Now Assist for Knowledge: Real-world use cases and lessons learned?

[shutoishizu]

Hi Community,

We are currently evaluating how to leverage Now Assist for Knowledge Management,

and I’d like to learn from real-world experiences rather than just feature descriptions.

In particular, we are looking at the following use cases and would appreciate any insights, challenges, or lessons learned.

1. Auto-generating Knowledge from Incidents

- Are you using Now Assist to generate knowledge articles directly from resolved incidents?

- At what point in the incident lifecycle do you generate the article (on Resolve, on Close, manually triggered, etc.)?

- Does Now Assist (or any OOTB capability) actually check for *existing or duplicate knowledge* before generating a new article, or is this something you had to design separately?

2. Knowledge maintenance and updates using AI

- Beyond initial creation, are you using AI to help with maintaining or updating existing knowledge articles?

  (e.g. improving wording, consolidating similar articles, updating outdated content)

- If yes, what does that process look like in practice?

3. KPIs and value measurement

For those already using Now Assist for Knowledge:

- What KPIs or metrics are you using to justify or measure the value?

  For example:

  - Reduction in manual effort for documentation

  - Knowledge reuse rate

  - Incident deflection or MTTR impact

- Were there any metrics you initially expected to track, but turned out to be difficult or unreliable?

4. Things that were harder than expected

- Any pitfalls, governance challenges, or unexpected side effects?

  (e.g. knowledge quality, duplication, ownership, review workload)

We are especially interested in practical experiences (what worked / what didn’t),

rather than purely theoretical or demo-based scenarios.

Thanks in advance for sharing your insights.