AI Journalism

Helping sports journalists to create real time match reports using AI insights to better understand complex statistics during tennis events  

Overview

Today sports journalists report on multiple matches during an entertaining grand slam event, they are dependent on understanding the matches statistics to create a story that is absorbing but more importantly statistically accurate. Under pressure to publish reports and articles first to the fans we have created a AI assisted product to bring the insights to the media team in real time.

Role

- User research

- Workshop facilitation  

- User interviews

- UX Design 

- Prototyping

Duration

1.5 months

• It was found the social media team updated to their channels of each set of the important matches and therefore creating a statistic graphical download of these snapshots in the match was an asset   

Broadcasting restrictions were found to be a limitation to accompany some of the statistic insights that were generated, moving towards graphics to represent a majority of these

• Users wanted guidance on the the top suggested statistics that they could review straight away

• The importance of comparing previous performances of players, how their statistics are changing through the tournament was a main part of journalists articles 

The challenge

Infosys hold the data partnership with the Roland Garros for the tournament, providing statistics on all matches throughout the tournament. With access to this data combined with video assisted match highlights there is an opportunity to further assist the RG team when creating match reports, articles and content for their site and social media channels.

• Today journalists email infosys technical team for advice and suggestions on key match insights, time consuming to be completed manually

• The media team follow multiple matches a day and struggle to stay ontop of these in order to report accurately the pivotal parts of the matches analysed at a deeper level

• Journalists have up to 20 minutes to publish these reports so any time saved in creating content is a major help for overall quality

Interview & Testing insights

Report templates

We gave journalists the opportunity to select from a variety of templates that are formatted based on the topic they are writing about for example a match report is very statistics based compared to an interview or tournament feature which is imagery led.

Live match downloads

Screenshot 2020-03-10 at 13.53.52.png

A feature to be used by the social media team to download key statistics during the match, using preselected template imagery for each player. These can be filtered to compare statistics for each player and during end of set evaluations.

AI insights 

Each match generates insights once completed, these are categorised clearly for the user to navigate between, users can view the video highlight of the insight to further understand this statistic. Graphics are created along with copy describing the insight for the user to download directly or add to a larger report to later edit.

AI insights 

Screenshot 2020-03-10 at 13.56.51.png

Editable master reports

Several of these insights are added to a larger report which are saved in 'Your reports' here the user can edit this template to add further copy and narrative  and can customise components to suit their preference. The report can be downloaded or shared to the teams back office to be checked by editors before publishing.

Follow multiple matches

Several of these insights are added to a larger report which are saved in 'Your reports' here the user can edit this template to add further copy and narrative  and can customise components to suit their preference. The report can be downloaded or shared to the teams back office to be checked by editors before publishing.

Main takeaways

 6

User interviews

8

Usability tests

 4

      User journeys

• Requesting to speak to the wider team during initial user research engagement in order to find out what the full functionality of this product could be 

• Questioning assumptions of key stakeholders and how the product initially was pitched, this was challenged and through user centricity and coherent playback of findings attitudes were changed

• Understanding the technical environment the product will work within fully in order to avoid wasted design effort for integration proposals, getting the right people in the room early on