QuikCal: Construction Site Scheduling Application
QuikCal is a scheduling application that uses image recognition and AI to convert construction site job whiteboards to a google calendar. Many construction companies use whiteboards to organize deliveries of building materials to jobsites. However, these whiteboards are messy, are not remotely accessible, and don’t support a log of deliveries to a jobsite.
Moreover, existing enterprise grade project management software is expensive and often toocomplicated to be quickly and effectively utilized by construction workers on a jobsite. QuikCal addresses this problem by using SMS messaging to eliminate the need for complicated software to digitize a whiteboard. A user can take a picture of a jobsite whiteboard, text it to a project phone number and our system will ask them follow u
p questions (like a personal secretary) and finally add the events to a google calendar which can be shared with site administrators and contractors. Our solution became a finalist in the 2022 Pratt Student Design fair. Since then, we have tested our application on three different jobsites through working with the SKANSKA construction company.
Adding AI Technology:
Based on our first couple of trials, we noticed that the most significant challenge with our system was making sure data we received from contractors was correctly formatted so our application could present it on a calendar. To solve this problem, we developed a new version of our software that uses generative AI to interpret user input from contractors and sort that information into the form template. By using AI to handle “edge cases” (i.e. misspelled words or incorrectly formatted dates) and having it intelligently correct/interpret this information, we have found that the textbot is much simpler to use and much more versatile than before. We can now handle full paragraphs descriptions of deliveries in plain text and our system can parse out the relevant information.
The current version can employ image recognition as well as natural language processing AI, so that a user can take a picture of a work order form and text it to our phone number, and our software will automatically read the text and fill out a form (it can then ask the user to correct any mistakes before submitting to a calendar). We are also integrating our software directly with Google Calendar so that foremen and site admins will be able to add our calendar directly to their google calendars on their phones (meaning that only one site admin needs to use our website to configure the project/approve deliveries).
History:
Through Duke EGR 101 and EGR 102 engineering design classes, I worked with small team of Duke students to develop custom web/mobile app for SKANSKA Construction Company to schedule deliveries of building materials to construction job sites. Our application aims to replace job site planning whiteboards with a digital solution. Many construction sites have been slow to adopt enterprise-grade planning software due to the complexity of existing solutions. Our solution combined an online calendar with a “textbot” chat assistant to make scheduling a delivery or checking a delivery schedule accessible for contractors who don’t have access to, or are unfamiliar with web apps. Our solution can be managed using a web application to allow contracts and site administrators to view and update the job calendar, but can also be fully accessed without an internet connection using only SMS.
• Finalist in Pratt School of Engineering 2022 Student Design Fair
• System built using HTML, CSS, node.js, and react.js. Developed companion Twillio text-bot to make scheduling deliveries and accessing delivery schedule possible over SMS (without internet access)
• Piloted Application at three different active SKANSKA job sites. Met with site managers to collect user feedback.
Original Beta Application:
