Inbound Email Ticket Routing – Classification & Assignment Logic

mdshahvez11
Tera Contributor

4 weeks ago

Hello Everyone,

I have a problem statement for which I’m looking for a solution or a practical workaround.

Currently, all inbound emails are received through a single address, and there is no predefined mechanism to decide whether an email should create an Incident or a Request, or determine the appropriate assignment group.

Since users can write any subject or body text, the challenge is to design a rules-based approach that can:

  • Classify tickets using a predefined but manageable keyword list, sender profile, or structured email format.

  • Handle unmatched cases by routing them to an Unclassified queue for manual triage.

The main concerns are:

  • There could be a very large number of possible keywords, making it difficult to maintain an exhaustive list.

  • Even if a list is created, determining the assignment group logic based on these keywords still remains a challenge.

Looking forward to your inputs on how we can balance accuracy with maintainability in this logic.

0 Helpfuls
  • 365 Views
5 REPLIES 5

Stefan Reichelt
Tera Guru
Tera Guru

4 weeks ago

I wonder if Predictive Intelligence could help you here. Do you have a subscription to use it?

0 Helpfuls

mdshahvez11
Tera Contributor
In response to Stefan Reichelt

4 weeks ago

@Stefan Reichelt 

We’re currently on Yokohama ITSM and not using AI/ML or Predictive Intelligence for this case.
Our aim is to design a fully rules-based approach for inbound email routing — something maintainable without dependency on intelligent classification engines.

Any guidance or examples around structuring such deterministic logic would be really helpful.

0 Helpfuls

Laveena-Agarwal
Kilo Sage

4 weeks ago

Hi @mdshahvez11 

In this case , data needs to be analyzed first.

So first the inbound email data needs to be stored in ServiceNow tables and then you can use "Predictive Intelligence" to create solution definition like classification, cluster, similarity.

2. Training and Education - Generally there are communication at organization level to educate the users and trainings needs to be done.

0 Helpfuls

mdshahvez11
Tera Contributor
In response to Laveena-Agarwal

4 weeks ago

@Laveena-Agarwal 

We’re currently on Yokohama ITSM and not using AI/ML or Predictive Intelligence for this case. Our aim is to design a fully rules-based approach for inbound email routing — something maintainable without dependency on intelligent classification engines.

Could you please confirm if Predictive Intelligence is included as part of ITSM by default or if it requires a separate module/license?

0 Helpfuls