Research Overview
Research is how you teach the AI about your business. You point Instruqt at a company or product, an AI agent crawls public sources, and the result is a structured profile you can review, refine, and publish.
Published research becomes the reference material the assistant draws on when it generates a lab: your terminology, your products, your writing style. It isn’t a hard prerequisite — you can generate a lab without any research at all — but a lab grounded in real company and product context is far more useful than one written from generic assumptions.
You’ll find research under AI Context in the management interface (/manage/<organization>/research).
Companies and products
Section titled “Companies and products”Research is organized around two kinds of entities, shown on separate tabs:
- Companies — the primary research entity. A company profile captures business context, domain knowledge, and brand or style information.
- Products — research focused on a specific product. A product can either belong to a company (and appears nested under it on the Companies tab) or stand on its own as a Standalone product on the Products tab.
When you research a company, the agent can automatically discover related products and create product research items for them, so you don’t always have to add products by hand.
The research lifecycle
Section titled “The research lifecycle”Every research item is either a draft or published, shown as a status pill in the list:
| Status | Meaning |
|---|---|
| Draft | The profile is still being worked on — by the agent, by you, or both. It can be edited freely. |
| Published | The profile has been finalized and is available as context for lab generation. |
The typical flow is:
- Start research — add a company or product by filling in a short intake form.
- Let the agent work — the AI crawls public sources and streams documents into the editor as it writes them. A draft opens directly in edit mode, so you can follow along and take over between the agent’s turns.
- Review and edit — read the generated profile, annotate it for the agent, or edit it yourself.
- Publish — finalize the research so it can be used to generate labs. You can re-open a published profile later and publish it again.
flowchart LR A[Add company<br/>or product] --> B[Agent researches<br/>public sources] B --> C[Review & edit<br/>the profile] C --> D[Publish] D --> E[Generate a lab<br/>with AI]
Why research matters
Section titled “Why research matters”Spending a few minutes on research up front pays off in the quality of every lab you generate afterward:
- Accurate content — the AI works from a real profile of your company and products instead of guessing.
- On-brand output — a generated style guide keeps tone and terminology consistent with your brand.
- A point of control — reviewing and editing research lets you correct or steer what the AI knows before it writes a single lab.
- Reusable context — published research is picked up automatically for every lab you generate, so you only have to get it right once.