For years, software teams have relied on estimation and project planning to answer a simple question: When will this be done?
The reality is, that question has never had a reliable answer.
Long before AI, traditional project planning assumed we could define scope, estimate effort, and execute against a plan. In practice, that rarely happened. Estimates were off. Plans shifted. Teams adjusted as they went.
AI is making that harder to ignore.
The reality is, that question has never had a reliable answer.
Long before AI, traditional project planning assumed we could define scope, estimate effort, and execute against a plan. In practice, that rarely happened. Estimates were off. Plans shifted. Teams adjusted as they went.
AI is making that harder to ignore.
Planning Was Always on Shaky Ground
Traditional planning depends on a chain of assumptions:
That chain breaks early in real-world software work.
Agile improved this by shifting focus away from precision:
Estimation didn’t go away, but it became less about accuracy and more about direction. Over time, teams developed consistent throughput, which allowed for forecasting at a higher level.
That consistency is what made Agile planning workable.
- Requirements are understood
- Effort can be estimated
- Timelines can be built from those estimates
That chain breaks early in real-world software work.
Agile improved this by shifting focus away from precision:
- smaller increments
- faster feedback
- continuous adjustment
Estimation didn’t go away, but it became less about accuracy and more about direction. Over time, teams developed consistent throughput, which allowed for forecasting at a higher level.
That consistency is what made Agile planning workable.
AI Increases Variability
AI hasn’t just made development faster. It’s made outcomes less predictable.
Take a typical piece of work, like implementing a feature or integrating with an API.
That work might:
You often don’t know which it will be ahead of time.
AI also introduces a new failure mode: it can produce output that looks correct but isn’t. Teams can spend significant time moving forward before realizing they’re off track.
The result is a wider gap between best-case and worst-case effort.
Take a typical piece of work, like implementing a feature or integrating with an API.
That work might:
- take an hour with the right prompt and approach
- take days of iteration, rework, or false starts
You often don’t know which it will be ahead of time.
AI also introduces a new failure mode: it can produce output that looks correct but isn’t. Teams can spend significant time moving forward before realizing they’re off track.
The result is a wider gap between best-case and worst-case effort.
This Is Really a Planning Problem
Estimation feeds planning. When estimates become less reliable, plans lose credibility.
Teams fall into a familiar trap:
This tension existed before, but AI amplifies it.
At some point, the problem stops being “how do we estimate better?” and becomes “how much should we rely on planning at all?”
Teams fall into a familiar trap:
- assume work will go quickly → miss expectations
- assume it will take longer → appear inefficient
This tension existed before, but AI amplifies it.
At some point, the problem stops being “how do we estimate better?” and becomes “how much should we rely on planning at all?”
Not Everything Got Faster
AI accelerates parts of development, but not the full system.
Teams still spend time on:
These don’t compress at the same rate as coding.
So while execution speeds up, delivery doesn’t always follow. Time shifts into areas that are harder to predict.
Teams still spend time on:
- understanding requirements
- coordinating across teams
- validating solutions
These don’t compress at the same rate as coding.
So while execution speeds up, delivery doesn’t always follow. Time shifts into areas that are harder to predict.
Consistency Still Matters
Agile has always relied on consistency… not in estimates, but in throughput.
Teams that work at a steady pace can forecast using averages over time. That’s why Agile discourages overtime and deadline-driven spikes. They create burnout, instability and make forecasting worse.
AI doesn’t change that.
Individual tasks may vary more, but over time, patterns still emerge. Some work will be fast. Some won’t. A steady system still produces usable signals.
That’s what allows forecasting to exist at all.
Teams that work at a steady pace can forecast using averages over time. That’s why Agile discourages overtime and deadline-driven spikes. They create burnout, instability and make forecasting worse.
AI doesn’t change that.
Individual tasks may vary more, but over time, patterns still emerge. Some work will be fast. Some won’t. A steady system still produces usable signals.
That’s what allows forecasting to exist at all.
Where This Leaves Us
You still need planning. You still need estimates.
Over time, the work tells you the truth. You see how much gets done, how steady the pace is, and where effort expands or contracts.
That’s always been the Agile mindset.
AI increases the spread between fast and slow work. It doesn’t change the need for steady progress and real feedback.
Teams that lean into that will stay predictable. Teams that rely on upfront precision will keep chasing it.
Over time, the work tells you the truth. You see how much gets done, how steady the pace is, and where effort expands or contracts.
That’s always been the Agile mindset.
AI increases the spread between fast and slow work. It doesn’t change the need for steady progress and real feedback.
Teams that lean into that will stay predictable. Teams that rely on upfront precision will keep chasing it.
About the Author
Andrew Anderson is the President of Latitude 40 and a seasoned technology leader with over two decades of experience in software development and process improvement. He helps organizations achieve operational excellence through practical, low‑risk strategies that deliver measurable results. His work combines technical expertise with a commitment to agility, guiding teams toward smarter solutions and sustainable growth.
About Latitude 40
Latitude 40 works with companies that are tired of over-engineered solutions and unreliable plans. We build custom software and help teams move away from rigid project thinking toward more adaptive, reality-driven execution. Our focus is simple: reduce risk, improve flow, and help organizations deliver meaningful results without unnecessary complexity.
If you want to see what a fast, responsible start could look like for your organization, we would be glad to walk through a practical first step.
If you want to see what a fast, responsible start could look like for your organization, we would be glad to walk through a practical first step.
RSS Feed
