Telia
Telia, the "Tone of Voice" generator
Telia's brand voice was fragmenting across writers, languages, and channels. TRY built an AI tool on Azure GPT-4 and React that checks written content against Telia's brand guidelines, offering automatic corrections and adaptive guidance by content type, so copy reads consistently everywhere it's written.
Inconsistent brand voice across many writers, languages, and channels meant Telia's audience wasn't experiencing one cohesive brand, different teams, different tone, same company.
TRY built a tool on Azure GPT-4 and React that analyzes written content against Telia's brand guidelines directly, offering both automatic corrections and the reasoning behind them, adapting its suggestions depending on whether the content is a social post or long-form copy.
No numeric results have been published for this project, the qualitative outcome, as TRY has stated it, is that written communication across Telia now stays consistent and recognizable, wherever in the organization it's produced.
Challenge, solution, results
Challenge
Differences in style across channels led to fragmented messaging, making it difficult for Telia's target audience to experience a cohesive brand voice.
Solution
An AI tool built on Azure GPT-4 and React that analyzes written content against Telia's brand guidelines, offering both automatic corrections and explanatory guidance, adapting suggestions by content type.
Results
Strengthened brand presence by keeping all written communication consistent and recognizable. No numeric results have been published for this project, stated qualitatively here on purpose, not backfilled with an invented figure.
Frequently asked questions
- What problem did Telia have with brand voice?
- Differences in style across writers, languages, and channels meant Telia's messaging was fragmented, its audience wasn't experiencing one cohesive brand voice.
- What did TRY build to fix it?
- An AI tool on Azure GPT-4 and React that analyzes written content directly against Telia's brand guidelines, offering both automatic corrections and the reasoning behind them, adapting its suggestions depending on whether the content is a social post or long-form copy.
- Were measurable results published for this project?
- No numeric results have been published for this project, the qualitative outcome, as TRY has stated it, is that written communication across Telia now stays consistent and recognizable wherever in the organization it's produced.
