Now Assist Skill Kit Use Case Library

[Eliza]

Welcome to the Now Assist Skill Kit Use Case Library!

This collection of use cases has been carefully curated to spark your creativity and demonstrate the vast potential that lies within Now Assist Skill Kit.  Each example presents a unique scenario and walks you through the steps required to build it in your instance.

If you are new to Now Assist Skill Kit, you may find reviewing the below material helpful before exploring these use cases:

• Now Assist Skill Kit FAQ
• AI Academy: Introduction to Now Assist Skill Kit
• How to approach building custom generative AI solutions
• Product Documentation

We will be continuously updating this guide, so feel free to subscribe to this article to get updates when we upload more.

Use Case	Description
Ask questions and receive answers for a portfolio	This use case focuses on project/portfolio management, and involves building a custom skill that allows for managers to interact with project/portfolio records by asking questions, with the LLM analyzing the data and returning a response.
Automate the creation of change backout plans	Join @ChrisSteinke2 as he builds a custom skill that uses data within a change request to generate a backout plan.
Detect Negative Survey Sentiment and automate the response	We will use NASK to send the survey responses to the Now LLM which will predict the sentiment of a survey. If the sentiment of the survey is negative, we will automate the creation of an incident record so we can address the issue quickly
Knowledge Article Coach	We will create a custom skill that reviews the contents of our knowledge article, and to provide suggestions on how to improve it. This feedback will be stored in a Feedback Task attached to the knowledge article.
Knowledge Article Compliance Review	For this use case, we will use NASK to ask the LLM to generate a new knowledge article using branding and compliance policies outlined within our instance. We leverage both the GRC and Knowledge Management Advanced plugins in this scenario.
Knowledge Article Categorizer (Can also apply to incidents, cases, etc)	NASK is used in this use case to build a custom skill that gathers the existing categories for knowledge articles (KA) and then evaluates the content of KAs to assign them a category.

[raj96]

Hi @Eliza ,

This is some great content and really appreciate you sharing different use cases.
Could you please help me with a below scenario:
My Organization does not have any knowledge articles, They want to leverage the Now Assist Tool kit to loop through all the incident records and automatically create a bunch of distinct knowledge articles.

Here are couple of concerns I have:

1. if I pass around 20000 incident record data as below to Now LLM:
Example:
Inc1: 

Short description, description, work notes, resolution notes, resolution code etc...

Inc2: 

Short description, description, work notes, resolution notes, resolution code etc...

So on upto inc20000 will the LLM prompt be able to handle that data and provide a response of 1000 articles

Here is what I am trying to get at, is there a limit on how many words the prompt can accept and is there a limit on how many words the response can output.


Apart from the above what is the best way to implement a solution for an organization that wants Now assist to look at all its incidents and generate distinct knowledge articles. It would be great if you could consider this use case and do a demo. Or even if this is something that can be presented as OOB Skill. Will help a lot of organizations that are having challenges with no knowledge articles.

Thanks in Advance!!
Regards

Raj

[Eliza]

Hi @raj96,

Lovely to hear this guide was helpful for you!

There are token limits set by each LLM provider that will likely prevent you from implementing your idea. You can learn a bit more in the FAQ.

I would recommend following the process outlined in the Knowledge Article Categorizer example, where you select a group of incidents and run the skill once per incident.

However do note that each time you run a generative AI skill (even custom ones) you will consume assists. Running this against 20,000 incidents may consume a fair chunk of your allocated assist pool, so do check with your account manager on this topic first.

As a way of lessening the number of times you have to run the skill (and thus consume fewer assists) would be to use predictive intelligence in a subflow to group similar incidents, and then generate a single knowledge article off of those.

I do recommend reaching out to your account rep on this topic however - they should be able to give you guidance based on the exact details of your instance and license.

Thanks,

Eliza 🙂

[mig_gar]

Will there be a Use Case/ additional articles to assist with building out a Ground Truth NASK?

[Chandana E S]

Hi,

Thank you for creating this page. I am trying to create a skill in HRSD that suggests the resolution of the case based on the input. I have created a text box to output the suggested solution on the case form. Created a sub-flow to link these two and a UI action on the Case form, but facing the below issues. Can you please assist:

1. The skill when i am running it using Run test button is giving the same solution every time even though a different case is selected.
2. Also, can you please share what all variables need to be given in the flow designer including input and output? Appreciate quick help!