This brief is for the IT team scoping a pilot. It sets out what you receive, what your team does, and what we need from you to confirm the fit.
In short: two files loaded into your existing AI environment, nothing installed, nothing connected back to us.
What you receive
Two components, delivered as files you hold. First, the cultural intelligence: a plain-language knowledge base your AI draws on, currently covering Germany, the United States, and fifteen business topics.
Second, the behavioral instructions: a system-prompt layer that tells your AI when and how to use that knowledge. Nothing else is installed.
How it runs
Both components run inside the AI environment you already operate. There is no separate system to buy, host, or connect, and no infrastructure to set up.
Understand Culture is not connected to your environment and receives no data from it. Your people’s questions, messages, and documents never leave your systems.
What your team does for a pilot
Scoped to the pilot department, a single administrator:
- Creates a dedicated assistant, project, or agent for the pilot group.
- Loads the cultural-intelligence document, a single consolidated file, as its knowledge source.
- Places the behavioral instructions into the instruction or system-prompt field.
- Restricts access to the pilot users.
- Runs a short test with realistic scenarios to confirm the AI surfaces the right material.
- Confirms that the platform’s own usage reporting is available, which is the basis for the active-user count.
Effort and who is needed
On platforms with a projects, assistants, or agent layer — for example Claude for Enterprise, ChatGPT Enterprise, Microsoft Copilot, Google Gemini or Vertex, or Amazon Bedrock — a pilot is a one-person, roughly one-day task, and that person needs administrator rights rather than deep engineering.
In most of these environments the intelligence can sit in the assistant’s context as a single document, so on these platforms that is all it takes.
Only where a platform requires the knowledge to be chunked into a retrieval (RAG) index does the setup call for your AI-platform engineers and more time. The hands-on work is a person-day.
Allow separately for any security or governance review your own process requires, which can add calendar time the technical work does not.
Security and data
Because everything runs inside your environment and nothing is sent to us at UC, the deployment stays within your existing data-residency and security boundaries.
There is no external data flow to review, no vendor connection to secure, and no telemetry or licence check that reports back to us.
It works as a set of files you hold rather than a service you send work to. Your team can verify this directly rather than take our word for it.
What we need from you to scope it
So that we can confirm the fit and effort, please tell us: (1) which AI environment the pilot teams use, (2) whether that environment can hold a single knowledge document in its context or requires chunked retrieval, and (3) the pilot audience and the approximate number of users.
Commercials
The price is €150 per active user per year — an active user being a colleague who has received at least four responses enriched with cultural intelligence. You pay only for those people.
Access is scoped by your administrators, team by team, so you begin as small as you wish and widen from there. Full commercial terms are set out separately. This brief is for technical scoping.
Support and updates
When we improve or extend the intelligence, you receive the new version and your team loads it the same way. Where helpful, we adapt the behavioral instructions to your specific platform so the cultural layer triggers reliably.
Back to Start with pilots.