Value of AI in ServiceNow

[JakubGiza]

Dear all

I’m looking for ServiceNow community feedback to shape AI Decision Intelligence, which helps ServiceNow customers better understand the AI-to-value proposition, select better LLMs for their use cases, and measure the value they get from AI capabilities used in ServiceNow.

The idea came from a problem where companies can’t clearly establish ROI or value in general from AI utilization. Relying on token spending is cumbersome, and comparing it to value is almost impossible. Managers need a tool that speaks their language, and tokens are not the proper vocabulary.

I have been building a solution on ServiceNow that does not replace ServiceNow AI Control Tower or ServiceNow’s native governance capabilities but it extends them by acting as an independent referee and coach for AI decisions.

It currently brings together three connected areas:

1. Understanding AI Inventory on your instance - an automated, structured overview of the AI capabilities native to ServiceNow, the business activities, connected AI providers, and available models - including gaps in configuration, ownership, pricing, and evidence.
2. AI Model Comparison - customers can compare different models (LLMs native to ServiceNow and external models) that are suitable for a particular business activity, going beyond general benchmarks to consider factors such as output quality, latency, cost, the need for human review, and organizational feedback.
3. Business Value Hub includes metrics such as Velocity of Value, Use-Case Heat Map, LLM adoption, effectiveness, user feedback (in the future, a global report showing which ServiceNow use cases, and in combination with which models, work most effectively), correction effort, cost, and actual outcomes.

I am seeking feedback on whether this direction is valuable for ServiceNow users.

I am grateful for any feedback

Thank you, and have a great day

Jakub

[Deėpak Sharma]

hi @JakubGiza 

This tackles one of the biggest pain points platform owners and enterprise architects are dealing with right now around Now Assist and GenAI adoption.

Right now, leadership asks what they're actually getting from Pro+ or consumption packs, and the out-of-the-box dashboards mostly show vanity metrics—executions, token counts, or raw acceptance rates. A developer or platform lead can't take token burn to a steering committee to justify license costs; they need to show deflected hours, reduced MTTR, and minimal rework.

Happy to help! If this resolved your issue, kindly mark it as the correct answer ✅  and Helpful and close the thread 🔒 so others can benefit too.

Warm Regards,

Deepak Sharma

Community Rising Star 2025

[JakubGiza]

Hello @Deėpak Sharma 

Thank you so much for taking the time to respond. IMHO the missing element, as I see it and you mentiond them too, is the link between an AI-enabled activity and an outcome that the organization actually cares about like time saved, a shorter MTTR, less correction or rework effort, better resolution quality, higher levels of automation, or a financial impact.

That is also the reason why I regard HyperVe AI Decision Intelligence as something complementary to AI Control Tower. The overarching objective is to convert the technical and practical evidence that is already available into a form that different roles in a company can use in order to answer the question of whether this AI capability is in fact generating value and whether the right model is being used for the task.

If you were making the case for the value of AI to managers today, which metrics would you find most essential?

Once again, thank you for your feedback. It was very helpful.

Thank you
Jakub