Now Assist Article Generation: Maximise GenAI in the Now Platform

Knowledge articles are a massive time saver for your team. Now, thanks to Now Assist, you can take your knowledge base to the next level with GenAI.

ServiceNow's May 2024 Store Release introduced Now Assist in Knowledge Management. This feature promises to significantly improve the creation, management and use of knowledge articles, driving faster resolution times and enhancing self-service capabilities. 

Here's an overview of what Now Assist offers and how you can prepare your ServiceNow knowledge base to get the most out of it.

Key features of Now Assist in Knowledge Management

Implementing Now Assist stands to transform how you create and manage knowledge articles in the following ways:

• Faster turnaround time for knowledge article creation
  • Now Assist accelerates the creation of knowledge articles by leveraging AI to draft articles based on case details. This reduces the time agents spend on documentation, allowing them to focus on resolving issues.
• Increased self-service and deflection
  • By making relevant knowledge articles more readily available, Now Assist empowers users to find solutions on their own. This enhances the self-service experience and reduces the number of tickets submitted to support teams.
• Faster resolution time on tasks
  • With reduced knowledge gaps and more accurate articles for troubleshooting, tasks can be resolved more swiftly. And greater efficiency = improved customer satisfaction and operational effectiveness.

Questions to ask before implementing Now Assist in Knowledge Management

To ensure a successful implementation of Now Assist in Knowledge Management, you need to understand your current challenges, desired outcomes and metrics for success. 

Here are some key questions to consider:

Understanding your challenges

• Knowledge gaps: Do you currently have articles that don't exist but are needed? What specific knowledge gaps are you facing?
• Article quality: How would you rate the quality of your existing knowledge articles? What improvements are needed?
• Self-service effectiveness: How effective is your current self-service system? Are users able to resolve issues on their own, or do they often need help?
• Resolution times: Are lengthy resolution times a problem in your organisation? What impact does this have on your overall efficiency?
• Escalations: How frequently do preventable escalations occur? What are the common reasons for these escalations?

Identifying desired outcomes

• Quality improvements: What improvements in article quality do you expect to see with Now Assist? How important is it that articles are viewed and used (attached to records)?
• Enhanced self-service: What are your expectations for increased self-service interactions and deflections? How will this benefit your organisation?
• Resolution time reduction: By how much do you hope to reduce resolution times? What targets are you aiming for?
• Escalation reduction: What are your goals for reducing the number of escalations? How will this improve your overall service efficiency?
• Other KPIs: Are there other key performance indicators you are currently measuring that you expect Now Assist to impact?

Measuring success

• Current measurements: What are your current measurements of a successful knowledge programme? How do you track and communicate these measurements?
• Value communications: What are your plans to communicate the value of article generation with Now Assist versus manual methods?
• Establishing baselines: Have you established any baselines before launching generative AI? How will you measure improvements against these targets?

Preparing for knowledge article generation

To prepare for the integration of AI in knowledge article generation, it's crucial to evaluate and establish strong governance processes. Here's a guide to help you get started:

1. The quality of knowledge is linked to case quality

• Enhance documentation standards: The quality of your knowledge articles is closely linked to the quality of case documentation. Ensure that your case documentation meets high standards, as this will directly impact the effectiveness of the AI-generated articles.
• Train agents: Incorporate knowledge generation best practices into your agent training programmes. This helps agents understand the importance of creating detailed and accurate case documentation.
• Leverage useful platform features: If you're not already, start using templates for knowledge articles to ensure consistency. Explore Coaching for Learning to improve agent skills in documentation and article generation.

2. Human verification is vital

• Evaluate your quality review process: AI can generate well-written articles, but human oversight is crucial to maintaining accuracy and relevance. Regularly refine your quality review process to ensure AI-generated content meets your standards.
• Communicate the value of knowledge creation: Educate your agents on the importance and benefits of creating high-quality knowledge articles. Highlight how these articles improve business outcomes and operational efficiency.
• Embrace feedback: Ensure there is a mechanism for consumers and agents to flag issues and provide feedback on articles. Regularly review and act on feedback to continuously improve the knowledge base.

3. Launch with pilot groups and scale on your timeline

• Launch with pilot groups: Begin by implementing Now Assist for Knowledge Management with pilot groups to control the initial rollout. This helps identify and resolve issues before a full-scale deployment.
• Train agents on content standards: Ensure your agents are well-versed in organisational content standards. Regularly update training materials to reflect best practices and new features of Now Assist.
• Scale on your timeline: Use insights and lessons learned from the pilot phase to plan your scaling strategy. Scale up at a pace that aligns with your organisational capacity and readiness, and continuously evaluate the process to meet your goals.

4. Evaluate governance processes before a full implementation

• Dealing with duplicates: Train agents to search thoroughly before generating new articles to avoid duplications. Implement processes to identify and eliminate duplicate articles before rolling out Now Assist.
• Prioritise validity and relevance: Ensure your articles contain valid, updated information and are relevant to the current needs. Regularly review and update articles to maintain accuracy and relevance.
• Establish review processes: Develop robust processes for reviewing content to ensure its relevance and validity. Focus initially on high-value articles and maintain a cycle of review and improvement.

Best practices for effective article consumption

To ensure your knowledge base is effective and easily consumable by both AI and users, follow these guidelines.

Clean up and prepare articles for consumption

To ensure optimal AI search results and article consumption, focus on these areas:

• Focus on articles with high view and use counts [kb_use], and those that have been recently updated. Ensure that these articles are up-to-date and maintained by the correct ownership group.
• Articles should ideally have a ‘Valid To’ date set to one year or less, and higher-value articles should be subject to more frequent reviews. Regularly check the authorship and ownership group to maintain accuracy.
• Use the Meta field and relevant keywords to improve AI search results. Avoid content that cannot be summarised effectively, such as links within the article body, attachments, images, videos and tables.

Knowledge base architecture guidance

• Avoid duplication to reduce maintenance and potential conflicts. Regularly review your knowledge base to eliminate outdated or redundant articles.
• Titles should be specific, descriptive and free of duplication. Include relevant search terms to improve discoverability and think about how users might search for the content.

Article content guidance

• Create hyper-focused articles covering one topic fully. AI cannot summarise content from external links or sources, so ensure all necessary information is contained within the article.
• Use clear plain-text language, avoiding jargon, technical terms and abbreviations (unless defined within the article). Simplify language to ensure clarity and comprehensibility.
• Provide definitive answers, ensuring answers are complete and self-contained within the article. Use complete thoughts rather than fragments and provide context where necessary.

Article formatting guidance

• Minimise reliance on images and videos as they cannot be summarised by AI. If you use images and videos, provide supplemental text that describes the content.
• Use article templates for your knowledge base to ensure consistent content capture and adherence to formatting standards.
• Utilise basic styling, such as bold, italic and underlined text for emphasis. This can be useful for the AI's understanding of the content, but LLMs cannot process advanced styling such as custom fonts and colours, so these should be avoided unless they are only used for aesthetic reasons.
• Organise your content with header tags and lists (ordered or unordered) to help both AI and users break down complex information.
• Write short and focused paragraphs to enhance readability and coherence. Avoid lengthy content that can overwhelm the AI and result in unclear summaries.
• Use appropriate HTML tags such as <h1> to <h3> for headers and <p> for paragraphs. This also helps with both accessibility and SEO.

Get ready to maximise the benefits of Now Assist in Knowledge Management

Now Assist represents a significant advancement in Knowledge Management on the Now Platform, offering the tools to create high-quality, relevant articles and summarisations. 

By following the guidelines and best practices outlined above, you can leverage AI to enhance your knowledge base and drive better self-service and faster resolutions. 

Prepare your organisation today to fully benefit from the capabilities of Now Assist in Knowledge Management.

If you need help implementing Now Assist, want to find out more or even see it in action – reach out to us today and we’ll help guide you on your journey to using GenAI in ServiceNow.