Skip to main content
All CollectionsPlatformFAQ
Brand Assistant Admin's FAQ
Brand Assistant Admin's FAQ
Updated over 2 months ago

What is the Brand Assistant?

The Brand Assistant is a chatbot that allows your brand portal end users to interact with your brand guidelines in a conversational way. Think of the Assistant like an extension of your brand team that can take care of answering the everyday questions your users have about your brand.

You can find the Brand Assistant demo below 👇

What can the Brand Assistant be used for?

Your users can ask the Brand Assistant about the content in your brand portal. Below are a few examples. How well these examples work will depend on the content available in your brand portal:

  • Content questions: In which cases can I use a logo lockup?

  • Checking copy: “Review the copy below and point out any areas for improvement based on our writing guidelines. [Insert your copy]”

  • Generating copy: “Create copy for [insert scenario] based on our writing and tone of voice guidelines”

  • Instant translation: The Brand Assistant supports over 100 languages out of the box. If your native language is different from the language your brand portal is written in, you can ask the Brand Assistant questions in your native language and it will translate the relevant content from the brand guidelines. Example: Assuming your brand guidelines are in English and your native language is German, you could ask “In welchen FĂ€llen kann ich ein logo lockup verwenden?” (In which cases can I use a logo lockup?) and the Assistant will provide the answer in German.

Depending on the content of your brand portal, many more use cases may be possible and we invite you to try them out.

How does the Brand Assistant work on translated guidelines?

Currently, the Brand Assistant will always take into account the default language and base its answer on the content available there. Content available only in the translated versions will not be considered for the answers. Nevertheless, the Assistant can answer in over 100 languages, depending on the language of the user prompt. For example, if the user asks a question in German, the Brand Assistant will respond in German even though the default language of the guideline is English.

How do I enable and disable the Brand Assistant?

The Assistant can be enabled and disabled by account admins in the account settings. Once enabled for the account, account admins still need to toggle on the Assistant for each individual guideline where they want the feature to be usable. Get more information in this help article.

Are there special terms connected to the use of the Brand Assistant?

Yes, when you enable the Brand Assistant on your account, you will need to accept the Beta terms and privacy notice.

How can I control which users have access to the Brand Assistant?

You can control which guidelines the Brand Assistant is available. All logged in users who have access to this guideline will be able to use the Brand Assistant on it. If you’d like to restrict the test user group, we recommend copying the guideline you’d like to test on and only inviting the users to it that you’d like to include in the testing. Then you can enable the Brand Assistant on this guideline via the usage dashboard.

Does the Brand Assistant respect targets and permissions that I’ve defined on my guidelines?

Yes, the Brand Assistant respects all targets and permissions, meaning your users will only receive answers based on information that is available to them. This means that two users with different permissions and/or targets may receive different answers due to the content the Brand Assistant uses to generate the answer.

Is the Brand Assistant available for non-logged in users on my public brand portal?

No, the Brand Assistant is only available for logged-in users at this point.

What does “Beta” mean in the context of the AI Brand Assistant?

The Brand Assistant is still under development. We regularly release changes, improvements and new features and you may occasionally experience minor bugs or performance issues because of this. During the Beta phase, your feedback is particularly important to us. Read more on how to provide feedback below.

ï»ż

What data does the Brand Assistant consider for its responses?

The Assistant considers text data available in your brand guidelines (e.g. tables, do and don’t, annotations, etc.) Images, icons, PDFs etc. are currently not part of the Assistant’s knowledge base.

What technology does the Brand Assistant rely on?

The Brand Assistant uses Azure OpenAI GPT 4o as a large language model in the background. For monitoring, we use Langfuse. Both services are hosted in Europe.

Will any of my data (guidelines or prompts) be used to train the model?

No, none of your data will be used to train the model. We use a technique called Retrieval Augmented Generation (RAG), which allows the Assistant to be aware of your guideline content without the content becoming part of the model’s general knowledge.

Do any of the services powering the Brand Assistant use personally identifiable information (PII)?

No, unless the user shares PII in the input (prompt). E.g. they write: “Hi, my name is John Doe”. The Assistant is not being trained on any of your users' data or questions.

Can I connect my own large language model to the Brand Assistant?

This will need to be evaluated on a case by case basis, depending on the model you’d like to use. Please ask your Customer Success Manager to put you in touch with the responsible product team.

Do any of the services powering the Brand Assistant install cookies?

No, no new cookies will be installed for the use of the Brand Assistant.

Are there any extra costs associated with the Brand Assistant?

No, during the Beta phase, the Brand Assistant is free, although usage limits may apply. When the feature leaves the pilot phase, we will update you on pricing.

Are there analytics available for the Brand Assistant?

There are no Analytics currently built into the product but we can provide certain statistics on request, such as number of conversations or average rating of answers (how many users rated answer with 👍 or 👎) some examples of questions, answers and user comments received.

How do I know whether the Brand Assistant is providing accurate responses?

The Brand Assistant works best with specific and detailed questions about your brand. However, AI outputs may be wrong or misleading. Therefore, we recommend validating the accuracy of content generated by the Brand Assistant before relying on it. To this end, the Assistant provides the link to the source document for verification with every answer it provides and users are alerted to the fact that AI outputs may be wrong.

Where can I learn more about the roadmap and be updated on future developments of the Assistant?

Your customer success manager can talk to you about the roadmap. Future updates will be announced via our regular product release notes.

I have more questions or feedback, who can I speak to?

Please reach out to your customer success manager who can put you in touch with the responsible product team.

Did this answer your question?