The AI assistant is a chatbot built inside the Workflow app’s enhanced proofer. While making comments on a proof, your team can use it to boost its creativity by having it generate fresh copy or design suggestions.
Turn on AI assistant¶
The AI assistant is an opt-in feature. To turn it on for your site, first submit a case to DAM Customer Support.
The AI assistant will not be turned on until we perform a review of your contract to determine if it falls within your service agreement. If changes are needed to allow the use of AI with your organization, we will provide steps to update your terms before enabling the feature.
Ask AI assistant a question¶
In the Workflow app, you can add comments and use the AI assistant if you are a reviewer of a deliverable. Open a deliverable and click Review proof.
Then, click the add comments icon.
There are two ways to ask the AI assistant a question:
- Select the AI assistant icon in the comments tools and then give the assistant a commend or ask a question.
- Add a mention of “@AI” in your comment.
Try asking the AI assistant to generate:
- Alternative headlines
- New ideas for visuals
- Personalized content for a particular audience
- Suggestions based on competitor content
The AI assistant will generate an answer and post it under your comment. Replies added by the AI assistant are marked with a special icon, so it is easy to differentiate them from reviewers' comments.
Frequently asked questions ¶
What powers the AI assistant?
The AI assistant uses ChatGPT (v. 3.5) by OpenAI via their API. The model is primed for creative and marketing use cases so that the chatbot provides responses that are more informed.
Are responses from the AI assistant brand-safe and accurate?
Much has been written about generative AI technology providing inaccurate or inappropriate information. The AI assistant is built on similar technology, and we cannot guarantee the information it provides will be factually accurate. It does not contain additional filtering for trademarks or creative commons licenses, nor for language that may be offensive to other cultures or subgroups. As a result, we encourage you and your team to verify responses from the AI assistant before including the information in published materials.
Will the AI assistant respond with language that’s offensive?
Having been built on a large multimodal language model, the AI assistant will respond to any question you ask it. As a result, if you ask a question and expect an offensive response, there’s a reasonable chance that you’ll get one. Given that Acquia is not responsible for the questions your reviewers may ask, Acquia is not responsible for the answers that the AI assistant delivers in return.
Our advice is, if your reviewer wouldn’t be comfortable asking a colleague the question in a company-wide messaging channel, then they should avoid asking the AI assistant that question.
What does the AI assistant do with the questions I’ve submitted? Will other companies have access to this data?
According to OpenAI’s API terms of use, “data submitted through the OpenAI API is not used to train OpenAI models or improve OpenAI’s service offering.” In addition, OpenAI states that they do not share content with third parties for marketing purposes.
For each request, the only content the AI assistant sends to ChatGPT via OpenAI’s API is the query in the comment thread itself. Acquia does not send any personal information such as the name of the reviewer or their email address.