly 25 meaningful entries and ask a sharper question about every task: what would ‘done’ look like to the person who asked for it? In her profession as a publisher, that would mean a finished draft. If the request was to pitch it, done means the pitch was sent.
She revealed how she retrained her AI assistants. “Here are the changes I’m making. First, I want a daily list that reflects available capacity. For assistant execution, I’m aiming for roughly 25 meaningful entries on a weekday, while preserving real deadlines and necessary daily operations. The number is a working limit for my system, rather than a productivity prescription for everyone.” Its purpose was to force her system to prioritise tasks instead of treating everything as equally urgent. “If everything is due today, the list has stopped helping me decide what to do today.”
Flexible work also moves into a backlog with a regular checkpoint. Items waiting on someone else get a waiting status rather than a place at the top. Related steps, such as research, outreach and follow-ups for a single campaign, are grouped under one execution item. The details stay available, but they stop competing for attention. “A genuine deadline deserves space that has not already been occupied by 40 vaguely urgent reminders,” she wrote.
Proof before checkmarks
Her second rule is to group related tasks under one main item instead of listing each one separately. For instance, research, outreach and follow-ups for a campaign could be managed under a single task, keeping the to-do list shorter and easier to manage without losing important details.
Third is, of course, evidence. If the assistant says an email went out, she wants to see it in the sent folder. If something was published, she wants the live link. If a file was delivered, she wants the usable file and a record of the delivery. It’s not that she does not ‘trust’ her agents. Rather, she believes evidence matters. “That proof also helps prevent duplicate work. A system that remembers what actually happened can stop proposing the same action with slightly different punctuation,” she wrote.
Fourth, a ‘blocker’ needs to be specific. “‘Unable to complete’ gives me another mystery to manage. ‘The form requires an answer only you can provide’ gives me a decision. ‘This browser session cannot access the account’ identifies an access problem,” she said. She wanted the AI to explain clearly why a task could not be completed, whether it needed her input or had an access issue. This would help her solve the problem quickly without delaying other tasks.
Finally, Schwartz measures outcomes. How many pitches went out? Which replies still need an answer? What was accepted? Which deliverables reached the client? What remains unfinished? “I care about those numbers because they describe the work I was trying to accomplish. A magnificent list of possible actions describes the system’s imagination,” she said.
Schwartz is not arguing that AI should do less. She still wants an assistant. What she wants is one that rewards finishing, checks its own claims and says plainly what is left. By her measure, a good day has fewer items on the list and a clearer record of what was sent, published or delivered.