Let's connect together
Join us to make our over 890K member community even better. Create your free account and make the world work better for everyone.
Join the Community
Date: Tuesday, June 11th, 2024, 9 am – 10 am PT | 12 pm – 1 pm ET
Topic: Now Assist in Knowledge Management
Join us for a series of live sessions focused on Knowledge Management topics such as getting started, processes and roles, best practices, analytics, release updates, and more!
The academy occurs every second Tuesday of the month at 9 am PST/12 pm EST. We'll have product experts on hand to provide demonstrations, explain concepts, provide guidance, and answer questions.
Register for sessions here.
In this session we'll cover the current features of Now Assist in Knowledge Management that was released on May 9th. This session will include plenty of time for Q&A and product feedback as we want your guidance on how to make generating articles better for future store releases.
Resources
Knowledge Managers Discussion on Community - https://www.servicenow.com/community/knowledge-managers/gh-p/knowledge-managers-
Now Learning training - https://nowlearning.servicenow.com/lxp/en/pages/lxp-search?id=search&q=knowledge%20management&spa=1
Customer Success Center - https://www.servicenow.com/success.html
ServiceNow Impact - https://www.servicenow.com/impact.html
Custom Training and Adoption - https://www.servicenow.com/services/training-and-certification/custom-training-and-adoption.html
Consortium for Service Innovation and Knowledge-Centered Service (KCS) resources - https://www.serviceinnovation.org/kcs/
About Live on ServiceNow®: Live on ServiceNow is our latest event series to help you deploy, adopt, and achieve value faster from your ServiceNow solutions. Each event features an interactive discussion with a product expert, so be sure to come prepared with your questions!
When will the recording for the June 11, 2024 session be available?
Here's the transcript of the Q&A from the event:
Now Assist in Knowledge Management, or knowledge generation is included with the Now Assist for CSM, ITSM, and HRSD applications as of May 2024. You must be entitled to these applications to be able to use this functionality.
The updated date does factor into the relevance score for AI Search. Re-publishing an accurate and still relevant article with a more recent date can help maintain its relevance in search results. See the AI Search product documentation, and our AI Search and Knowledge Management best practices webinar for more information.
The knowledge generation feature as of May 2024 targets the agent persona. Only new articles can be drafted from an incident, case, or HR case. This is a 1:1 relationship. Articles need to be updated manually as of this release. We plan to expand functionality and add capabilities for more personas in future releases.
Now Assist requires a specific licensing agreement. Contact your ServiceNow account representative for more information on licensing requirements.
Now Assist is designed to work with the KCS (Knowledge-Centered Service) template if you have the required KCS plugins and properties installed and configured in your instance. See the KCS product documentation for more information: Activate Knowledge Management KCS Capabilities, Activate KCS Integration for Incident Management.
Knowledge generation uses the approval or publish workflows associated with the knowledge base, so if you have approval workflows set up for your KCS processes, you’ll continue to use those workflows.
Not as of the May release. We’ll communicate more about this functionality in future releases. Agents should utilize Agent Assist and Recommended Actions to search early and often for knowledge articles prior to creating an article for a task.
Now Assist uses AI and natural language processing (NLP) to interpret and refine case details, reducing the time agents spend manually updating information and ensuring that knowledge base articles are clear and accurate. It is strongly recommended that agents and authors review the generated article for errors and correct any information found in the article prior to publishing.
Knowledge Management v3 is the current version, but be sure to check your instance for updates to Knowledge Management and its associated plugins.
No, knowledge generation at this time will not currently scan for duplicates before generating a new article. We do have licensed Predictive Intelligence solutions to show similar articles when authors are working on the knowledge form itself (which is not correlated to our generative AI functionality at this time) and when agents are searching for articles within Agent Assist.
Not at this time. As of the May release, Knowledge generation is only available for Customer Service Cases, HR Cases, and Incidents. We’ll update our communications as the capability expands in future releases.
PDI stands for Personal Developer Instance, which is a free instance of ServiceNow provided to developers for learning, experimentation, and development purposes.
Not currently, it generates plain text with basic styling. Currently, our Q&A model interprets plain text and basic styling to summarize articles with Now Assist in AI Search and Now Assist in Virtual Agent. See this support article on using images with articles.
Yes, ServiceNow provides various metrics and analytics tools within the Knowledge Management module to measure deflection, effectiveness, and usage of knowledge articles, including views, feedback, and relevancy scores. See the product documentation on the knowledge dashboards, as well as our success indicator dashboards.
'Most used KBs' typically refers to the most frequently accessed and utilized knowledge base articles within the knowledge base.
I believe this question refers to Agent Assist as there is no feature called KM Assist. Agent Assist in a workspace uses Zing search, which is an older and more limited search functionality from AI Search that is commonly used in our portal solutions. We recommend implementing Recommended Actions within your workspace to use AI Search solutions.
It is recommended to keep the ownership group to a minimal number of knowledgeable and trained individuals to ensure consistency and quality in the management of knowledge articles. This helps maintain a high standard and reduces the risk of outdated or incorrect information being published.
Having a 1:1 ratio of ownership group members to domain experts may be a good setup for your organization. Keep in mind that ownership groups were created to reduce the 1:1 reliance on an author for an article in case the author cannot perform the duties needed for article maintenance such as review, approvals, and feedback review. You may find that the 1:1 domain expert relationship may produce bottlenecks, but if your domain experts are able to manage the workflow effectively then this strategy may work for your organization, and you can use the ownership group to scale and add more domain experts as needed.
There are no hard and fast rules for review, it depends on the ability of your organization to perform the reviews. We recommend identifying high-value (high use and views) articles and ensuring there is at least a 6-month or quarterly review on these articles to ensure accuracy and relevancy. You may find this cadence works for all articles in your knowledge base, or you may extend the review period to yearly for articles that are not as highly trafficked. The cadence depends on your organization and how efficiently your authors or ownership groups can review articles.
Access to generate articles using Now Assist is controlled through 2 avenues:
Access to the ServiceNow knowledge generation feature can be obtained by entitling and installing the Now Assist for ITSM, CSM, or HRSD applications and ensuring the appropriate licensing and configurations are in place. Contact your ServiceNow account representative for more details.
Knowledge generation uses the context from the incident, case, or HR case the agent is creating the article from. The NowLLM model itself is trained on large datasets. For more information on our models, see the product documentation. Knowledge generation uses the Mixtral-8x7B-Instruct model as of this recording.
Currently, knowledge generation supports the Standard and KCS templates. We have roadmap items to expand to more templates in the future.
Knowledge generation is currently supported on incidents, cases, and HR cases. Usage depends on your licensing for the related applications: Now Assist for ITSM, CSM, and HRSD. Contact your account representative for licensing questions.
We’ll have more information on this in future communication, currently you’ll need to continue using your organizations knowledge governance strategies for training agents on checking for articles prior to creation, and managing duplicate articles.
Knowledge generation as of May 2024 is geared towards the agent use case for generating articles from incidents, cases, or HR cases. We will support more personas and use cases in future releases.
The author of an article can be changed. This process does not differ when using knowledge generation. See this support article for more information.
The primary language for article generation in Now Assist is English as of May 2024, contact your account representative for more information on future supported languages.
Not currently, keep an eye out for future communication on this feature.
Generative AI is not field mapping. We do have the capability of using field mapping to grab verbose what’s in the resolution field and map it to a knowledge article template without generative AI. Generative AI takes the context of the short description, description, activity (work notes and comments) and resolution/close notes to draft the article. The more context you must provide generative AI the more defined outcome you will see. Generative AI can use what’s in the resolution field within a resolution section of an article, the context isn’t solely limited to this field.
User criteria are defined at the knowledge base and knowledge article levels. For more information, see this product documentation on user criteria.
The TinyMCE editor does provide options for spell check, generative AI takes this a step further by producing drafts that are -free from spelling errors and typos.
The availability of the Now Assist panel for generating articles depends on user roles and permissions. It may be restricted to certain ITSM users who have the necessary permissions to create and manage knowledge articles. The users must have Can Contribute user criteria to the respective knowledge base, as well as the roles defined in the configuration for Now Assist panel.
No, knowledge generation only generates the body or text of an article. Categories must be set manually at this time.
You cannot modify the prompt for Now Assist knowledge generation at this time.
You can change the text within the article, but if the question can the article be set to use a new template, this is not possible at this time. Technically, each article template is a unique table that extends kb_knowledge. Changing a template for an existing article means moving it to another table. While the sys_class_name attribute of an article record can be changed, you risk losing the values in the fields that are specific to the old template that do not exist in the new template. To avoid data loss it’s suggested to create a new article rather than attempting to convert templates.
The article title is usually generated based on the content and context of the record it is derived from, but it can be edited and customized as needed to ensure clarity and relevance. It is not a 1:1 field mapping.
Now Assist for Knowledge Management is geared towards the agent persona as of May 2024.
@Ashley Snyder
Thanks for sharing release information around Knowledge Management.
Is this feature designed to be used synergistically with Knowledge Demand Insight, or is it intended to be an alternative to Knowledge Demand Insight in the future?
I need this information to determine whether or not I should use Knowledge Demand Insight going forward.
Join us to make our over 890K member community even better. Create your free account and make the world work better for everyone.
Join the Community