Main content

Unifying knowledge to drive agent productivity - Salesforce's latest Service Cloud enhancement

Stuart Lauchlan Profile picture for user slauchlan May 7, 2024
Knowledge is power - but knowledge is to be found in unstructured formats in multiple locations. Time for some unification.


Salesforce has announced Unified Knowledge, a complementary gen AI-enabled offering to its Data Cloud which can pull in unstructured data from third party sources, such as Google Drive, SharePoint and brand websites. 

The primary use case for Unified Knowledge, which was developed in partnership with Zoomin Software, was cited as Service Cloud, assisting agents to provide better service to customers. But its data integration capabilities can also be applied to Sales Cloud, Health Cloud, Financial Services Cloud and Field Service apps. 

Gen AI features include: 

  • Knowledge Answers. Answers can be generated in chatbots based on a holistic view of information available to a brand. 
  • Search Answers. Answers generated in response to queries, either from agents or customers.
  • Einstein Copilot for Mobile Workers. Non-office-based workers will be able to have conversations with Einstein Copilot, again based on an expanded information set.

Kishan Chetan, General Manager of Salesforce Service Cloud, pitched this as an evolutionary development: 

Service for the longest time has been using AI in a very great way to schedule the call center agents to be sophisticated about making sure that the right rep with the right skills works on the right cases…But we've we've doubled our investments with the advent of generative AI and pivoted the entire company and all of that in our Einstein 1 platform, where we have all this fantastic data within our Data Cloud, all of the applications and our underlying platform. And Service is built on top of it and uses all the richness of that. The goal for us is really with Service to help make agents super-productive or deliver deflection and self-service in a very seamless way. 

What empowers all of this is knowledge - if you don't have the right knowledge, if you don't have the right information about the product or the process, then all of these [customer] experiences fall short, whether it's a user who's trying to find something on a mobile app or on the browser, or an agent providing feedback. That's the reason why this becomes even more important for us…The biggest problem with a lot of AI projects, as exciting as it is, is really data, right? There's a significant difference between an AI project which is consumer-oriented - let's go and do something on a consumer app  - versus business AI, where having the right data becomes super-crucial.

He added: 

If you look at the consumer internet or a lot of the consumer tech right now, without the unstructured data that's there we wouldn't be having ChatGPT the same way, we wouldn't be having Service the same way.

Use cases and data

Chetan pointed to Field Service engineers as a good case in point to demonstrate the need for being able to pull together knowledge sources: 

Several of our largest customers have been technicians who come to your houses, fix problems, install internet, etc. And these Field Service technicians have to keep up with a whole bunch of new products and services. Having that at their fingertips becomes super-crucial. These are the big use cases, where using knowledge just powers all of these use cases and that's where this is a big part of our investment and a continued investment area for us.

There are two parts to the data needed by enterprises, said Chetan: 

One is the structured stuff that sits in your database, such as order histories and your traditional data, and that’s where Data Cloud comes in in a really big way. But then there’s the unstructured data, which is super-useful for Service. Getting all of this data together, organizing it and making it meaningful for a service rep or customer who's coming to find information about your product or services, that becomes super-crucial. So data becomes a big part of it and that’s the reason we've really focused on Unified Intelligence. 

The reality of the matter is that data is going to be in a bunch of different places. It's not going to be in one place. We can't assume that everybody's going to come into Salesforce and say, 'Hey, this is where your data is going'. Data is going to be different places. So how do we get data from different places, - in the case of knowledge, the unstructured data, which is a document which could be in a shared file, shared folder - bringing that all together and using that data to then drive better case resolution, driving better customer satisfaction, making sure an agent who's working on a case is more productive.

As to what scale of productivity increase is involved here, Chetan is talking big numbers: 

These are significant improvements you're speaking of, 90%+ improvement in terms of productivity, and that's huge. One of the things that I always hear from customers is, they're all growing with the great economy, they can't hire fast enough, so getting this type of productivity for them is super-crucial.

He pointed to a gaming customer’s experience with the beta version of Unified Knowledge, where a reply that used to take 15 minutes can now be dealt with in under a minute: 

I was just talking to another customer recently. Their biggest problem was they are growing at 10% They can't hire enough people. How do they get enough productivity from their existing people, so that they can actually solve more and more of the cases that are coming their way? Getting the right answers, getting them quickly, solving the issue, drives customer satisfaction, drives productivity for the user. That’s really big. 

What Salesforce’s latest offering brings to the party is “the plumbing” and Unified Knowledge, according to Chetan, but there’s more: 

We do what it takes to use that data to deliver better generative AI. So if I'm engaging with the customer and I want to have a conversation, having the right contextual information about the problem that the customers raised to us about the product that we have, that becomes super-crucial, and that's where grounding your AI on the right knowledge is important. If you're going into ChatGPT and ask a generic question, you get a generic answer. But the more you give it, the better answer you'll get and that's where our trusted AI becomes super-crucial. 

He concluded: 

It really is not just reacting to customers, but it's also being more predictive, more pro-active, and using that to go and educate the agents.  All of these have complex products, products that keep changing, evolving, and it's really hard for these agents and our customers to keep on top and that's where the knowledge really helps.

My take

There's certainly a strong business need in play here. According to Salesforce's 2024 State of Service report, service organizations are investing in AI to support their agents, but more than three-quarters of executives (76%) struggle to scale AI effectively, largely because of data disconnects. 

A grey colored placeholder image