Prelight FAQ: quick answers about the governed brand record

Before you ask

the short versions
Does this replace our DAM? No. It sits above the one you already have and reads from it. Nothing migrates and nothing gets renamed.
How long until it is useful? The ingest itself takes minutes; we walk you through the read-back on your own files in about a week. No migration project, no naming convention to agree first.
What does it cost to start? Most start on a paid 90-day pilot, credited in full against year one. Basic is $9,000 a year; talk to us for Agency and Enterprise — see the table.
Do our creators need a paid seat? No. Creators get a free prelight account. You share a kit, they connect their own AI tools, and every post is checked against the record before it ships.
Who can see what? Access is scoped by user roles, including partners and licensees. Same permission schemes, as on the app, enforced on the MCP surfaces.
Do you train on our assets? Never — not us, not anyone else. It is a term in the contract, not a setting in an account.
What happens if we leave? You take the whole record — assets, descriptions, rights and the decision trail — in an open format, and we hard-delete on request.
We are a small startup. We've got you! You are a small team hustling to move faster. You are the marketing team, you are the counsel, you are the HR. Talk to us to get the pilot price for a longer period.

Can we not just…

the honest answer
…prompt Claude or ChatGPT? For one task, yes. But a prompt lives in one chat, holds a handful of files, knows nothing about your licenses and starts from zero tomorrow. Keep the assistant you like; give it the record.
…upload our brand book to a custom GPT? That gives one tool a static PDF. It cannot see your media library, licenses or corrections, and every other tool, agency and licensee still needs its own copy.
…connect an agent to our DAM? Then files are reachable, not understood. The agent still opens undescribed files to find one, with no rights check or durable decision trail.
…build this ourselves? You could. What you would own is the software, the data model and the IP-safety layer: secure media retrieval, permissions, provenance, rights clearance and compliance. Models, tools, laws and standards keep changing, so it needs ongoing maintenance. That makes it a product, not a side project.
…wait until AI settles down? Your teams, agencies and the internet already make AI content with your brand. The question is not whether machines read it. It is what they read from.
…start small? Yes. Point us at one folder. The ingest takes minutes, and we walk you through the read-back on your own files in about a week.