Choose your industry. We research what could help customers understand, choose, buy, use, and keep using LLMs — across product, marketing, sales, pricing, and retention.
What would help your customers understand, choose, buy, use, and keep using your LLM?
For model developers, agent platforms, API providers, and AI-native software companies. We research the full human adoption cycle — from product interaction and narrative to willingness to pay and retention.
When does additional LLM capability create enough customer value to justify the compute?
For chipmakers, cloud providers, inference platforms, and AI infrastructure teams. We connect real customer workflows and perceived value to reasoning budgets, context, latency, modalities, and compute — testing when additional capability creates enough customer value to justify serving cost and potentially support sustainable pricing.
Which LLM interaction works best for your customers on everyday devices?
For smartphone, PC, wearable, XR, robotics, and other device makers. We research what helps your customers use multimodal LLMs naturally and effectively, and when text, voice, vision, on-device inference, or cloud support is worth using.
What would help your users delegate more useful work to LLM agents without losing control?
For enterprise software, cybersecurity, productivity, and workflow companies. We research what users need to delegate useful work to agents with appropriate confidence, permissions, explanations, and oversight.
What helps your users get more value from LLMs when understanding and behavior matter most?
For hospitals, medical technology, digital health, mental-health, psychology, and education organizations. We research what patients, clinicians, learners, and other users need to understand, trust, remember, and act appropriately on LLM-supported information.
The common requirement is a consequential communication problem between an LLM system and real customers, employees, or professionals.
Propose a research question