LLM content & AI prompting
Unlocking $18 million in annualized revenue with LLM-driven content
The JustAnswer chatbot initially relied on a set of hand-crafted, hard-coded responses triggered by user input. These responses performed reasonably well, especially with help from our machine learning team. However, we saw the potential to go even further by using large language models (LLMs) to generate unique, real-time responses tailored to each user. This approach promised a more dynamic and engaging interaction, unlocking untapped revenue opportunities.
Boosting conversions with hyper-personalized chatbot gambits
By using LLM-generated responses to craft hyper-personalized prompts, we’ve seen a boost in user engagement and conversion rates. These tailored responses make users feel acknowledged and understood, building rapport and increasing the likelihood of conversion.
Transforming non-converting chats into valuable leads
We took things a step further by harnessing AI to extract key data from non-converting chats, allowing us to sell these as leads in specific legal subcategories. By collaborating with partners, we identified essential information they require and designed AI prompts to gather this data with minimal additional questions, preserving the high-performing funnel. First, AI scans the user’s initial message to check for required details. If necessary, it then prompts for missing info—only asking when essential to reduce user drop-off. Finally, AI formats the data for seamless sale to partners. This lead-generation test alone has delivered a new $18 million in annualized revenue, even before optimizations.