Whether it’s federal elections, local council elections, board of directors elections in large corporations, or any other large (or even smaller) format elections, processing ballots presents some challenges and inconveniences. Anyone who has ever been called upon to sort ballots in referendums knows how miserably long it takes and how tedious it is (especially if you were out too late the night before…). If at the same time there are municipal, cantonal and federal referendums to be held, the day is over. In this article, you can read that the processing of ballots nowadays also works differently, namely to a large extent automatically, and what various advantages this has to offer.

In times of COVID pandemic, absentee voting only increases

Parashift Live-Webinar

Various aspects and challenges in elections

The goal of an election is to be able to process all ballots in the shortest possible time in accordance with the requirements and in a legally secure manner so that the result of the election can be announced as quickly as possible. However, as we have seen most recently with the 2020 U.S. elections, the reality is significantly different, and can hardly be called a best-case scenario. The problems that arise not only in the U.S., but in all major elections, are many and complex. In addition to security issues and the already immense logistical challenges, the COVID pandemic has added another significant one: Far fewer ballots are being cast in person at polling places, and many more voters are voting by absentee ballot, leading to delays even just to receive them, not to mention processing them.

Security is guaranteed by the automated capture and processing of the ballots – which cannot necessarily be said of manual processing

The manual processing of ballots and its problems

Once in the counting centers, things could theoretically move quickly. However, manual processing directly puts a stop to this. Ballots must first be roughly presorted. The first step is to define whether the ballot is valid, invalid, or questionable. This already requires multiple checks by numerous people. After the rough presorting, further distinctions have to be sorted:

  • Checkboxes
  • Listed names
  • Name lists completed by hand
  • Signatures

Accuracy, speed, low error rate – all this and much more is promised by intelligent, AI-based OCR for ballot processing

Automated processing of ballots

The fact is that elections are always extremely time-critical and data-sensitive, but at the same time they want to be very results-oriented. In order to meet the high demands of elections, some important points must be followed in the processing of ballots:

  • The process must be transparent, secure, and accessible to ensure auditable elections
  • The solution must be easily integrated into the existing election management system
  • Ballots must be verified at multiple stages and pass through all pre-defined rules
  • The solution must be easily scalable based on volume

All of these requirements can be achieved with the integration of intelligent, AI-based OCR (Optical Character Recognition) into the election management system. Here’s how automated ballot processing with intelligent, AI-based OCR works at a glance:

  • Automated receipt and recognition of ballots, by mail or online
  • Quality improvement and page separation, which is highly relevant for ballots
  • Automated classification, for example in the case of multiple elections at the same time
  • Double verification, which is essential when processing ballots and for the security aspect
  • Automated data extraction from the ballots; handwritten or electronic, name lists completed by hand, checkboxes ticked, and so on and so forth
  • Structured output of the extracted data from the ballots

With the automated and intelligent processing of ballots, a high cadence can be targeted and thus an attractive ROI achieved

What works for city and state elections also works for board elections and the like


One of the big advantages of intelligent, AI-based OCR is the ease of scalability and thus the high volume of ballots that can be processed in a time-critical manner. This makes it perfect for solutions of state elections with several hundred thousand voters, but at the same time, especially large corporations benefit from the integration of intelligent, AI-based OCR, for example, in board elections and general meetings. Furthermore, the automated processing of ballots is significantly less prone to errors and thus the risk of fraud is also considerably reduced.