For this business automation case, we engaged a whole, standalone section of a large company’s business. Many of its most critical tasks, like customer delivery transactions, were still being run with paper, with widespread delays and losses in recognizing revenue and collecting cash. The project achieved automation of the point of sale activities with a delivery transaction and customer management app able to be run both on mobile devices and on computers. To this we added to this a management and reporting application to take care of all support functions.
The application and consequent change in processes has completely eliminated this.
Built a standalone app to support customer management and transactions, using a mobile stack which allows it to be deployed on mobile devices as well as being used on browsers. The front end is a Javascript/Ionic one, while there is a relational database on the back end to allow for easy search and reporting and to store all activity and customer information specific to this business. This is all information that drivers used to collect and manage on paper, so not only has this eliminated paper transactional activities, it has also simplified the way drivers track information about their customers.
The app operates independently of transactional systems and can run key functions offline. This has two advantages: the system can operate faster than it otherwise would if it had to connect to backend business systems and it can be developed independently of a major integration project. It has been successfully deployed on several cloud platforms, including Heroku and Azure, providing a further level of flexibility for a business which may want to migrate its applications between different cloud providers from time to time.
Integration is done through batch processes which output key info to transact business once a day and loads completed transactions at day’s end. These processes can be automated either through standard input and output capabilities for an ERP system or by using RPA-like functions to automate extraction and loading of data. While this approach clearly does not allow for real time visibility of transactions in all systems, this was not critical for the problem being solved here. On the other hand, avoiding direct integration massively simplifies a project like this, allowing us to focus on usability and speed of the application without needing to worry about integration points and the constraints they would impose on the system.
The project was delivered from concept to final handoff in just under a year. This included six weeks of initial development for an MVP, after which the system was deployed for three users initially, then a whole subfleet, over a period of four months. The software was released weekly through this period with updates to functionality and fixes to bugs, growing to the point where it was able to smoothly support the business. At that point the project went into full rollout and deployment to all users; at the same time more features were added as possible, in a less disruptive change mode for users, with releases happening every two to four weeks. It has been in full production for over two years with no software issues at all.
