The drumbeat of pre-Dreamforce announcements from Salesforce continues today with a trio of workflow automation tools. There's a new MuleSoft RPA offering, due to ship next year, based on the recent acquisition of German RPA specialist Servicetrace. Available in pilot, Einstein Document Reader is a machine learning tool that reads data from documents such as driving licenses and I-9 employment eligibility forms. Finally, Digital Process Automation today brings together several existing components so that organizations across industries from financial service and healthcare to telecoms and public sector can build branded digital experiences without the need to write code.
Released as part of the Einstein Automate portfolio, these new tools continue the long-running expansion of Salesforce's workflow automation and low-code offerings, as customers seek to automate previously manual processes. John Kucera, SVP Product Management, who leads the Automation Platform, emphasized the ongoing trend in a pre-briefing with diginomica earlier this week:
We're seeing a huge increase in adoption, whether it's for our guided workflows, whether it's for triggered batch jobs, whether it's for Einstein recommendations. We're seeing huge increases quarter over quarter that hasn't let up. And that's been the case for years. And that's the backdrop on why we continue to invest heavily.
MuleSoft RPA for disconnected data sources
MuleSoft RPA will extend MuleSoft's existing API-based integration pathways with the ability to bring in data from sources that can't present an API. It will be able to access data from disconnected sources such as spreadsheets and PDFs, or from older systems that don't have APIs. Kucera gave the example of a large financial institution that has a dozen different ways of receiving a change address request, including receiving a written request or a fax. MuleSoft already integrates various different systems into Salesforce Service Cloud, where a case is created and the Salesforce Flow automation tool triggers actions based on the policy type, the type of request, and handling the various attachments. But there are four systems which don't have API connections, which is where the RPA tooling comes in. He explains:
They work with RPA for that last-mile automation to make those requests in those other four systems. So it's really the combination of automation, integration, and this robotic process automation that lets them fully automate, end-to-end, these requests.
For many processes, it's often better to automate them as they stand in this way, rather than embark on a more fundamental digital transformation. Kucera elaborates:
What we've found from our customers is that for some of their high-volume, mission-critical [processes], they absolutely want to reimagine it, they want to digitize it completely end-to-end and invest that effort. And then for some of their more rare processes, where there might not be as many times that they occur, they want to still have the ability to automate it, but without potentially the amount of change management and effort.
What we like about this approach is that RPA can handle those scenarios where you might not have that frequency, in a cost-effective manner. You don't have to reimagine the process. You can translate it the way that it is. You can automate where it's at.
We basically have these two options for folks, whether you want to use these enterprise-grade tools to fully reimagine and automate the processes, or where there's bits of it that you want to automate as-is. I think that they're very complementary, and that this allows our customers to take that different approach depending on which process they're automating.
Einstein Document Reader integrates into Salesforce Flow, where a user can point-and-click to set up actions on the data being read in from documents. Out-of-the-box workflows mean that data extracted from documents "can be integrated seamlessly with Salesforce tools without writing a bunch of code," says Kucera. There's a big requirement for this type of capability in industries such as financial services and healthcare for know-your-customer verification and the like.
Digital process automation for industries customers
Digital process automation packages up several existing components so that industries customers have a low-code way to build consumer-friendly digital processes for everything from mortgage applications to self-service telecom bundles. It's based on technology originally developed by industry solutions specialist Vlocity, which Salesforce acquired last year. At the core is OmniStudio, an integrated suite of low-code tools for creating forms and other components, bringing in data from API connections, doing transformations, and other operations. It also adds decision tables, a new capability which lets business analysts create business rules in a spreadsheet format and then upload them — for example in a healthcare scenario, it could detail the drug interactions for a suite of medication, to ensure doctors are presented with the correct next steps. Other components include document automation and a business rules tool that's been developed for public sector. Kucera cites the example of a large government entity that has purchased this solution to replace several thousand PDF-based forms it currently uses to manage various processes. He explains:
They wanted to be able to create these experiences and reimagine the forms into automated processes, without using 20 developers to do so. They believe they're going to save more than 14 developers' worth of full time, just in the creation of it — in addition to the great experience for the consumers, efficiency for the employees in the process, and the integration to all of the other systems.
The announcement is notable for omitting any mention of the recently acquired collaboration tool Slack, but Kucera says the automation team is doing plenty enough with Slack. He elaborates:
Slack is critical to our automation strategy. As an example, I have two teams focused on how can you reimagine Salesforce out-of-the-box processes embedded in that highly productive experience that is Slack. Things like, how can you do approvals — the prototypical thing that is a little bit of work that you don't always need a lot of context for, that you want to be able to do quickly in line in an efficient experience, like Slack. And so those are the types of experiences that we're thinking about and reimagining.
We're also embedding other actions and capabilities. So we're working with our Service Cloud team, as an example, to make it easy to create a swarm channel. And it's not just about having buttons in Slack to go say, 'I want to go create a channel in order to be able to help this customer.' It's letting our customers customize that experience that you also see in Slack. So what's really unique and novel about our automation approach in Slack, is that it's not only bringing those experiences to Slack from Salesforce, doing the posts and the contextual messages. We plan to have capabilities that let you actually customize the rich experience *in* Slack without having to write code, which we think is really differentiated and really powerful.
In every organization, there are any number of processes that haven't changed for ages, generally because they're never considered important enough to gain new investment. Finally, digital tools are reaching a level of maturity where even these processes can now be automated, filling in the 'white space' of mundane operations that never previously merited this kind of attention. This is particularly true in industries that have traditionally been regulated or habitually rules-bound, and it should be no surprise that Salesforce sees plenty of opportunity to bring automation here.
The challenge will be striking the right balance between ad-hoc automation of existing processes versus more thoughtful reimagination of digital operations. Sometimes, as Kucera says, the right choice is also the quick-and-easy route of automating the existing processes. But in other cases, it avoids facing up to the more difficult challenge of grasping the opportunity to adopt a new, more efficient process. As always, the technology is the easy part. What's hard is grappling with the human issues of how to manage change.