We’ve all heard the hype around AI, so we won’t add to it here. In a nutshell, powerful new AI and machine learning capabilities are raising interesting prospects around automation and efficiency. As a result, all the major technology players are integrating this technology into their products.
At Skuid, we have been experimenting with AI tools that help teams complete business tasks more rapidly, with an additional focus on improving UX. Now, we’re incorporating new conversational UX patterns. On a Skuid page, users work with data from Salesforce objects, Hubspot, a Postgres database, or some other main system. But the Skuid UI layer allows integration with other data and services in a single experience, enriching your business data and making it easier to reach your objectives. Those services can even include AI tools.
Here are a few of the ways we’ve envisioned AI helping with this sort of assistive integration:
- In an “Account Creation” flow for a sales organization
Use ChatGPT to draft a description and an initial sales strategy for a new company.
- In a marketing system
Leverage ChatGPT to identify better ways to discuss product opportunities and draft outreach communications.
- In description fields, email text, action plans, and other longer elements
Edit the tone and voice of your materials using ChatGPT or use the tool to create an initial draft or outline.
And these are just a handful! We’re confident there are hundreds more potential AI applications out there.
Integration with Skuid
OpenAI has been an early leader in publishing thought-provoking capabilities for LLM Interaction. Because they have implemented standards-compliant REST data sources for interacting with their models, Skuid can easily trigger these models and consume data returned from them.
Recently, our R&D team spent a few days in a “hack-AI-thon” and drafted several interesting integrations using the OpenAI models. So, we thought we would share them here! We’re also publishing this material on our Skuid Labs repo, a place for interesting experiments and bleeding-edge examples of how to use Skuid. Feel free to check out these examples and incorporate them into your own Skuid pages and business applications.
1. Have AI draft your description fields
Our first example asks the AI Models to draft initial values for any field being managed on your Skuid page. It creates a suggested prompt based on information from a contextual record (like an Account Record), lets the user edit that prompt, and submits it to the AI model. Then it allows for review and iteration of the response before finally copying the output to the specified field in your application.
The beauty of this application is that it’s generic; it can be called by any field within your Skuid page and takes information from the field, model, and suggested prompts as its input parameters.
Check it out here.
2. Create a menu of actions powered by GPT
A second example generates a list of actions that can be taken based on field data. Each of these actions submits a contextually-aware prompt to the GPT models, in some cases allowing the user to further modify the prompt before submitting.
3. Add sample data to your Skuid applications
Our third example leverages the Open AI API to create sample data for your application. Use it to populate a test bed or a sandbox table with data based on what is included in your Skuid model.
Again, please feel free to use these pages as a reference. You’ll need to create an account with OpenAI and retrieve API keys in order to access that data source, but the instructions in our repo will walk you through the steps.
A quick caveat
AI services are still new, and teams are only now establishing best practices for using them safely. That’s why it’s so important to read up on AI security. Ask yourself: When your users enter a prompt, where does it go? Is there any sensitive data that could be exposed as a result of training a large language model (LLM)?
Look for LLM products that can run securely and train on your org’s data without being widely accessible. Tools like Scale AI will be critical as privacy-sensitive organizations work to leverage LLM tools for their end users.
Support from Skuid
Building a Skuid app that uses an AI data service? Have another use case that combines business data and OpenAI? We’d love to hear about it. Let us help you turn your idea into an application. Just send us a note at email@example.com or tag us on LinkedIn.