For working with tables, Microsoft Excel is easy to use and flexible as compared to database systems. For this reason it is widely used when manual processing of information is needed e.g., for tasks like data cleaning, checking, input, gathering from different sources, etc. These tasks are hard to automate using basic functionality of Excel, as it is tailored to numeric calculations.
For decades that Excel has been a de-facto standard for working with tabular data, numerous attempts have been made to address this problem. As an end user dealing a particular routine task, like phone number formatting, you can search and try different Excel add-ins that claim to perform a task you need. Alternatively, you could get a programmer to write a VBA script, which is known to be painful task in itself. Both approaches may require quite some time, which makes them infeasible for relatively quick (less than a workday) tasks. Additionally, both approaches require a good formal understanding of what exactly has to be done, which can be problematic with tasks like "highlight wrong addresses" or "see if any values are off".
In the age of IBM Watson and Siri, it shouldn't be like that. Users should be able to show example tasks to the system and get it to mimic them.
comnsense is a tool that runs in the background of Excel and recognizes patterns in user's actions and data. If it is able to recognize a pattern in actions, it "autocompletes" all values it can deduce. When there is a pattern in data, it highlights the values that are unlikely. If a user does not agree, they can just press the "Undo" button to remove the suggestions, just like with an Excel Autocorrect feature. comnsense will learn from this occurrence and never bother them on a similar occasion.
The system operates on a level of string and number statistics and is not task- or language-specific. Our prototype was validated on typical contact data tasks. Please check our website for intuitive examples!