Chat with documents and translate text in Tulip apps: Introducing AI trigger actions

We're excited to preview AI Trigger Actions, the first step in our journey to a Frontline Copilot in Tulip. AI trigger actions bring the power of AI to Tulip applications, allowing engineers to build powerful functionality into their applications to support operators in their facilities.

To add logic to your applications without writing code, Tulip uses triggers, blocks of if-then logic that can perform actions based on input or context from operators and other data. AI trigger actions allow you to chat with documents, parse text, and return summarized information and hints to your operators. Let's look at a few different AI actions to get you started:

1. Extract Text lets your operators or engineers upload an image of a document, label, or traveler, and pull specific text out of the image and into an app or table.

2. Answer Questions lets you upload .PDF files and use AI to query the documents and get answers back about the content of the files. If you have an SOP or machine maintenance manual, an operator can quickly ask a specific question and get a summarized answer so they can get back to the task at hand.

3. Find References in your .PDF files. This goes hand in hand with the Answer Questions action and provides a reference page in your documents for an operator or technician to search for an answer. If a piece of equipment needs to be cleaned, they can quickly find the next steps to do just that.

4. Translate. In situations where multiple languages are spoken or used in your plant or process, the Translate trigger action allows a user to enter text into their Tulip application and have it returned in a specific language, reducing the time it takes to look something up outside of Tulip.

The Tulip platform will always be human-centric, ensuring that your team has the right tools to get the job done, and our approach to AI doesn't change that. Context is key, which is why AI-triggered responses appear with references to the original text, allowing your operators to easily verify the answer or dig deeper into the material to find instructions or more information. This keeps humans in the loop and helps operators quickly navigate through documents, manuals, and SOPs so they can get back to the task at hand.

If the platform does not have enough information to provide an answer, it will let your operators know that it cannot respond. We also use large language models that don't require training, so you can get up and running quickly, and make your information more accessible for your team.

Tulip's investment in AI with Frontline Copilot brings powerful AI functionality to the platform's intuitive no-code trigger editor. AI trigger actions will be available in beta to select Tulip users as we work to refine and gather feedback. If you're interested in testing these features and providing feedback, please email us at copilot@tulip.co.