AI startup to dramatically improve truck scheduling logistics
There is dire need to improve the operational complexity of third-party logistics (3PL) and freight brokerage
providers by automating data gathering, analytics and creating a fast decision matrix. Day-today operation of
any 3PL requires middlemen to constantly intervene and manage the scheduling process. While doing that, these
companies can employ hundreds, if not thousands, of operations employees to input the data and then many several
brokers who put these deals together every day.
Here’s a typical daily work flow : 3PL sales staff everyday keeps in touch with the shippers who want to move loads on a particular day,
and at the same time information about the empty trucks is also collected and entered into the logistics software. This software will generally
do the matching of truck availability, locations and the shipper requirements. But much of the work here is mindless data entry, and it’s easy to
see why we think there is an opportunity to do things more efficiently.
There is also tremendous scope to use IOT devices that can track the equipment using GPS and transmit using cellular or satellite networks.
AI implementation on Medical Image Analytics
We think that augmenting existing analytics software using AI tools, such as Machine Learning techniques like convoluted neural network,
decision trees and deep learning would allow further optimization and classification to assist decision making for physicians and researchers.
Opportunities to develop computational methods and algorithms using latest AI technologies would help in faster analysis of biomedical data.
AI and AR with HoloLens for Product and Process development
We are envisioning a startup in optimizing business data in 2D and 3D, using AI technologies to create the augmented and virtual reality solutions. The opportunity for
machine learning is to identify each real object, learn it and then place the virtual object from the database that matches the current environment conditions.
There are various possibilities of accomplishing this, such as converting the real object into binaries and instantly comparing it with the database and then fetching it. Challenges are also on the object size, as the services
are written with some limitations on traction size .. so when objects are brought in (form of binary) we need to restrict the size, or services may fail.
There are many challenges in precisely matching the conditions of the real objects with virtual objects …