_
_
ServiceNow
You’ve probably heard the buzz around ServiceNow’s Now Assist product which enables Generative AI on the ServiceNow platform. If you have been living under a rock for the last 6 months, catch up here.
There is no question that Now Assist will bring value to your organization. From the time reduced by summarizing written text within tickets, to the increased efficiency of GenAI writing code for your developers. Setting up these value adds and the other’s currently available OOB can be intimidating at first glance, but we are barely scratching the surface. The scariest thing is if I don’t implement the Now Assist Generative AI foundational product now, do I risk realizing value fast enough as future features are released because I waited too long?
The lead time to upgrade a plugin is much shorter than implementing an entire ServiceNow product. As new GenAI features and skills continue to release, let’s make sure you have everything set up so that you are ready to realize value by reducing costs and improving quality faster than the competition.
Installation
The first step is the most basic. Getting plugins installed is like watching paint dry. But the question is, which plugins are required, and which dependencies are needed? What happens when there is a conflict?
Now Assist plugins are broken out by Product Workflow. Does the team know which plugins to request and install? Which internal stakeholders should be engaged to make these decisions?
Integrations
Integrating with OpenAI, OpenAI or Google? What if Security policies do not permiss the integration? Maybe deploying a LLM behind your firewall and integrating it with ServiceNow is the only option. How are you going to navigate those conversations? Who is going to handle the configuration on both sides of the equation to enable a secure integration?
Now Assist can also be integrated with AI search and Virtual Agent. Of course, it’s beneficial to have Generative AI assisting search results on the platform and assisting the virtual agent with its responses to your user base. But, does the current team have the bandwidth to correctly enable this integration?
Activating and Configuring Skills
Within each plugin exists sets of skills. Here are some of the main skills:
The skills need to be activated and configured prior to use. Triggers should be set informing the skill when it should execute. Input and output parameters should be defined. AI model parameters can be adjusted if applicable for the underlying use case while considering the target model’s strengths and weaknesses. Generic prompts should be set, especially while leveraging text-to-code functionality to inform a model to take internal coding standards into account as it delivers code to your instance.
Are you confident in being able to understand what skills should be configured and how they are best configured to meet the needs of a particular use case?
Models also have token limitations, your team will need to be aware of how to side step the limitation with recursive summarization techniques.
Security and User Access
You will need to determine who is allowed to configure Now Assist features on the platform as well as who will consume that functionality. It’s all Role Based Access Control, you will have to map functional roles to the technical roles provided within the Now Assist product.
Sensitive data is another concern that must be accounted for. The team will need to understand how to detect and mask sensitive data. Can the team do this efficiently or effectively? What is the risk if they don’t?
Get Started
RapDev is the best way to get started with ServiceNow GenAI and Now Assist.
Now Assist is powerful while each small component has to be carefully thought out and precisely tuned in order to extract the most value from it. RapDev has the answers to all of the questions above and the thousand others to long for a blog post.
Stop wondering how to take advantage of Generative AI on ServiceNow and simply get the advantage by talking with RapDev today! Reach out to our team at chat@rapdev.io
.png)
