Problem statement
- AI-generated responses don't match the host's brand voice or communication style
- AI responses contain telltale signs of AI writing (e.g., excessive em dashes, generic phrasing)
- AI lacks company-specific context that would make responses more relevant and personalized
- Guests can detect when AI is being used, which undermines trust
Current workaround
Hosts must manually edit most AI-generated responses to align with their brand voice and remove generic AI language before sending to guests.
Impact
- Erodes guest trust: Guests have asked if AI is being used, which damages the perception of personalized service
- Time-consuming: Editing responses takes extra time, reducing the efficiency gains AI should provide
- Limits AI adoption: Hosts avoid using the AI feature as much as they'd like because responses require significant editing
Proposed solution
Enable hosts to customize how the AI generates responses, similar to how Claude or ChatGPT allow custom instructions. This could include:
- A settings area where hosts can define their brand voice guidelines
- The ability to specify what language or phrases to use (and avoid)
- A way to provide company-specific context so the AI understands the business
- Training the AI to write in the host's specific style, not just generically
Unlock
Once implemented, hosts would be able to:
- Use AI-generated responses without needing to edit them
- Save significant time while maintaining their unique brand voice
- Scale guest communication without sacrificing quality or authenticity