Reetika Fleming — Executive Research Leader, HFS Research[00:21]
Hi, everyone. I’m Reetika Fleming, Executive Research Leader at HFS. Welcome to another videocast. Today I have the pleasure of welcoming Abhijit Shroff from Soroco to chat with us a little bit about interaction data and the dark side of the moon, and how do we illuminate that. So, Abhijit, I’d love for you to introduce yourself, and then let’s get into it.
Abhijit Shroff — Product Growth Leader, Soroco[00:47]
All right. Thank you, Reetika, and thanks to HFS for the time as well. Hello, everyone. This is Abhijit Shroff. I’ve been with Soroco for the last 4.5 years, working on all of our go-to-market initiatives, and I’m responsible for product growth for the company. I’m happy to share with you, as Reetika mentioned, what we do in the space of interaction data at Soroco and how our customers resonate with it, so that we can share some real-world examples of how this technology is relevant to you as well. So let me begin by taking a sneak peek. All of us have understood interaction data from the construct of how you and I interact with social media sites, be it LinkedIn or Facebook or Twitter, or for that matter any other social media site, on a day-to-day basis, right? We all have our own corporate office time or me time in terms of how we interact with social media sites. So what is it that we collect at the heart of it? We collect clicks and keystrokes, which is basically how I interact with an application, be it Outlook, or when I interact with SAP, or when I interact with an Excel sheet, and I’m doing—let’s say I’m interviewing someone, I’m part of a hire-to-retire process. Now, when I collect clicks and keystrokes, I’m able to capture the gist of how I did that entire process just by looking at clicks and keystrokes. And this interaction data is extremely valuable. The interaction data footprint which we leave within enterprises on a day-to-day basis is 70 times more than what you would leave on a social media site, right? So if our team of, say, 2000 people across my department, I would be leaving close to a billion interactions per year on a day-to-day basis, right? And that is something which is very valuable but which is untapped within organizations. This ability to capture this untapped interaction data to make business sense of how organizations are working, how teams are working, doing either activities or processes—the science is basically where it requires deep tech. And Soroco, at the heart of what it does, applies AI models on top of this interaction data, which is collected by deploying agents on end users’ desktops, to understand how a team works. And its AI models demystify those clicks and keystrokes, which are extremely noisy, into something which is tangible, which can be understood by business leaders or CXOs to state, hey, these are my finance operations processes, and this is how much effort my team is manually spending on it, which is what we call a cost to serve. And the ability for them to understand the ways of working to do one simple task and say, Georgia versus Washington versus California — in terms of saying somebody’s manually raising an invoice, how do these people across different locations work on the same process in potentially the same way or multiple different ways, and what is the cost to serve, which is the effort spent by the team member? In order to understand how much effort is being spent to manually raise an invoice, and then compare and contrast the variations to drive their digital transformation initiatives. So can you optimize the process? Can you automate parts of it? Can you suggest ways in which user training can be imparted to various stakeholders to drive more efficiency in the process? This is what the value of interaction data means to an enterprise. I just gave you an example of how we would apply it to problems of operational excellence, but you could also apply it meaningfully well in terms of solving technology problems, in terms of solving what we call process optimization problems or process reengineering problems. So there are multiple ways in which interaction data is applicable and valuable for enterprises, to different buying centers, be it a CDO, chief digital officer, be it a chief data officer, be it a chief information officer. Interaction data is giving the behavioral aspects of how teams interact with software to drive business operations at scale across a large enterprise, and helps them optimize it. So that’s precisely what Soroco’s interaction data foray is all about. Our product is called Scout.
Reetika Fleming — Executive Research Leader, HFS Research[05:08]
Yeah, that’s what’s so uniquely interesting about this category, right, with enterprise software in process intelligence that we’ve been tracking for a few years now. The world really is your oyster, right? Process optimization, that’s great, but that’s just the starting point, right? There are so many interesting ways in which you can use interaction data to improve your top line, to serve customers better, to have employees have a much better working experience in their day-to-day job, right? Take off the soul-crushing, task-based activities — you know, this is not how this process should run, right? We hear that so often. So what are the most interesting new ways in which this interaction data is starting to be used by your customers?
Abhijit Shroff — Product Growth Leader, Soroco[06:23]
It all depends on how much context you collect. And that’s where the secret sauce lies — the ability to collect more context from those clicks and keystrokes. Now, there are multiple use cases which we have been working on with so many customers. We are understanding the ways in which the interaction data can be applied to varied use cases. Another use case which a CXO of a large firm gave us was: I have this mega exercise which we do every year, because we keep buying tech and we don’t know what part of the tech is being utilized all the time for what purpose. So every year I have a budget to buy software and I also have a budget to retire software and generate savings. Using the interaction data, you have the ability to collect the most granular bit of information around how your applications are interacted with. You can deploy this at scale across thousands of users to understand what the application usage patterns are, what the clusters of business or non-business-critical applications are which are being used. Within those business or non-business-critical applications, which are the features which are being used — so you can actually get your entire application metadata usage catalog, which you have no other way of figuring out apart from interaction data, right? And then you get to know, and you can make a truly informed decision, a data-backed decision, saying, hey, these are the candidates of applications which are potential for sunset versus elimination, versus I need to invest here because this is where my data integration is lacking and maybe I can integrate it here. So that’s another category of use cases. A third very interesting use case, using the same kind of features which I just mentioned above, is customers are leveraging interaction data to drive one of the biggest goals for any CIO who’s in the business of leveraging SAP applications, which is to drive S/4HANA migration. And that is a journey which is a multi-year project and has billions of dollars for large organizations invested in it. How do you get a catalog of all the activities your teams are performing, and how do you ensure that business continuity is not impacted when you actually decide to migrate or not migrate a set of code pieces, right? Which processes do you give priority to when you migrate? How do you ensure that in the S/4HANA world, when you do your fit-to-template workshops with your teams, you are arriving at the right set of use cases or right set of scenarios to ensure that your S/4 testing before you roll it out is complete? It is about cataloging all of the team’s activities. So if you have a day in the life of how a team uses SAP, or a week in the life of how a team uses SAP, and then map it to what can happen in the S/4HANA world, your testing and rollout is all the more complete. It’s not 100%, but it is a lot more than what you were doing before based on conjectures, because you have your actual usage statistics. And then when you deploy it, the biggest question which the business sponsor has is, OK, is this getting adopted? Because you might run two applications in parallel, the old world and the new world. So what is plaguing the adoption? How can I accelerate it? What features should I continuously use? Is it actually meeting my business KPIs? Is my cost per invoice lower in the S/4HANA world versus my cost per invoice in the earlier world because of XYZ interventions which I have taken? Can you measure the cost-to-serve effort for all of these things?
Reetika Fleming — Executive Research Leader, HFS Research[09:52]
That’s interesting, because you’re getting so closely tied now to very, very large technology transformation programs, right? I just feel like this set of technologies, due to this benefit of radical transparency, visibility into your operations, it needs to be step one for any transformation program, right? Whether you’re trying to automate something, you’re trying to migrate to a new platform, you’re trying to rationalize your portfolio — wherever you want to go, you need to first understand where things stand across different markets, teams, and unearth fact-based insights on what is actual best practice versus assumed, ideal processes that you might design up. So I’m loving the different directions that this is going in. One of the biggest evolutions that we’re seeing with AI-based technologies is obviously GenAI. So we’d love to hear what Soroco has been cooking up on the GenAI front. How do you make this stuff easier for enterprises to consume?
Abhijit Shroff — Product Growth Leader, Soroco[11:03]
In all of the examples which I gave earlier in terms of helping understand how teams are working — how can you leverage generative models to summarize, not just show, hey, here’s a step of the process on this application and the screen. Scout is now able to document every activity in a human-readable, business-understandable format akin to a user guide, end to end. So that is one thing which we are doing on our product. The second thing is, as you rightly said, everybody is adopting generative AI, right? And rightly so, because if you look at the automation world, what you were automating was human-machine interactions. If you look at the GenAI world, where you are trying to apply prompt engineering, it is basically again to move human-machine interactions, right? Now, CIOs, CDOs, chief digital officers, they’re spending billions of dollars in terms of getting the Copilot licenses from various vendors and injecting it. It started as a pilot maybe 18 months ago, and many customers are putting in dollars, but they do not know the true value of whether they’re actually getting the ROI, and that’s something which is amiss. So one of the things which Scout’s interaction data technology can actually help pinpoint is: what is the usage of GenAI models in different parts of your organization when you’re using, say, a ChatGPT from Microsoft or a Salesforce Copilot, etc.? What is it that it is being used for? Where is it deployed? What are the ways in which people are using it, in the context of which process it is getting used? How is Copilot actually contributing — more importantly, not contributing, which you think should happen but is not happening, right? So the CEO of a very large Fortune 10 firm basically told us, this is one of my major problems in terms of understanding where my investment dollars are going. We are spending, but I’m not able to see the ROI. Can you help us measure this? So these are the early set of use cases where interaction data is trying to create an impact in helping organizations drive the adoption — or rather, the right-size adoption — of Copilots.
Reetika Fleming — Executive Research Leader, HFS Research[13:12]
Yeah, that’s so interesting, because I hear it all the time — yep, we’ve turned it on. I don’t know who’s using GenAI in our organization, and is it really having an impact, and how do we materially measure this? Also because it’s such an emerging set of skills to work with GenAI in your day-to-day job. What are the best practices? We don’t really have them yet. OK, we are up on time here, but I love this conversation. We talked about the disconnection debt that you guys see implementing your product. We talk about process and culture debt. So I think we’re very much in alignment on — we need to throw so much more data at this. So, OK, I know there’s so much more we can chat about on this subject. But if people are curious, please do come message us on LinkedIn, reach out to myself, to continue this conversation. But thank you so much for watching. Thanks, Abhijit.
Abhijit Shroff — Product Growth Leader, Soroco[14:06]
Thank you all. Thanks, Reetika.
