Best practice for screenshot-based procedures in Advanced Editor

Kohei Tominaga1
Kilo Sage

3 weeks ago

Hi, Everyone!

We typically create How-to articles with a screenshot after each step.


In Advanced Editor, we can reproduce this layout by alternating text and image blocks. However, doing so divides one procedure into many separate text blocks. My concern is that this could affect how well AI Search or Now Assist understands the article as a single, continuous procedure.


Has anyone tested or used this structure in practice?


I would be interested to know:

  • How you structure screenshot-based procedures
  • Whether separate text blocks affect AI understanding
  • Whether one screenshot per step is common in your organization or region

Any examples would be very helpful.

  • 1,143 Views
5 REPLIES 5

sachinvic
Giga Guru

3 weeks ago - last edited 3 weeks ago

@Kohei Tominaga1 

 

 

Example-Based Procedure Guide.png

 

Helpful? Please mark the response as Accepted Solution ✅ and close the thread  — this helps future readers find the solution faster.


Regards,
Sachinvic
✨ArchX Certified | ✨ 16x ServiceNow AI MVP

0 Helpfuls

Kohei Tominaga1
Kilo Sage
In response to sachinvic

3 weeks ago

Hi, @sachinvic 

Thank you for the detailed example. The visual layout is very clear.

 

It looks like the example may use a table or column-based block, with each step placed in a separate row or container.

Wouldn’t this still separate Steps 1–4 structurally, even though they appear to be one continuous procedure visually?

 

My concern is not whether the text is searchable, but whether AI Search or Now Assist can recognize the four separate sections as one ordered procedure.

 

Could you also clarify whether this recommendation is based on an actual implementation or validation test using Advanced Editor, AI Search, or Now Assist, rather than a conceptual or AI-generated example?

If it has been tested, could you share whether the test confirmed that Steps 1–4 were recognized as one continuous, ordered procedure after indexing?

Eoghan Sinnott
Mega Sage
In response to sachinvic

3 weeks ago - last edited 3 weeks ago

I’m concerned that your original response reads as a generic best-practice answer rather than evidence addressing the question that was actually asked. Your reply states that instructions should be available as text rather than contained only in screenshots, but that was not the question.

The question was whether multiple Advanced Editor text blocks are still understood by AI Search or Now Assist as one continuous, ordered procedure. Your original answer said there was no need to avoid that structure, but the follow-up then confirmed that the example was conceptual and had not been validated using the relevant products.

Presenting an untested assumption as a definitive answer, then asking for the response to be accepted as the solution risks misleading community members. This is especially concerning when a "16x ServiceNow AI MVP" credential gives the recommendation additional perceived authority. The question should remain open until someone can provide documented guidance or actual test results.