GanttGenerate
Turn a sentence, a drawing set, a scope document or a contract into a detailed, logic-linked schedule, in minutes rather than weeks.
About us
Someone has to answer, so what’s the plan? We build for them.
Whoever signs off on the schedule carries it for the length of the job. GanttAI exists so that plan can be built from whatever the project already knows, checked before it goes out, and read by everyone who has to work to it.
What we do
We help teams plan projects better
GanttAI generates, verifies, reviews and compares construction schedules using AI trained on how projects are actually delivered, not on generic project-management theory.
Verify
Run any schedule through 36 checks, the DCMA 14-point assessment plus our own 22, to find what’s missing, unrealistic or out of sequence. Your company’s own rules run on top.
Review & compare
Open someone else’s schedule, whatever format it arrived in, pull out the activities in your scope, and see every one that moved, stretched, shrank or disappeared between revisions.
GanttAI Agent
Ask anything about the schedule in plain English and get an answer you can act on, with the working file to go with it. It shows the reasoning behind the answer, so you can check it rather than trust it.
Zooming out
Every building is coordinated by a schedule and often it cannot be understood.
A project is dozens of separate companies with no shared system between them, and the one artefact coordinating all of it is the schedule: when money moves, when crews mobilize, who is liable when something slips. It’s still written by hand, by a few specialists, in tools designed decades ago. Everything downstream inherits that.
01
Make the schedule better
It should come from the job in front of you, not a template of one someone else built. Describe it, and add the contract, drawings or scope if you have them. Then 36 checks run before a human sees it.
02
Make it readable
A plan only coordinates people who understand it, and today that is a handful per project. Ask in plain English, get an answer out of a 4,000-line file. The asset was never the file. It’s everyone knowing the same thing.
03
Fit the way jobs already run
The industry runs on P6, MPP, Excel and PDFs and won’t stop because a tool asked it to. We read and write what companies already send each other: no new process, no retraining, no waiting for every trade to join one platform.
Each part stands on its own. Together they change how many people on a project actually understand the plan they are working to, which is the part that matters.
Want to get involved? →
Building in this space, operating in it, or backing it: the same form takes all three.
What we believe
Three things we won't compromise on
Everything else is a trade-off we are willing to argue about. These three are not.
Your data stays yours
Schedules, drawings and contracts are private and never exposed. Nothing is pooled with other customers, and nothing trains anyone else's model. A competitor can’t benefit from your job because you ran it through us.
You understand your own data
Every activity carries a confidence value and shows the reasoning behind its date. If the model isn’t sure, it says so. You should never have to take an output on faith, or defend a number in a meeting you can’t explain.
It works the way you already work
Construction software has a long tradition of asking the job to change shape around it. We meet you where you are: the formats you already send, the process you already run, the words you already use. If using GanttAI means retraining a team, we have built the wrong thing.
Who’s behind it
Chirag Soni
Founder
“An engineer and PMP with 14 years on site, most of it running schedules on portfolios worth billions. I have built them by hand under deadline, defended them in the OAC meeting, and watched them go stale the week after issue. GanttAI is the tool I wanted on those jobs: a plan the whole project can read, not a document a handful of people can.”
Where we are
Based in Toronto, Canada
We build GanttAI out of Toronto and test it on jobsites across North America. Built by people who have spent their lives on site.
Let’s collaborate
Build this with us.
Small team, real customers, no layer between the people writing the code and the people running jobs with it. Both roles are part-time and based in the Greater Toronto Area. If neither fits but the problem does, write anyway.
Business Development Representative
Greater Toronto Area · Part-time
You’d be the first commercial hire, which means the job is to find out who needs this most and why, not to work a script someone handed you. Talking to general contractors and trades, sitting in on demos, and bringing back the objections that change what we build next.
- Comfortable cold-starting a conversation with a project manager
- Can hold your own on a jobsite and in a sales call
- Construction or ConTech experience is a real advantage, not a requirement
Data Scientist
Greater Toronto Area · Part-time
Construction schedules are messy, inconsistent and rarely labelled, and the interesting work is turning them into something a model can actually learn from. You’d own extraction quality, duration and logic prediction, and the confidence values users see on every date.
- Strong Python, and real experience with models in production
- Happy working where the data is dirty and the ground truth is arguable
- NLP or document-extraction background is a plus
Let’s solve built-world problems →
Tell us what you would want to work on. A resume helps, but it isn’t the point.