Skip to main content
GPTfy - Salesforce Native AI Platform

Classify Cases Before Anyone Opens Them

GPTfy extracts product, type, language, and sentiment from unstructured case text, populates Salesforce fields for automated routing, and retrieves live knowledge base answers — all before an agent touches the record.

Manual case triage slows routing and wastes agent time. This demo shows how GPTfy reads unstructured case text, populates structured routing fields, and searches any knowledge base — so cases arrive in the right queue pre-analyzed and pre-answered.

Capabilities shown in this demo

Intelligent Routing

  • Reads subject and description text to extract product, case type, language, and sentiment.
  • Populates Salesforce case fields so all native routing automation can act on structured values.

Multilingual Case Triage

  • Detects the language of the case submission automatically.
  • Routes submissions to the correct regional queue without requiring bilingual manual triage.

Knowledge Base Retrieval

  • Searches any knowledge source — Salesforce Knowledge, WordPress, PDFs, or custom docs.
  • Returns a recommended resolution to the case record before the agent opens it.

Flow-Based Automation

  • Routing and knowledge retrieval can run automatically on case creation via Salesforce Flow.
  • Agents see pre-classified, pre-answered cases the moment they open the queue.

Use this video when

A mid-sized support team wants to automatically route incoming cases by product and issue type without manual triage

A global enterprise needs to route support submissions in multiple languages to the correct regional queues

A service operations lead wants to attach a recommended resolution to every new case before the first agent touch

A team using Salesforce Knowledge, PDFs, or an external documentation site wants AI to surface the right answer on the case record

An admin wants to enable routing automation via existing Salesforce flows without writing Apex code

A sales ops team wants to apply the same intelligent classification logic to incoming lead or opportunity records

Frequently asked questions

GPTfy reads the unstructured text in a case's subject and description, then extracts structured information: the product referenced, the case type (technical, billing, etc.), the language of the submission, and the customer's sentiment. It populates these as Salesforce field values on the case record, unlocking native routing automation like assignment rules, flows, and escalation.

GPTfy detects the language of the case submission automatically. For organizations with global support teams, this means a case submitted in Italian, Hindi, or Arabic can be routed to the correct regional queue without requiring a bilingual agent to manually triage it first.

Once fields are populated on the Salesforce record, all standard Salesforce automation becomes available — assignment rules, flow triggers, escalation rules, and queue routing. Because the values are now structured field data, any automation that works on a Salesforce record can act on them immediately.

After routing, GPTfy can query a connected knowledge base using the case content. The demo shows this running against a documentation site, but GPTfy's RAG approach is technology-agnostic — it works against Salesforce Knowledge Articles, WordPress, PDFs, or any other knowledge source. Results are returned directly to the case record.

Yes. GPTfy supports Flow-based automation, meaning intelligent routing and knowledge retrieval can execute automatically the moment a case is created — before any agent touches the record. When the agent opens the case, routing fields are already populated and a recommended resolution may already be attached.

The demo uses Service Cloud cases, but the same capability applies to leads, opportunities, or any custom Salesforce object that contains unstructured text. Any object where you want to extract structured attributes from free-form descriptions can benefit from Intelligent Routing.

Ready to see this in your Salesforce org?

Book a 45-minute session and we'll walk through this use case using your own data.

Video transcript
So now let me jump into — let's say the deal closed. We are going through stuff now. We talk about cases right now. We make this transition from sales to support and we are going, alright, here's a case. The customer opened one. So, we are going to see a few very interesting things here. One of the fundamental challenges with cases is case routing. Particularly if you are a large or a mid-sized company, you have stuff coming in. You may want to know what product and what function. Is it technical? Is it a billing issue? What's the challenge that the customer has? And then if you are a large global multinational, you may have to support multiple languages. You may go, hey, based on the language, I may have to send it elsewhere. It's in Italian, I need to send it to my Italy case queue, it's in Hindi or Tamil or whatever language — if you are in India, if it's in Arabic, if it's in whatever language you want — all of that information. So, let's try that with this, right? This is a simple case. It's just got a bunch of information here. I'm gonna go here and run the Intelligent Routing. And what it's gonna do with this is it's gonna pull all the information out of the subject and the description here and populate some of this information. So when I click it, it's gonna look through this — it knows there is a product here. So it's populated the product information here. It thought it was technical support. So technical support — it's got English, it's figured out there's no negative sentiment here, things like that. But notice that we not only extracted information out of the case, sent it to AI and got back a response — which you see here — but we then also took this response and populated fields on this record. The minute we populate these fields, it's a record. You can apply all kinds of automation, you can utilize it with routing rules, you can use it with flows, you can use it with escalation — whatever you would like to do. And it's the same for leads. If you were routing leads this way, you could route opportunities this way. You can route any custom object this way. So, this intelligent routing is something you can apply anywhere by using AI. The key is that all of a sudden we have moved into the sort of unstructured world — this description could have been 350 characters, long bunch of lines, or a lot of information — it could have been an email to case. All of that data AI can process for us very quickly and give us this. So that's great. I know this person is asking a question about data masking, which is a product that we make — cloud compliance or sandbox data masking — and they wanna know what's new. So at this point, if you want something to go against your knowledge base, at cloudcompliance.app/docs, we have a knowledge base. Now, when I click on that button, what it's gonna do is it's actually searching my knowledge base using AI. And it comes back with this information, and this is coming live from my knowledge base. So, it's actually pulling from this knowledge base right here, and it really is technology agnostic. So, it doesn't care if my knowledge base is a Salesforce knowledge base, if it's based on WordPress, if it's some homegrown thing, if it's a bunch of PDFs — it really doesn't care. You can have any kind of knowledge base and go against it. So, you have the opportunity to surface a lot of that information right inside the case record. We support automation, which means once a case is created, you can automatically do these two things even before a person in your organization touches this case. We can run GPTfy, automatically populate all the routing, automatically populate a recommended resolution — and then we can share it with whoever that group is. And they, when they look at it, so much of their time is saved.

Last updated: February 2026