Project Managers Are Drowning in Updates. So We Built AI to Help.
Project management is not broken because PMs are bad. It is broken because too much of the job has become manual. We built AI into our platform so PMs spend less time chasing updates and more time leading the project.
Genti Osmanaj
Head of Tech
Project management is not broken because PMs are bad.
It is broken because too much of the job has become manual.
- Chasing updates.
- Checking task statuses.
- Writing summaries.
- Preparing reports.
- Following up with developers.
- Checking budgets.
- Checking deadlines.
- Checking who is overloaded.
- Checking if a client request is still inside the scope.
- Checking the same thing again the next day.
At some point, the Project Manager stops managing the project and starts managing the noise around the project.
At Tetbit, we wanted to change that.
Not by removing the PM. That would be a mistake.
Projects still need people who understand clients, priorities, context, trade-offs, quality, and timing.
But we also believe PMs should not spend half of their day collecting information that already exists somewhere inside the system.
So we started building AI directly into our internal platform.
AI Should Not Replace the PM. It Should Remove the Repetitive Work Around the PM.
There is a lot of talk about AI replacing roles. We do not see it that way. At least not for project management.
A good PM is not just a person who moves tasks from one column to another. A good PM understands the client, the team, the deadline, the budget, the business goal, and the pressure behind every decision.
AI cannot fully replace that. But AI can help with the parts of project management that are repetitive, slow, and easy to miss.
For example:
- What changed in this project today?
- Which tasks are blocked?
- Which team member has too much work?
- Which project is going over budget?
- Which deadline is at risk?
- Which client request needs escalation?
- Which tasks need a follow-up?
- What should be included in the weekly update?
These are not questions that should require manual digging every time. The system already has the data. AI helps turn that data into useful answers.
AI Project Summaries
One of the first things we worked on was project summaries. Every project has a lot happening inside it.
- Tasks are created.
- Tasks are completed.
- Comments are added.
- Files are shared.
- Bugs are reported.
- Deadlines move.
- Priorities change.
The problem is that this information is usually scattered. The PM has to open the board, read through tasks, check comments, ask the team, compare what changed, and then prepare an update for the client or leadership. That takes time.
So we built AI summaries into our internal platform. Now the PM can quickly understand what happened in a project without manually reviewing every detail.
The AI can summarize:
- What was completed
- What is still in progress
- What is blocked
- What changed since the last update
- What needs attention
- What should be communicated to the client
The PM still reviews the summary. That part matters. AI prepares the first version, but the PM adds context, adjusts the message, and decides what should actually be sent.
AI saves time. The PM keeps control.
Daily Standup Support
Daily standups are useful. But they can also become repetitive very quickly.
- What did you do yesterday?
- What are you doing today?
- Are there any blockers?
The answers are important, but collecting them manually every day becomes a burden.
So we started connecting daily standups with the actual work inside our platform. Instead of relying only on manual updates, AI can help prepare a first version of the daily status based on task activity.
It can detect what moved, what did not move, where comments were added, what was completed, and what seems stuck. This gives the PM a better starting point.
Instead of asking everyone for the same basic information, the PM can focus on the real questions:
- Why is this blocked?
- Do we need to change the priority?
- Does the client need to be informed?
- Is the estimate still realistic?
- Does someone need help?
Workload Balancing
One of the hardest parts of project management is knowing when someone is overloaded. Most teams notice overload too late.
- The task is already late.
- The developer is already stressed.
- The designer is already blocked.
- The PM is already explaining delays to the client.
We wanted to catch this earlier. So we built workload visibility into our internal tool and connected it with AI.
The system looks at active tasks, complexity, estimates, deadlines, sprint capacity, and team availability. This helps the PM see when workload is not balanced.
For example:
- One developer may have fewer tasks, but they are all complex.
- Another person may have many tasks, but they are smaller.
- A designer may look available, but is waiting on feedback from three clients.
- A marketing person may have too many client deliverables in the same week.
AI helps identify these patterns faster. It can suggest where the pressure is building and where work may need to be reassigned. The final decision still belongs to the PM and Team Lead.
AI can see the workload. People understand the context.
Budget Tracking and Early Escalations
A project going over budget is rarely a surprise if you are watching closely. The signs are usually there.
- Too many unplanned tasks.
- Too many revisions.
- Too many bugs.
- Too many unclear requirements.
- Too much time spent on something that was estimated too low.
The problem is that teams often notice this when it is already too late. So we added budget tracking and AI-assisted escalations into our internal project workflows.
The system can compare estimated hours, logged work, project progress, task complexity, and scope changes. Then AI helps flag when something does not look right.
For example:
- This project is using hours faster than expected.
- This task is taking longer than estimated.
- This client request may be outside the original scope.
- This project may need a budget discussion.
- This deadline may no longer be realistic.
- This issue should be escalated before it becomes a bigger problem.
This is not about blaming the team. It is about protecting the project.
Better Client Updates
Client communication is one of the most important parts of project management. But it is also one of the most time-consuming.
A good client update needs to be clear, honest, and useful. It should explain what was done, what is next, what is blocked, and whether anything needs the client's attention.
The problem is that writing these updates manually takes time, especially when a PM is handling multiple projects. So we started using AI to help prepare client updates.
The AI generates a first draft based on the project activity and structures the update clearly:
- What we completed
- What we are working on now
- What is waiting for feedback
- What risks exist
- What decisions are needed
- What comes next
But again, the PM stays in control. The PM reviews the message, adjusts the tone, removes unnecessary details, adds context, and decides what the client should receive.
AI helps with the draft. The PM owns the communication.
Scope Change Detection
Scope creep is one of the biggest hidden problems in software, design, and marketing projects. It does not always happen dramatically. It usually happens slowly.
- One small request.
- Then another.
- Then a "quick change."
- Then a new feature.
- Then a new page.
- Then a new flow.
Then suddenly the project is no longer the project that was originally estimated. We wanted to give PMs better support here.
So we started building AI-assisted scope change detection. The system can look at new tasks, client requests, comments, and project changes, then help identify when something may be outside the original agreement.
It does not automatically reject anything. That would be too rigid. But it gives the PM a signal:
- This may need scope clarification.
- This may require a new estimate.
- This may need client approval.
- This may affect the deadline.
- This may affect the budget.
Human-in-the-Loop Project Management
The most important part of our approach is simple.
- AI can suggest.
- AI can summarize.
- AI can detect.
- AI can warn.
- AI can draft.
- But people decide.
This is how we think AI should work inside project management. The PM should always stay in the loop, especially when the decision affects a client, a budget, a deadline, or a team member.
- AI should not approve a scope change alone.
- AI should not blame a developer.
- AI should not send sensitive client communication without review.
- AI should not make final budget decisions.
- AI should not replace leadership judgment.
But AI should help the PM see things earlier and prepare better. That is the balance. Automation where it saves time. Human judgment where it matters.
What Changed for Our PMs
The biggest improvement is not that AI does the PM's job. The biggest improvement is that AI reduces the operational noise around the PM.
- Instead of manually checking every project detail, the PM gets a clearer overview.
- Instead of writing every update from scratch, the PM starts from a structured draft.
- Instead of discovering risks late, the PM gets early warnings.
- Instead of guessing workload, the PM sees real pressure across the team.
- Instead of chasing information, the PM can focus on decisions.
Not replacing the PM. Making the PM sharper.
The Future of Project Management Is Not Fully Automated
We do not believe the future of project management is a fully automated AI system running everything alone. That sounds impressive, but it is not realistic for serious client work.
Projects are full of context. Clients change their mind. Priorities shift. Budgets matter. Quality matters. People have good days and bad days.
Some delays are technical. Some delays are communication problems. Some problems require a direct conversation, not another automation.
That is why the future is not AI instead of PMs. The future is AI-powered PMs. Project Managers who have better tools, better visibility, better data, better alerts, better summaries, better control.
At Tetbit, that is what we are building inside our own platform. Because we believe the companies that win will not be the ones that simply "use AI." They will be the ones that redesign their operations around it.
And project management is one of the best places to start.
