MatrAIxGitHub ↗Join us
Study intake//simulate before reality
01 / 09

Basic information

// who and why

Completion progress 11%
Step 1

Basic information

Tell us who you are and what you want to work on. You will choose a research study or an integration next.

Do not upload secrets or personal data. Do not enter passwords, API keys, customer-level records, patient data, or other personally identifiable information. Secure access can be arranged separately.
Would you be willing to sign a mutual NDA? If you would rather have an NDA signed before sharing anything specific, say so here and keep the rest of this form high-level — we will follow up to get one in place first. An NDA is what lets us work from the real artifacts (internal metrics, unreleased builds, evaluation data) instead of a sanitized description. We can send a standard mutual NDA, or work from your template.
Step 2

How would you like to use MatrAIx?

This intake is designed for individual research studies. Platform integrations and recurring automated workflows use a separate technical-support process.

In a few sentences, describe what your organization does, the relevant product, service, or research context, and what you want to work on with MatrAIx. You can provide detailed study materials later.
What do you need? Pick the closest fit. Choosing AI / agent model evaluation also lets you ask for the underlying data as an asset you keep.
Step 3

Outcomes and success

Tell us how you will decide whether the study is useful. The sharper the threshold, the more decisive the result.

Write it as a threshold you would act on, not just a topic. For example: “more than 50% of the cohort prefers variant B”; “task completion above 80%”; “B beats A by at least 5 points on purchase intent”; “fewer than 10% hit a blocking error”; “no regression against the current model version.” If you do not have a number yet, say what a good result would look like in words.
Any number the result will be measured against — a current conversion rate, a previous survey, a human eval score, an internal benchmark, or an expert estimate. Leave blank if there is no baseline yet.
For example: whether to ship the redesign, which of two models goes to production, whether to fund the next phase, or which segment to target first. Knowing the decision helps us design a study that can actually settle it.
Step 4

Desired deliverables

Describe what you would like to receive when the study is complete. Pick everything that applies — most engagements ship more than one.

Which of these would be useful? Optional Tick whatever is relevant; the box below is where you describe anything specific.
Anything specific about format, level of detail, or who the audience is. For example: scored results plus the failure modes behind them, telemetry we can join to our own analytics, or a deck an executive team can read without context.
Step 5

Confidentiality and governance

Help us determine whether the study needs additional review, safeguards, or a private runner.

Does this study involve confidential data or materials? Do not upload confidential data through this prototype. We can arrange an appropriate sharing process later.
After redaction and de-sensitization, may we include this task in the open-source MatrAIx benchmark?
For example: minors, health or wellbeing, elections, vulnerable populations, personal data, licensed identities, legal constraints, prohibited claims, or content-rights requirements.
Step 6

Related materials

Share any related materials you can provide with us for reference.

Upload concept documents, questionnaires, presentations, screenshots, or other relevant files. You may select multiple files: up to 20 files, 10 MB each and 50 MB in total. Do not upload credentials, personal data, or other secrets.
Step 7

Review your intake

Confirm the study details, then submit. We will follow up using the work email from Basic information.

Intake submitted

Thank you. Our team will follow up using the work email you provided.

Opens your browser’s print dialog — choose “Save as PDF” as the destination.