Frame aims to understand business dialog within each organization's sphere of communications. To accomplish that, our natural language understanding (NLU) framework is informed by three philosophies that run counter to prevailing trends.
While most NLU research focuses either on automated response to individual queries (e.g., intent systems, bots) or manipulating single-author texts (e.g., translation, summarization), Frame is focused on identifying speaking behaviors and meanings that arise with multiple speakers.
A common AI strategy is to perform mass data aggregation via a platform, then resell the learned domain model to platform users. Frame is instead focused on generating customer-specific insight, and we measure ourselves by how quickly and easily we do so. We've designed our production and data pipelines to strictly segment by customer, supporting on-prem deployments if needed.
Customer-facing teams are subject-matter experts at their conversations, but are also operationally taxed teams that must prioritize conducting conversations over research. Frame's job is to make the most of their expertise without disrupting existing processes or demanding too much time — at any point in our adoption process.