Your Growing Backlog May Be Evidence of Leadership Failure
When recording everything becomes more important than deciding what matters
A vulnerability is discovered.
Create a ticket.
A customer requests a feature.
Create a ticket.
A meeting produces an idea. A monitoring tool detects an anomaly. An audit recommends an improvement. AI analyzes the organization and identifies 147 opportunities.
Create 150 more tickets.
We have designed organizations in which creating work is easier than deciding whether the work matters.
That was already a problem. AI is about to make it explosive.
Work can now be created at machine speed
Most backlog systems were designed for a slower world.
Discovering a problem took effort. Analyzing information took time. Writing a detailed proposal naturally filtered out some of the noise.
That friction has disappeared.
AI can transform one sentence into an initiative, divide it into epics, generate user stories, add acceptance criteria, identify risks, and propose test cases before a human team has decided whether the original idea was any good.
Our ability to generate possible work is becoming almost infinite.
Our ability to determine what is valuable, safe, and necessary is not.
This is not merely a capacity problem.
It is a leadership problem.
Capture is not commitment
Modern organizations operate as though recording something creates an obligation to complete it.
It does not.
A customer request is not a commitment.
An idea is not a priority.
An alert is not an incident.
A recommendation is not a decision.
A possibility is not a promise.
The moment something enters a backlog, however, it begins to acquire moral weight. Someone asked for it. Someone noticed it. Someone might eventually ask why it was never completed.
So it survives.
Deleting it feels like rejecting the person, ignoring the risk, or admitting that the organization never intended to act.
The backlog grows not because every item remains valuable, but because adding work feels responsible while removing it requires judgment.
Eventually, it stops being a plan.
It becomes an archive of decisions leadership was unwilling to make.
“Agents will take care of it”
To some extent, they will.
Agents will investigate findings, compare information, implement routine changes, run tests, and close work without involving people.
That is valuable.
But agents do not make organizational capacity infinite. They move the constraint.
Someone must still decide what agents may change, which goals they should optimize, how conflicting priorities are resolved, and which consequences are acceptable.
Agents can execute good decisions at extraordinary speed.
They can also execute outdated assumptions, confused priorities, and unnecessary work at extraordinary speed.
And if an agent can safely detect, resolve, verify, and close something without human involvement, perhaps it never needed to become a visible backlog item. It was machine activity not a human obligation.
The danger is agentic bureaucracy: one AI generates work, another prioritizes it, a third implements it, and a fourth reports impressive throughput.
People may never see the work until they must explain why it happened, repair what it harmed, or accept consequences nobody consciously chose.
The system will look intelligent.
The organization may still have no idea what matters.
Your backlog may be protecting leadership
Creating a ticket is psychologically comforting.
The concern has been documented. It has an identifier, a status, and perhaps somebody’s name in the owner field.
If something goes wrong, the organization can say:
“We raised the ticket.”
“We assigned it.”
“We discussed it.”
These statements sound responsible. But recording a problem is not the same as accepting responsibility for solving it.
Sometimes a ticket exists primarily to transfer accountability.
Security transfers the vulnerability to engineering. Customer support transfers the request to product. Leadership transfers the transformation action to a working group.
Everyone follows the process.
Nobody creates capacity.
The ticket becomes a parcel of institutional guilt passed between teams until it lands with someone who already has more work than time.
If it remains unfinished, that person appears to be the problem.
The system that manufactures unlimited demand escapes examination.
The people who care most carry the most
Backlog overload does not affect everyone equally.
The most conscientious people remember the waiting customer, the unresolved risk, and the action with an owner but no allocated time.
Because they are reliable, more work finds them.
Because they care, they struggle to ignore it.
Because they repeatedly rescue the organization, leaders underestimate the real cost of the system.
Eventually, urgent work becomes their identity.
They become the person who handles every incident, understands the old system, or catches whatever falls between teams.
Organizations call them heroes.
But when the same people must repeatedly become heroic, leadership should stop celebrating and start redesigning the work.
When one person is always fighting the dragon, do not praise their bravery without asking why nobody is rebuilding the village.
A backlog must be allowed to forget
A healthy backlog is not a warehouse for everything an organization has ever considered.
It should create clarity about what matters now.
That means:
❤️🔥 Separating historical evidence from active work
❤️🔥 Revalidating old items instead of preserving them automatically
❤️🔥 Making new priorities replace existing priorities
❤️🔥 Allowing requests, risks, and ideas to expire
❤️🔥 Keeping routine machine activity out of human work queues
❤️🔥 Protecting capacity for prevention, not only urgent response
❤️🔥 Explicitly deciding what will not be done
Deleting an obsolete ticket is not deleting knowledge.
Closing a request is not rejecting a person.
Choosing not to act is not automatically negligence.
Sometimes it is leadership.
Because the real work is not processing everything that enters the system. It is distinguishing signal from noise, choosing what deserves attention, and protecting the capacity to do it well.
A growing backlog is not necessarily evidence of ambition.
It may reveal that your organization can collect but cannot choose.
That it can assign work but cannot create capacity.
That it can remember every fear but cannot release one.
So ask yourself:
Is your backlog helping people see today’s reality or preserving every fear, idea, and possibility your organization has ever recorded?
If it is the latter, do not demand that people process the pile faster.
Do not place another AI on the conveyor belt.
Stop creating obligations without choices.
Wake the dragon.
Look beneath the mountain.
And redesign how work enters the system. 🐉🔥
Very well said. I appreciate the thoughtfulness and the insight. It is very helpful to reflect on these things.