A pilot is built to get you to value fast. Your team uploads the data package in one to two hours, your engineers work with Imubit's builders on your own data, and each step hands you something you can use. You leave with a working open-loop model, a map of your biggest gaps and a plan to close them.
We judge every pilot by how soon you capture value, and we work to shorten that on every one. Nothing waits for the end. Each step leaves your team with something it can use.
Technology is easy to come by now. Knowing what to do with it is the hard part, and your team already knows your plant.
This is not a hands-off process. Your engineers build alongside ours, in DLPC® Builder or from the AI tools they already use, and they keep what they build.
You choose the problem you most want solved, and we agree one success test with you, so everyone knows what value means before work starts.
Your team prepares and uploads the data package in one to two hours, as it is: our engineers screen and clean it in the platform. Nothing is installed and no workshops are needed, so work on your data starts straight away.
We trace where the plant falls short of the plan, follow each gap to the recurring decision behind it, and rank the gaps by value, so the first model goes where it is worth most.
Your engineers and our builders build a working open-loop model of the unit. It is tested on data it has not seen, your engineers certify it, and it runs as advice before the pilot ends.
The same model opens other questions: a soft sensor for a late lab result, one target across several units, or a what-if before a turnaround.
You keep the map, the model and a plan to close the gaps. Closing the loop on the unit is the next step, inside the envelope your engineers set and on top of your APC and DCS, so the value keeps building.
Built on your data, tested on data it has not seen and certified by your engineers. It runs as advice.
Where the plant falls short of the plan, the recurring decision behind each gap, and what each is worth.
Which decisions to model next, which to take to closed loop, and in what order.
What you build on the models is yours. Closing the loop on a unit is the step after the pilot, under a subscription.
A few of the problems current pilots are working on. The same approach runs across refining, chemicals and polymers, cement, and mining and minerals.
RefiningThe pod builds A site-wide map of the gaps between plan and plant, and a calciner model that trades coke burn against fuel gas and steam while it holds the coke density specification.
CementThe pod builds Strength models for each curing age, from the chemistry and fineness the lab already measures.
See the applicationCalciningThe pod builds A density target for every kiln that meets the site's blended specification with the least gas, from the models already running each kiln.
See the application
Oil productionThe pod builds A model of separation efficiency against oil rate and chemical injection, built from the separator's own history.
Prepare and upload the data package, which takes one to two hours, and agree one success test at the start. Your engineers then work with Imubit's builders on your own data.
No. There are no workshops and nothing to install. Work on your data starts as soon as the package is uploaded, and the model runs as advice during the pilot. Nothing writes to your plant.
Into the Imubit platform, which runs on AWS and Google Cloud and encrypts all data at rest and in motion. Imubit has completed a SOC 2 Type II examination, and the report is available under NDA. When a unit later moves to closed loop, its controller runs on-premise inside your plant network.
Data security policyYou keep a working open-loop model, a map of your biggest gaps and a plan to close them. What you build on the models is yours. Closing the loop on a unit is the next step, under a subscription.
Tell us the problem you most want solved, and we will show you how a pilot gets you to value on it.
