Portant AI Content
Portant AI - generate AI content snippets
Last updated
Portant AI - generate AI content snippets
Last updated
Portant AI inserts Chat GPT-powered content snippets automatically within your documents. This is a super powerful feature that enables you to generate AI content like:
Contract agreement clauses e.g. non-disclosure agreements, service agreements, or lease contracts.
Terms and conditions e.g. payment terms, termination clauses, intellectual property rights, and dispute resolution mechanisms.
Executive summaries
Solution descriptions
Implementation plans
Product or service descriptions
Portant AI enables you to create a document template with consistent formatting and content and then insert AI-generated content at specified sections within your template. The best thing about them is that you can use source tags within the prompt, so the content is bespoke for every document you make.
In this article, you’ll learn how to:
How to add Portant AI to your document
How to construct your prompt
When you open a document template, you will see Portant AI at the bottom of the source table here:
When you click to add an AI content field you will then enter details for your prompt.
A prompt is a written instruction or query provided to an AI model to guide its generation of a response or output.
There are two things you need to enter:
A Field Name for your Prompt - this is what will be shown in your Source Table and in the {{tag}} within your document.
A Prompt - this is the written instruction or query provided to guide its generation of a response or output. You can use fields from within your source to construct your prompt. This will make the prompt content dynamic for each document generated or each time the workflow is run.
When you have constructed your prompt, you can click Save.
The final step is to enter the prompt in your document just like you would enter any other tag.
In most cases, it is a good idea to add a Review Block to your workflow after AI content has been generated. A review block allows you to review and edit the document generated mid-way through a workflow.
We asked Chat GPT to tell us how to construct a good prompt, and this is what it returned:
Specify the Context: Start your prompt by providing relevant context or background information. This can help guide the model and ensure that it understands the specific topic or scenario you're referring to. The more specific and detailed the context, the better the generated response can align with your requirements.
Pose Clear Questions: If you have specific questions or need precise answers, frame your prompt as a clear and direct question. This helps guide the model to provide focused and relevant responses. Make sure your question is specific and unambiguous to receive the desired information.
Adjust the Tone or Style: If you have a particular tone or style in mind for the response, explicitly mention it in the prompt. You can specify whether you want a formal or casual tone, a professional or friendly approach, or any other specific stylistic requirements. Providing this guidance helps shape the generated response to match your desired tone.
Request Examples or Step-by-Step Instructions: If you need the model to generate examples or provide step-by-step instructions, explicitly ask for them in your prompt. This can help the model generate relevant and illustrative content based on your specific requirements.
Set Constraints or Guidelines: If there are certain restrictions, guidelines, or constraints that you want the model to follow, clearly state them in the prompt. For example, you can specify word limits, exclude certain topics, or request the model to avoid using specific terminology. These constraints guide the model's behaviour and ensure the generated content aligns with your predefined guidelines.
Provide Sample Inputs or Outputs: If you have specific examples or desired outputs in mind, you can include them in your prompt. By showing the model what you expect in terms of format, structure, or content, you can guide it to generate responses that match your desired output style.
Here are some examples of prompts that include these points:
Prompt with Context, Clear Questions, and Constraints:Prompt: "In the context of a marketing campaign for a new tech gadget, provide three specific strategies to engage the target audience. Each strategy should be actionable and cost-effective, keeping in mind a tight budget of $10,000. What are your recommendations?"
In this prompt, the context is clearly defined as a marketing campaign for a new tech gadget. The prompt also includes clear questions: providing three specific strategies to engage the target audience that are actionable and cost-effective, considering a tight budget of $10,000. By specifying the context, asking clear questions, and setting the constraint of a limited budget, you guide the model to generate focused and practical recommendations.
Prompt with Style and Requesting Examples:Prompt: "Write a compelling product description for a luxury watch targeting high-end customers. Please use sophisticated language and highlight the exquisite craftsmanship, precision engineering, and timeless design of the watch. Provide an example that captures the essence of luxury and elegance."
In this prompt, the style is specified as using sophisticated language to appeal to high-end customers. The prompt also explicitly requests an example that captures the essence of luxury and elegance. By setting the style and requesting an example, you guide the model to generate a product description that aligns with the desired tone and showcases the key attributes of the luxury watch.
Prompt with Context, Customization, and Sample Outputs:Prompt: "You are creating a user manual for a new software application aimed at beginner users. The manual should include clear instructions on how to navigate the interface, perform basic tasks, and troubleshoot common issues. Provide step-by-step instructions for creating a new user account and logging in. Please structure the instructions using bullet points for easy readability."
In this prompt, the context is defined as creating a user manual for a software application targeting beginner users. The prompt also specifies customization requirements, such as including clear instructions, structuring the instructions using bullet points, and providing step-by-step guidance on creating a new user account and logging in. By setting the context, customization preferences, and requesting sample outputs in the form of bullet points, you guide the model to generate user instructions that are easy to follow and formatted appropriately.
We created Portant in 2021, and the feedback we have received since then has been very helpful and greatly appreciated. If you have any feedback, please feel free to send us an email at contact@portant.co or use our contact form here: https://www.portant.co/support
