Speechify subscribers paid AI assistants at more than four times the panel rate
Share with a paid AI-assistant receipt in the 12 months through August 2026
Share with a paid AI-assistant receipt
Source: YipitData U.S. e-receipt panel. Speechify subscribers n=3,098; active panel n=1.40 million. AI assistants were paid ChatGPT or OpenAI, Claude, and Perplexity; Gemini was excluded.
Chart data
| Series | Share with a paid AI-assistant receipt |
|---|---|
| Speechify: any paid AI assistant | 42.9% |
| Active panel: any paid AI assistant | 9.6% |
| Speechify: paid ChatGPT | 29.6% |
| Active panel: paid ChatGPT | 5.2% |
Does the overlap show that paid AI assistants replaced Speechify?
No. The overlap analysis shows only that the same subscribers paid Speechify and an AI assistant during the measured period. It does not establish which subscription came first or whether AI-assistant payment changed later Speechify activity. A retention comparison with fully specified cohort sizes would be needed to answer that question.
What does the overlap measure leave out?
The analysis observes paid receipts, not free AI usage or product activity. Gemini was excluded because its receipts could not be isolated cleanly, and OpenAI receipts may include some API spending. Annual billing and incomplete receipt capture also depress the absolute Speechify persistence rates, so the comparison is more informative than either level. Demographic differences were not fully matched.
Assumptions & Methodologies
We analyzed YipitData’s U.S. e-receipt panel. The cohort included 3,098 Speechify subscribers in the 12 months through August 18, 2026; paid AI covered ChatGPT or OpenAI, Claude, and Perplexity. Receipt activity is incomplete and measures paid relationships, not product usage. Free AI usage is not observed, Gemini was excluded because its receipts could not be isolated cleanly, and OpenAI receipts may include some API spending. Same-period overlap does not establish sequence, substitution, retention effects, or causality.