Project date: 1 May 2024
Project link: https://votespeaker.civita.co.uk/
Problem Addressed
Time-poor activists and campaigners everywhere face a trade-off between depth and breadth when it comes to political communications. They can spend the extra time to craft tailored narratives that reflect the particularities of specific audiences (and therefore lack broad relevance), or opt for more generalised boilerplate messaging and send less resonant messages with more reach.
VoteSpeaker serves to lessen the bite of that trade-off by using the power of AI to provide activists with political communications that are both quick to produce and tailored to the audience it seeks to persuade.
Approach & Implementation
Developed by Campaign Lab volunteer Jethro, VoteSpeaker leverages AI technology to generate efficiency gains without losing the human touch of political communication.
The first step is to enter your own details and those of the target audience , the most important of which are voter ID confidence and voter segment (VoteSpeaker can use a range of different typologies, including Mosaic and More In Common segmentations).
Having done so, the campaigner specifies the desired tonal style and political voice of the letter before inputting their core campaign message:
After that, VoteSpeaker does the rest:
The final output integrates several key elements: the user’s political framing based on local knowledge, essential political demographic data about the target audience, and an appropriate communication style tailored to maximise engagement.
Resources
-
- Code repository: https://github.com/Blue-Cardigan/votespeaker
Contact: [email protected]
7 May 2025