FAQ

How does FactAI Lab work with our existing LLM setup?

FactAI Lab builds AI automation around the tools and models you already use, and keeps a record of what the AI did. We do not sell a software kit or a self-serve platform today. In a pilot we agree with you how the workflow connects, and we start from your current model and prompts.

What accuracy do you achieve?

We do not publish a hallucination rate or an accuracy figure, because it depends on your documents and your task. In a pilot we measure your baseline error rate on your own workflow before and after, and report both to you.

What does the audit record contain?

We agree with you what each record keeps when we scope the pilot. The aim is that each output can be traced to the prompt, the source passages used, the model and settings, and the result of any check.

Is our data used to train your models?

No. We do not use your data to train models. How long data is kept and where results are stored is written down before a pilot starts.

Which regulations does this support?

We do not certify compliance, and we do not claim ready-made report templates for any regulator today. What we build is a record of what the AI did, so your counsel can check it against the rules that apply to you.

What does pricing look like?

We start with a 30-day structured pilot at a fixed price, so you can measure value before committing. The price depends on your workflow, volume and data requirements, so we do not publish a rate card. Book a call and we will talk through the pilot and its cost.