The Velocity Trap: How Working Faster Puts You and Your Company Behind

Employees keep moving faster and faster to prove their worth at their jobs in a volatile job market, but moving slower is an alternative option that may accentuate the advantage of showing up as a recovered human.

The ubiquitousness of the Agile sprint is killing the workplace.  For those unacquainted, the “sprint” is a short burst of work to be completed in less than a month by a team of coworkers.  This structure was created for software development, allowing software developers to break down pieces of a larger project into manageable pieces.  Unfortunately, the “sprint” work structure has expanded beyond software development and is popular throughout the technology industry, consulting firms, and more.

And it is killing workplace productivity, increasing worker burnout, and preventing work from remaining a sustainable activity.

Sprints were developed for short-term project completion with specific deadlines that also allow a recovery period once that project is completed.  Sprints can be effective for these short-term projects or in work with busy seasons.  When I worked in the budget office of the Defense Security Cooperation Agency, employees had to take on what would be an unsustainable workload for the month of September only.  Since the burst of work had a specific end date with recovery afterwards, it was possible to sprint for that month.

It is not possible to sprint forever.  The definition of a sprint is “to run or go at top speed especially for a short distance” or “a burst of speed.”  Sprints are short because they are unsustainable.  Nobody can run a marathon as quickly as they can sprint 100 meters.  But a growing number of companies adopt an Agile framework each year, usually poorly, and start encouraging their employees to start sprinting with no end date. By steadily demanding a faster pace, these companies fall into the velocity trap and hurt both employees and the company's bottom line.

Why are we sprinting?

The demand to sprint is increasing because the sprint model is becoming more commonly accepted, and AI is taking over the routine parts of our jobs.  My most recent job used the Agile framework for an ongoing policy review and revision project.  This is exactly the kind of long-term job that an Agile sprint does not help complete.  Reviewing long policy documents with groups of subject matter experts, recording all the revisions meticulously, and drafting a new document that all stakeholders agree covers nuanced policies is not an activity made for a sprint structure.  Even breaking the larger project down into pieces and claiming they are “sprints” does a disservice to the deep work involved in rewriting a policy.  While the project did not look like a project made for Agile sprints, it was forced to become that.

More and more white collar jobs revolve around the kind of work I described above.  A larger part of jobs are focused on deep work, the cognitively intense work we can only crank out for a maximum of four hours a day on our best days.  Our jobs seem to require more deep work than ever because AI agents are now capable of handling the busy work.

This shift towards more cognitively demanding work started before AI.  I streamlined daily reporting by writing macros to reformat reports, plug in formulas, and provide basic analysis of data received each day at a previous job.  Before writing those macros, a person took an hour each day to adjust daily reports.  With the macros, the same task took 5–10 minutes to complete.  Even before AI, technology allowed us to cut out 90% of the routine daily efforts.  Now, an AI agent could accomplish the job without human assistance, making that easy busy work obsolete.

This means that the slow parts of jobs are gone.  When I had a job, my ideal day involved a deep work morning, ideally getting to that full four hours of deep work.  Then, my tired brain needed to downshift.  Like everyone, I spent a short amount of time on emails, but I also planned activities like low-energy data management tasks that could be completed with a podcast running in the background.  I alternated between high cognitive tasks and low cognitive tasks to keep myself fresh and focused for the high cognitive tasks.

Now, the low cognitive tasks are gone.  All the work left for humans is deep work, but deep work is mentally taxing.  We cannot keep up the focus required for deep work for eight hours each day, but most jobs still require employees to work for eight hours.  Employers want to fill the spare time gained from AI automating our routine tasks with more deep work, but this keeps us constantly sprinting full speed.

But sprinting forever is impossible. We fall into the velocity trap when we believe faster is always better. We believe we just need the right coping mechanisms to maintain a new speed. But our current coping mechanisms do not help us avoid the trap.

Microshifting: An Unsustainable Coping Mechanism

Since it is not possible to sail through eight hours of deep work, some employees are working their eight hours with breaks to let them recover from one deep work block before diving into the next deep work block.  Employees might work from 7–10 a.m., take a break, work from 1–4 p.m., take a break, and then work from 7–9 p.m. in an effort to recover from each short burst of deep work.

This new approach to work is called microshifting.  While it allows a bit more recovery than trying unsuccessfully to complete deep work for eight hours straight or with a short lunch break, these short breaks still do not allow full cognitive recovery.  Microshifting lets us be more productive in secondary and tertiary bursts, but we are still not primed for multiple sessions of deep work in a day.

When we try to force extra deep work, we make mistakes.  Even with a couple hours away, a final push to get two hours of deep work in the evening is going to result in a lower quality of work than an ideal four- hour block approached with full recovery.  Additionally, spreading out work throughout the day leaves us less recovered and less likely to optimize a deep work block the following day.

Microshifting also prevents workers from having a life, family, or hobbies outside of work.  Outside of a quick yoga class between not-so-deep work bursts, there is no time for real non-work commitments.  This schedule also lacks time to truly be mentally “off” from work.  There is value in leaving work and living outside of it.  The act of leaving work behind switches our brain to another environment.  This is not only healthier for us overall, it also increases the likelihood of creativity at work.

Finally, microshifting has a dangerous tendency to lead to overworking when strict boundaries are not followed.  If that schedule of work from 7–10 a.m., take a break, work from 1–4 p.m., take a break, and then work from 7–9 p.m. becomes work 7–11 a.m., take a break, work from 1–5 p.m., take a break, and then work from 7–9 p.m., burnout looms closer, productivity drops, and creativity is nonexistent.  Deep work stops happening, and the worker’s health declines.

The Impacts of Microshifting to Maintain a Sprint

Sprints are unsustainable by definition, and the sprint work culture leads to higher burnout rates than companies that adopt a more consistent approach over time.  Teams that constantly sprint make more mistakes because deep work is not possible for longer than about four hours, it takes more time to fix those mistakes, and this often leads to even more hours of work overall.

Alternatively, allowing employees ample recovery time lets employees hit peak productivity in the hours they work.  They make less mistakes, creating less secondary tasks.  Getting it right the first time saves work hours, keeps employees primed to accomplish deep work more often, and makes the company more productive. 

Yet companies are moving away from models that provide employees with the recovery time they need to do the deep work only humans can do.

The Problem

AI eliminates the easy parts of our workday so we can focus on the cognitively taxing tasks.  We all have had a day filled entirely of monotonous work that AI can now do in seconds, and most of us are grateful to have AI take over those pieces of our day.  However, companies need to understand that employees cannot simply shift to do entirely cognitively demanding work for eight hours each day.

The current friction exists because there is a misunderstanding about how to spend the four hours of the day when our brains can no longer complete deep work.  Companies still want workers working, but defining what they should be doing is difficult now that the easy tasks are gone.

A secondary problem is the slow loss of understanding the work at its core.  As companies remove day-to-day tasks from all human portfolios and have AI complete them, we eventually arrive at a place where no human understands all the tasks that AI is now completing.  I saw this at a lower level in my previous job, the Agile sprint-obsessed policy document revision one.  A macro existed to organize data from Word in Excel in a specific format.  However, nobody on the team was part of the initial creation of the macro, and nobody had the skills to edit it if it broke or needed to be improved.  As someone with experience writing macros for Excel only, I expressed interest in receiving the advanced macro training that would allow me to understand the macro that organized Excel data in Word.  My previous employer decided there was not time within the workday for me to gain that knowledge.  Instead, that entire team of 40 people continues to run the daily risk of their work stopping because they do not have the time to train one person how to fix and improve a necessary technological process to complete their work.

My prior job had one potential chokepoint due to the team lacking knowledge of the specifics of an important part of the work process.  As AI completes more tasks, that opens businesses to more potential chokepoints that could hold up work.  Fewer standard operating procedures are being written because AI is taking care of small tasks, but that means fewer employees really understand the work process from start to finish.  

Not only does that risk the process breaking, it also prevents creative human ideas for process improvement because few employees have a grasp on the bigger picture of their workplace’s processes.  If nobody has a complete understanding of a company’s processes, that makes us reliant on AI to come up with better solutions as human workers become less skilled in process improvement.

The Solution: A New Work Day

Most of us can agree to dedicating 3–4 hours to deep work.  Humans are still responsible for the cognitively difficult tasks, and this will continue at least until Artificial General Intelligence is mastered.  The question that remains is what to do with the other half of the work day to allow workers to recover adequately while companies still get work hours from them.

First, I would set up minimal common touchpoints to address the collaborative human side of work.  I do not mean the daily “scrum” suggested by the Agile framework since scrums always last longer than they should, at work and on the rugby pitch.  Meetings tend to serve as opportunities for employees to talk a lot to prove they are busy.  This is a waste of time.  The most successful companies will have fewer, briefer meetings while assessing employee performance on the outputs from their deep work efforts.  Ideally, a team might meet once each week to problem solve and address any aberrations in their work.

Additionally, while companies agree that deep work should happen, they fail to schedule for it.  At times, my work schedule was filled with hour-long meetings with unfortunate spacing that prevented deep work blocks.  Protect at least three times each week for employees to have lone, four-hour deep work blocks.  This allows them to complete the outputs that make human employees valuable.

With the extra time that remains in a 40-hour work week, I would suggest:

1. Prioritizing consistent learning within regular work hours,

2. Discussing and documenting work processes to maintain institutional knowledge, and

3. Reducing hours worked.

Prioritizing learning and documentation of work processes prevents the chokepoint problems I addressed in the previous section.  Keep employees talking about the processes crucial to their institutions to keep creative ideas flowing.  When employees have ideas about process improvement but lack the technological skills to implement them, give them time in the work day to begin training to gain those skills.

Skill development has largely been an outside-of-work-hours effort in recent years.  But we had to pull daily reports, manually enter data, and other tedious tasks when that was true.  Since AI is doing those tasks, allowing employees to keep training for self-improvement in their careers will keep companies current as technology changes.  Learning is also a positive creative break for the brain after deep work activities, making it easier for employees to recover for their next deep work period.

Finally, since humans simply cannot complete deep work for eight hours a day, we should be looking at AI as an opportunity to reduce our work days rather than try to force our brains to do something they cannot.  The 40-hour work week was established to maximize productivity for factory workers.  AI has most of those jobs covered.  As we enter a new era of work where we are responsible only for the distinctly human tasks, it is time to reduce our hours worked.  Humans are still the only ones to complete the most cognitively challenging and creative tasks.  To keep our advantage, we need to fully recover between periods of deep work.  Reducing the total hours worked will make each hour of work more productive, benefitting companies that see increases in productivity and employees who experience better health, less burnout, and more time to live their lives.