Welcome to our Stellar Keynote Recap Series!
As part of the 1st Annual Pathfinder Community Space Camp & Generative AI Showcase, we’re hosting live sessions with Community MVPs and industry experts to share the latest developments in intelligent automation—especially Generative AI!—and provide you with learnings and resources to drive success at scale. If you don’t have the chance to attend the live session or want to come back to reference some of the critical mission information that was discussed, we’ve captured key intel from each session to share with you!
Day 5 Stellar Keynote is Your Pipeline Lifeline!
For our final keynote session, we’re throwing you a pipeline lifeline! Mark Goodaire, our Director of Automation & Business Transformation, chats about all the different approaches to pipeline, pipeline management, and leveraging tech for your pipeline with Adam Favreau, Sr. Manager of HR Technology at Boston Scientific, and Tim Long, Director of Digital & Data Business Services at The Venetian Resort in Las Vegas, and our live audience delivers some decisive poll responses.
Sourcing Automation Ideas for Your Pipeline
💫 Main Intel: There is not one right method for how you choose to source your automation ideas. Different companies have seen great success at scale using different methods i.e. crowdsourcing, leadership-led, process discovery technology, etc.
Tim: People have different opinions on how to best source automation ideas. Some people like to crowdsource from the bottom up, and some people like to work with the executive team from the top down. My sweet spot is right in the middle. First, we have a digital champion advocacy program with designated people from different functions who have the ability to affect change but also understand the day-to-day pain points for their team. We meet once per month to discuss different innovation idea. From there, we have a steering committee that convenes to review our backlog and archives and take action. We haven’t quite gotten to automated process discovery yet, but I hope to in the future.
Adam: We’re a federated model so all of HR development is done inside of our team, so we rely on the SMEs within HR to submit their ideas. We also go out and teach people what intelligent automation can do so they can start brainstorming ideas they could submit through a formal intake process. We have used technology for process discovery as well. We do a lot with s service center, so much of how we prioritize or target automations is based off volumes that the service center is producing because we pay per transaction, so if we’re targeting the highest level transactions, we’re going to hit the biggest numbers with that.
Mark: What I noticed at our annual Imagine conference last year was several organizations who showcased their automation programs were at different ends of the spectrum—one organization had great success scaling by starting in their boardroom and pinpointing key enterprise cross-functional processes, and another organization put up a portal to completely crowdsource and then selected from those submissions. So there’s definitely different approaches and not one right answer for getting those automation ideas.
SESSION POLL SAYS…
Question: Do you use technology to document your existing/as-is processes?
- 75% of the live audience reported they have NOT used technology for process discovery
Process Discovery & Security Concerns
💫 Main Intel: Process Discovery is a very slick, powerful tool that will give you back valuable insights in a short time, and also has strong capabilities to address any security concerns. But also, being sure to have very clear communication with employees about the intent of the exercise and guardrails around protecting their personal data, is essential to ease concerns and gain willful participation.
Mark: A few quarters ago, we actually put our own Process Discovery tool to work here at Automation Anywhere. We deployed the product to 7 or 8 roles globally and used it in and around our business development. The analysis was really interesting and the data record after only a couple of days already identified how people in different geographical regions were working differently and opened up opportunities for process standardization.
Was there pushback when we initiated process discovery—a sentiment of big brother looking over peoples’ shoulders? Yes, there were absolutely concerns and a nervousness. What we did was to work with our legal team to document policies around this that we could communicate to the whole workforce. We wanted to say we’re doing this because we want to identify inefficiencies, automation opportunities, and ultimately we want to drink our own champagne and be more effective in our work. Most concerns faded away after this honest communication. And then to come full circle, we shared outcomes at the end, we didn’t just disappear. We said, ”Thank you for your support. Here’s what we found, what do you guys think?“
What about sensitive data? This is a valid concern. In our case, depending on what we recorded, there might be some sensitive data. So firstly, we did not start out with a process that had the most sensitive data. Secondly, we worked within the product to ensure we had a full understanding of the process discovery tool and spent time configuring what data would need to be redacted and even what data we didn’t want to record. A lot of times people do personal things on their laptop, like check personal email or do personal web banking. We had a discussion with employees and let them know this is not what we’re trying to record. We gave them the option that if they needed to do something personal use the Edge browser but we’ll be recording the Chrome browser, for example. The product has strong capabilities from a security standpoint like redaction of sensitive data. Making this very clear to employees, as well as the fact that we weren’t trying to make their online banking more efficient, were great steps to take.
Tim: The product also gives employees the chance to review their submissions before they send it. So it’s really at the employees’ control, it’s not a big brother tool
SESSION POLL SAYS…
Question: What is the primary source of your automation ideas?
- 75% of the audience said most of their automation ideas come from the bottom-up - submitted by employees.
You Have Ideas in Your Pipeline…Now What?
💫 Main Intel: Finding a balance between prioritizing big ROI use cases and slotting time to deliver a few easy, quick wins on low-hanging fruit cases for your team is the ideal scenario for many organizations.
Adam: When we have a lot of ideas come in, most of the time we’re looking at ROI potential to evaluate what we actually execute and automate next. But many times, there’s some low-hanging fruit use case ideas where you know there’s not going to be a lot of development involved to get that quick win for the team and maybe save a couple of hours worth of work, so we try to squeeze those in while we’re developing a bigger process.
Tim: The low-hanging fruit is not something now right off the bat, but rather something I attack because I want to build trust and deliver quickly. But as you get into those “home runs”—the medium-complexity processes with huge returns—I typically like to do some low-hanging fruit cases when my developers have down time.
New Tech in Your Pipeline
💫 Main Intel: Staying ahead of the curve and testing new technology for its business impact is important, while also staying focused on your automation roadmap and keeping the lights on. The exciting part of new tech, namely generative AI at the moment, is the value that comes without a bunch of effort or additional overhead.
Tim: I am a huge experimenter. Any chance I get to do a proof-of-concept on a new technology, I’ll do it. You have to keep the lights on, you have to run your department, but you also have to stay ahead of the curve and test those new technologies to evaluate how they can impact business. With ChatGPT specifically, I think the power lies in its ability to enable average business users to build their own automations by just speaking in natural language. They can tell a bot to process an invoice and it will do it right away, or tell it to build an automation that does this-and-that, and it builds the code for you. There’s huge opportunity to enable the business to self-service.
Adam: This week we actually had a meeting and ChatGPT was brought up as to how it’s going to impact our world, so it’s definitely going to be one of those technologies we will look at from a proof-of-concept perspective. But like Tim said, the main goal is to keep the lights on and get through the automation back log we have currently, and then explore how we can leverage other technologies at the same time.
Mark: When I look at generative AI, it should be an accelerator for us. It should have good ROI in terms of taking a lot less effort to do our work. Instead of writing a bunch of custom code in Python or in a bot, it gives me a solution that is cheap and easy to use. You’ve got functionality built right into the product so you can easily consume ChatGPT, for example, and if it’s the right fit, it doesn’t add a bunch of overhead. The exciting part is the value that comes without a bunch of effort. But we do have to mitigate the risk as well—making sure we have the right guard rails and data sensitivity and policies in place.
SESSION POLL SAYS…
Question: What factors do you use to consider when determining which automations to build?
- No surprise here, 89% said they consider Immediate Business Value! Followed closely by Complexity/Effort and Potential Business Value
ROI Now vs. Future Potential
💫 Main Intel: Immediate business value is generally the driver when considering automations, with future potential business value a pleasant benefit.
Mark: In my experience, sometimes we forget about potential business value. We’ve had situations where, in the spirit of agile, we’re trying to get something into production that provides some immediate benefit, but on the roadmap the potential for it may be way bigger—we’re going to roll this automation out to 15 people but we know 100 more are waiting on it. Is that something you guys consider as well?
Adam: Yes. For us, a lot of what we push out is US first and then global. And if we can align globally before we even automation in the US, we look for that as well.
Tim: Me, not such and I’ll tell you why. Venetian used to be part of a global company and unfortunately the tech stacks at each company were very different, so it was hard to replicate processes across regions. We have seen some opportunity where we can reuse the same code across functions, but it’s very rare. We’re also a fixed property and we only have so many rooms, so it’s not something that I consider too much.
💫 Main Intel: There are many channels whereby you can deliver automations—from CoE, citizen developers, third-party partners, etc.—and the best method is what fits your particular organization’s needs.
Adam: We have a federated CoE, so each functional team—finance, HR, etc.—has their own CoE, and projects are prioritized based off of what each team’s needs are. In our HR CoE, 90% of the automations created by developers are sitting within HR. But there are also times when we have automations that are low-hanging fruit and someone that sits inside HR is more of a technical resource that can jump in and do a stretch assignment, similar to what I would call a citizen developer. The person has some technical capabilities and not a ton of experience in intelligent automation, but wants to learn more, so we let them jump on a project to gain experience using the technology as well. I call it “citizen development light.”
Tim: We are a centralized CoE, so my team serves the entire organization. We do have offshore developers that we can kick work over to, and we’re looking at getting into citizen development. We’re waiting to migrate to A360 because I don’t want people building code in version 11, but I do have one person who will be our first test subject for citizen development who specializes in QA automation and is currently getting trained on Automation Anywhere.
Mark: We’re always looking at what the capabilities of that channel are versus the automation. Each of our delivery channels have different capabilities. So, for example, even within our citizen developers, we have some who could join the CoE tomorrow. They're very, very capable. And then we have others that are quite new. So complexity is a key factor in terms of where we would assign certain work. And I'd say risk is another one. When you get into sensitivities around data and P, we like to keep that closer to the CoE just because of the sensitivities and the things that can go wrong around some of those initiatives. At the same time, we’re always trying to keep it as close to where the action is as well - to your point Adam. From a federated standpoint, we want to know where the expertise is on the process, where the expertise is on those systems, a that tends to be closer to the business units that are actually dealing with the processes.
Other Criteria For Prioritizing Use Cases to Automate
💫 Main Intel: It's not just about productivity - there are other types of business value that factor into how an organization prioritizes their pipeline. These factors may be tough to qualify, like compliance or risk mitigation, but are equally as valuable to an organization as something with a high quantifiable ROI.
Adam: During Covid, we scrapped our ROI prioritization model and moved to bandwidth support. We tried to free up capacity to full-time employees by creating automation, so sometime we’d push something ahead of a higher ROI case so we could help support the business.
Tim: In my business, we’re very close to the guests and we’re a service organization. Typically in our steering committees, we present ideas to one of the VPs and it’s ultimately their decision how they want to prioritize, but we do tend to prioritize anything that will result in guest improvement, because our number one focus is on the guest. We track our feedback from guests and if we see areas where we’re receiving bad scores and can build an automation to improve those scores, even if it’s not a great labor savings, we will attack that area over something else that might save a bit more labor.
Mark: It's not just about productivity - there are other types of business value that factor into how each organization prioritizes their pipeline. It might be customer satisfaction or even risk that needs to be mitigation. And these factors may be tough to qualify, but you know it’s something that has to be addressed. For example, if it's related to compliance or regulatory, or speed to market, there may not necessarily be a number tied to it, but it just floated to top priority and is of great importance. Can you relate to that?
Tim: When I was in the the financial side, our operation exceptions were a big deal. If we made an Op error, we would have to pay our customers a certain amount of money—like in the millions—to correct that error. So risk reduction in that job was our primary focus.
SESSION POLL SAYS…
Question: What tools/technology do you currently use to manage your pipeline?
- Top response was “Other” which our audience reported to include quite a diverse toolset: Sharepoint, DevOps, Smart Sheets, Custom Powerapps, among others.
- Second most popular response was Excel spreadsheets at 34%
Pipeline Management Technology
💫 Main Intel: There are several tools and methods to manage your pipeline and the recent release of CoE Manager has powerful capabilities to make managing your pipeline and tracking metrics a breeze.
Tim: We try to work Agile and I’ve found success in some of the Agile project management tools like Jira and DevOps. At The Venetian, we creat user stories that represent projects which have descriptions, ROI, and some value metrics tied to them. I bake that into my dashboard so I can see what I have in the pipeline, what’s in development, and what’s been completed. We used to base the program on repurposable hours, so there’s a call out for that in the dashboard, but now we focus on other benefits that are delineated, like use cases by department. Once we do the initial intake, we go over the process to see if it’s a good candidate and I assign it a confidence level that I base on complexity in relation to the value it provides. So if it’s highly complex, involved a lot of human intuition and process re-engineering, it gets a low confidence. The user stories get attached to tasks and enhancements and bugs, and my team can work in one environment and see everything from a reporting standpoint.
Adam: We use a form for intake that then goes into a Jira backlog, and then we move in sprints through Jira.
Mark: We just recently started using CoE Manager for our internal program and one of the nice things about it is that it’s integrated with our product, which means you can tie back planned versus actual savings and value. So we can say we expected this much value because the automation is going to run 100 times per month, but see that it actually ran 90 times or 150 times per month. That’s a great analysis we’ve been able to leverage.
SESSION POLL SAYS…
Question: Who do you involve in assessing potential automation ideas?
- 80% responded that they include the Requestor of the idea, followed by 69% involve a Process SME, and less than half involve a COE representative.
Assessing Citizen Developers
Adam: When assessing potential strong citizen developer candidates, we typically target people with proficiency in excel doing vlookups or pivot tables. From an automation perspective, when we do the original analysis, we assess how much level of effort there will be on the developer side, either low-hanging fruit or very difficult to automate.
Tim: I’m just getting into this, but the first step is I require them to become developer certified. I like people that have a technical background when just starting out, but I want to give them a process that should their bot break, it will have 0 impact on the business. That’s where we start. I also give them full reign to work with my professional developers to help them.