Our project management setup was a mess. And I don't mean "a little disorganized." I mean we had tasks scattered across three different Trello boards, milestone timelines living in a separate Gantt chart tool, time estimates in a Google Sheet that nobody updated, and status updates buried in Slack threads that were impossible to find after 48 hours. Every Monday morning started with the same ritual: "OK, what's everyone actually working on?"
It wasn't for lack of trying. We'd evaluated Asana, Jira, Monday.com, and half a dozen others. But the problem wasn't that we couldn't find a tool — it was that no single tool matched how our team actually worked. Every off-the-shelf solution forced us into someone else's idea of a workflow. Custom development? We didn't have a dev team to spare for weeks.
Then we tried something different. We built our own project management system on INFORMAT in under an hour. Here's exactly how we did it — including the parts that didn't go smoothly.
Why We Didn't Want Another Off-the-Shelf Tool
Before I get into the build, let me explain why this mattered. We're a small team — eight people across product, design, and marketing. Every SaaS project management tool we tried had the same problem: too much or too little.
Jira was built for engineering teams with sprint ceremonies. We're not all engineers. Trello was too simple — great for a to-do list, terrible for tracking dependencies or actual progress against deadlines. Monday.com was powerful but we spent more time configuring it than actually managing projects.
Here's what we actually needed:
- A single place to track tasks, milestones, and time estimates
- Custom statuses that matched our workflow (not someone else's)
- Simple enough that everyone would actually use it
- Something we could change as our process evolved
That last point was the killer. Every SaaS tool locked us into a data model. We wanted something that could flex with us.
The INFORMAT Approach: Describe, Don't Configure
Here's what surprised us about INFORMAT: instead of dragging and dropping widgets onto a canvas or wading through configuration menus, you just describe what you want. The platform generates the application — database schema, UI screens, business logic, everything.
We'd seen low-code platforms before. Most of them replace coding with an equally complex visual configuration system. INFORMAT does something different. It uses AI to interpret natural language descriptions and turns them into working software.
Was it perfect? No. But it was shockingly fast compared to every alternative we'd tried.
Our First Prompt: The Project Model
We opened INFORMAT and typed out what we needed. Here's the first prompt we used:
"Create a project management system where I can create projects, each with multiple tasks. Tasks need fields for title, description, status (To Do, In Progress, In Review, Done), priority (Low, Medium, High, Critical), assignee, estimated hours, actual hours logged, due date, and a comments thread. Projects need a name, description, start date, target end date, and status. There should be a Gantt-like view that shows tasks on a timeline, a Kanban board view, and a dashboard showing overall project progress. I also need role-based access so the project manager can see everything but team members only see their assigned tasks."
This is where we hit our first friction point. The initial output was close but not quite right. The Gantt view was generated as a simple list with date columns — not the visual timeline we'd imagined. And the dashboard showed raw numbers without any visual charts.
But here's the thing: fixing it took one more prompt, not a week of development. We just told INFORMAT what we wanted changed.
"The Gantt view is a good start but I need it to actually render tasks as horizontal bars on a timeline, grouped by project phase. Use different colors for different statuses. For the dashboard, add a pie chart showing task distribution by status and a bar chart showing estimated vs actual hours per project. Make the dashboard filterable by date range."
Ten seconds later, we had a proper Gantt chart and dashboard. Total time so far: about 12 minutes.
What the Generated System Looked Like
After those two prompts, INFORMAT had generated:
- A relational data model with projects, tasks, users, and time entries
- CRUD screens for creating and editing projects and tasks
- A Kanban board view with drag-and-drop status updates
- A Gantt chart view grouped by project phase with color-coded status bars
- A dashboard with pie charts (tasks by status) and bar charts (estimated vs actual hours)
- Role-based access control (PM vs team member views)
- Comment threads on each task
Let me be honest about what wasn't there: the UI was functional but not beautiful. The color scheme was the platform default, the layout was basic, and the mobile experience was usable but not optimized. If you're the kind of person who needs pixel-perfect design before you'll use a tool, you'll want to spend another 20-30 minutes on polish.
We're the kind of team that prioritizes function over form, so we moved on. But it's worth knowing that the trade-off for speed is that you'll want to iterate on the UI if appearance matters to your team.
The Time Tracking Module
Our original Trello-plus-spreadsheet setup meant time tracking was completely disconnected from tasks. Someone would log hours in a spreadsheet, and there was never a clear link between "this task took 6 hours" and "that's why the project deadline slipped."
We wanted time entries tied directly to tasks, with running totals that rolled up to the project level. Another prompt:
"Add a time tracking module. Each time entry should be linked to a specific task and include the date, hours spent, and a brief description of what was done. Show running total hours per task and per project. On the task detail view, show a log of all time entries. Add a weekly time sheet view that shows all team members' logged hours for the current week. Only the user who logged the time and the project manager can edit or delete time entries."
This worked on the first try — probably because the data model was already established and we were adding to it rather than reshaping it. The time sheet view was particularly useful; it showed us at a glance who had logged time and who hadn't.
Running time: about 25 minutes.
Where We Hit Real Problems
I promised I'd be honest, so here goes. We ran into three genuine issues:
1. The Notification System Was Too Broad
By default, every task update sent a notification to everyone on the project. The PM got flooded. We had to add a prompt to scope notifications: "Only notify the task assignee and the project manager when a task status changes. For comments, only notify people who have participated in the thread." Simple fix, but it was a miss in the initial generation.
2. Data Migration Was Manual
INFORMAT generated the system, but it didn't import our existing Trello data. We had to export Trello to CSV, then use another prompt to create a bulk import feature: "Add a CSV import function for tasks and projects. Map the CSV columns title, description, status, assignee, and due_date to the corresponding fields. Validate the data before importing." This added about 10 minutes. Not a dealbreaker, but worth planning for if you're migrating from another tool.
3. The Reporting Was Basic
The generated dashboard was fine for day-to-day tracking, but when we wanted a report like "which team members are consistently overestimating or underestimating their tasks" — something that required comparing estimated vs actual hours across multiple projects — we had to ask for it specifically. INFORMAT doesn't read your mind (yet). You need to know what reports you'll want and prompt for them.
Comparison: AI-Generated vs the Alternatives
Here's how the INFORMAT approach stacked up against the other options we considered:
| Criteria | INFORMAT (AI-Generated) | Off-the-Shelf SaaS (Jira/Monday) | Custom Development | Spreadsheets + Separate Tools |
|---|---|---|---|---|
| Time to working system | Under 1 hour | 1-3 days setup + configuration | 4-8 weeks (minimal viable) | Already in use (barely) |
| Customization | Prompt-driven, near-infinite | Limited to predefined fields/views | Complete control | Very flexible but manual |
| Cost | Platform subscription only | $10-30/user/month (premium features cost extra) | $15k-50k+ initial dev + maintenance | Low direct cost, high time cost |
| Team adoption | High (built for your workflow) | Medium (must adapt to tool) | High (built for your workflow) | Low (scattered, manual) |
| Iteration speed | Minutes per change | Days to weeks (process changes) | Days to weeks (dev cycles) | Immediate (but manual) |
| Integration capability | API access available | Extensive integrations (but may need extra cost) | Custom (anything possible) | None (manual copy-paste) |
| Mobile experience | Functional web mobile | Dedicated mobile apps | Depends on scope | Varies by tool |
| Data ownership | Full (exportable) | Vendor lock-in (difficult migration) | Full | Full (but fragmented) |
The biggest surprise for us was the iteration speed. When we realized we needed a "blocked" status halfway through our first week using the system, we fixed it with one sentence. "Add a Blocked status between In Progress and In Review." Done. No ticket, no sprint planning, no deployment pipeline. It was live immediately.
"We spent three months evaluating project management tools and never found one that fit. Then we built one in 45 minutes on INFORMAT. It's not that the SaaS tools are bad — it's that 'one size fits all' doesn't actually fit anyone."
— Our product lead, after the first week using the system
The Things That Surprised Us (Both Good and Bad)
The Good
Everyone actually used it. Because the system matched our workflow exactly — not some generic "best practice" workflow — there was no learning curve. People just started using it. Two days in, nobody was updating the old Trello boards anymore.
Changes were instant. When our design lead asked if we could add a "design review" step between "In Progress" and "In Review," we just told INFORMAT. It took longer to type this sentence than it did to make the change.
The Bad
Mobile is adequate, not great. The mobile web experience works, but it's clearly a web app rendered on a phone screen. If your team lives in the field or does most of their work on phones, this matters. There's no native mobile app.
Advanced features require deliberate prompting. Things like automated reminders for overdue tasks, time-based triggers, or complex cross-project rollups don't appear unless you ask for them explicitly. The platform does what you tell it to do — you have to know what you want.
Tips If You Want to Try This
Based on our experience, here's what I'd suggest if you're going to build your own PM system with INFORMAT:
- Start with the data model. Get the entities and fields right first. Everything else — views, reports, automation — builds on top of the data. We spent most of our first prompt on this and it paid off.
- Expect to iterate. Your first prompt won't produce a perfect system. That's fine. The key isn't getting it right on the first try — it's being able to fix it on the second. We went through about 8 iterations (prompts) before we had something we were happy with.
- Think about edge cases. The system INFORMAT generates handles the happy path beautifully. But things like "what happens when someone deletes a task with time entries?" or "can I reassign all of one person's tasks to someone else?" need to be prompted explicitly.
- Plan for migration. If you're coming from another tool, budget time to either build an import feature or manually enter your active projects. We underestimated this.
- Test with real data. We put 3 real projects into the system on day one. Within a week, we'd found and fixed about 10 things that didn't work the way we needed. Real usage surfaces real issues.
Final Verdict
So did we replace Trello, spreadsheets, and Gantt charts entirely? Yes. We're about six weeks in now and the old tools are gathering dust. The system we built isn't as polished as Jira or Monday.com — the UI is simpler, the mobile experience is basic, and we've had to prompt our way through a few unexpected scenarios. But it works the way we work, not the way some product manager at a SaaS company decided a team should work.
The hour we spent building it has saved us far more than that in eliminated context-switching. No more hunting through Slack for status updates or maintaining a separate time tracking spreadsheet. It's all in one place, and when we want to change something, we just describe the change.
That's the part that's hard to convey until you've tried it. Traditional software — whether SaaS or custom-built — treats change as expensive. New feature? That's a ticket, a sprint, a deployment. INFORMAT treats change as cheap. Want to add a field? Done. Want a new view? Ask for it. Want to restructure your entire workflow? Couple of prompts and you're there.
For a team of eight trying to get work done without drowning in process overhead, that's been the difference between a tool we tolerate and a system that actually helps.