The Process & Automation Podcast Episode #17 - How your business can benefit from AI chatbots -...

skill or expertise to address that's possibly beyond that of a chatbot. Or it could be that, you know, as you automate your Tier 1 resolutions for your help desk, of course, you then have capacity in Tier 2 and 3. So, you know, a journey I can think about is with that capacity in mind for human workforce is Somebody talks to the bot and say, I want to talk to a human. It pops up immediately what the person's telephone number is, and you get a phone call back straight away. Now, that's a great, great experience because I've chatted to a chatbot. It's created this, this, this context for, for that human agent to actually progress this case and resolve it. The human actually said, let me just speak to one of your operatives. The operative gets a message, they phone the person back straight away, they've got full context of what's been discussed, and it's a really good timely resolution, you know, of that particular problem. And again, if you look at the parameters of that journey, you know, it's all positive, you know, everything here is positive. There's no broken pieces of the process where things fall through gaps. You know, it's just a very smooth, you know, seamless journey up until that point where the actual technical issue, for example, is resolved.

Very good. Yeah, so as you said, there's a very, very broad spectrum of use cases there. These were just a few, few really good examples we thought. And yeah, so that brings us now into the internal-facing use cases. So the one before was sort of more customer-facing, external-facing, but you can use chatbots as well internally. And yeah, I think that's a completely new dimension I've never thought about before. We're going into this topic a while back. I always thought, okay, chatbots are only useful for customer service and customer journeys, but Yeah. So what do you think?

Well, I think, you know, internal-facing, I think that there's so many opportunities for the internal-facing domain as well, because of course, you know, chatbots are becoming increasingly capable. They are able to assist teams that work with internal employees and, you know, streamlining all of these workflows and saving human employee time. You know, for example, if somebody works remotely and streamlining that experience, you know, that actually ensures that the business keeps on moving forward and, you know, resources is not lost. And, you know, one of the biggest use cases we see is within HR, you know, and HR is a great place to deploy internal-facing chatbot because Let's look at a scenario. Employee just have a simple question, how many holiday leave or holiday days do I have left for the year? And you know, that's a pretty common inquiry. And then normally, you know, they will send an email to HR, somebody in HR will have a look at it. But if you have an HR bot that can field these sort of tier 1 requests, That chatbot can be connected straight away to the HR system. Once we know who the employee is, of course, it's all internal, so there's authentication. The chatbot uses its natural language processor, which is all AI-driven, to interpret, you know, what the actual intent is of a member of staff. In this instance, they are asking you know, how many vacation days do I have left? And the chatbot will give them the answer very quickly, very effective, to the point, you know, and the request is served instantly. Again, you know, that keeps everybody that engages with the bot away from these administrative duties to then wait for the HR email and chase the email potentially because HR didn't come back in a timely fashion. That's one example. Another example is logging or booking a new leave request. Very simple. When do you want to go? Obviously, how many days? And you can make that experience quite exciting as well for the employee. So we see a lot of use cases within Site HR for these internal-facing chatbot-driven assistance for staff.

Yeah, that's really good. And we see as well, so with the lockdown last year when this started, so there was a huge uptake and there still is Microsoft Teams and Slack, all these collaboration platforms, and all what we do there as well with our colleagues, we are communicating in the same way. So I'm writing a message and then something comes back. So it's a typical chat there as well, very context-driven in most cases. We talk about specific, in specific channels in Microsoft Teams about customer tickets which are coming through Teams these days, or specific projects, and And this allows us as well, so by these chatbots, they're not necessarily like this little icon there you see in the bottom right corner in your intranet or on your internet pages. So by combining the chatbot capabilities, probably you're aware of some of the bots firing up in Slack or so, so that they're taking care of a few things. So this is exactly where then this sort of chat capability comes in. You could have a really tight integration into your Microsoft of Teams to actually kick off that conversation. So you can be in the same place as always when you talk directly with your HR team. So the bot can take over exactly the same place and filter down these requests, and everything is in your context, in your own chat. Everything is stored, everything is, yeah, in the right place for you as an employee and for the HR staff. So yeah, it's very, very powerful by bringing the capability and the tools we already use a lot, and MS Teams is a really big one. And yeah, I think we also have that capability these days. We can plug that technology into our Facebook chats, into our WhatsApp chats, in our Telegram or whatever is out there these days. But this allows very, very seamless connection, isn't it? So with your team, same way as they have done before, And yeah, the next big topic really is, it's not just the HR. When we think about all the admin in the business, so the general business administration, there's so much stuff to do which is very time-consuming. And combining that with a chatbot, like scheduling meetings. So this is a good one. Or when, when should we, should we have our next team meeting or our next away day? So then, then the big thing starts, isn't it? So then usually I look into my calendar, I send out a doodle, everyone comes up with a was the right slot. And, but here we could actually schedule a meeting and let the chatbot do it, isn't it? And, uh, yes, it could be very clever.

It, it can do a lot of that kind of, um, sort of fiddly, you know, heavy lifting for you. So you could just focus on actually doing what you're supposed to do at work. And focusing on that. But yeah, I guess for general business administration, again, the use cases are vast. And I guess the reality here is that, you know, all of the backend systems exist, you know, to obviously schedule a meeting. And, you know, this is simply just a question of plugging that chatbot in and publishing that connection. If you want to book a flight, for example, you know, once we start flying again and the skies are open, you know, how convenient would that be to just use an internal chatbot to book your business meeting? It asks you all the questions, it does all the checks for you, you know, the chatbot is trained so that it covers that experience very seamlessly. And also beyond that, you know, once your trip you're coming close to your trip, you can just send you the reminders, ask you questions to make sure that you've, you know, you've got your copies of your itinerary. You know, so that's just another example. So there's so many sort of general arrangements and general admins or administrative tasks that this administrative, you know, assistant can provide you. You know, like you said, you could have integration with Teams and Slack, for example. So again, very broad, you know, a very broad set of use cases, you know, with inside business administration.

And a good thing is similar like the other platforms we're talking about here on this podcast. Yeah, you mentioned this, Arno, as well already. Yeah, they are very low-code focused, so it's more configuration. And yeah, and these technologies we are working with, so they already come up with use huge set of ready-to-go templates. So if you think about, okay, I need an assistant doing these flight bookings, yeah, there is already maybe a rough template there which is getting you started. A few more configurations obviously need to be done to adapt it to the right use case and to the department's needs, but generally it's a good accelerator there available. And this covers all sorts of use cases, not just for the chatbots. I think in general with all the local platforms, RPA platforms. So there's so many templates available to get started in a matter of days, really. Yes. So the next one.

Well, the next one is to do with management assistance. So this is where we have managers and they need to do, you know, a lot of approval-centric type tasks like budgets and requests. And usually, you know, they'll either receive an email with an attachment or they have to rely on an internal platform to access data. No doubt they would have to go to a calendar which has got a reminder for them to either approve the budget or to do something that's key, you know, for their job as a manager. So By deploying chatbots, you can negate the need for managers to do it in this inefficient way and really introduce new levels of productivity for them. So rather to wait for managers to return to the office or to when they're back from holiday, it's very easy for a virtual assistant chatbot just to pop up and really give them all the tools they need to have on the go to do these budget approvals and these management tasks, approve purchase orders, customer reports, you know, and many more types of things that managers does in a business. So again, you know, it is the same sort of topic where, or the same outcome where you free up these employees' time and, you know, you've you free them up to do other more intuitive things, you know, and really improve their happiness and their satisfaction as employees, right?

Absolutely, yeah. So this is pretty much the outcome for most of our automation projects, to really increase that sort of capacity we have in businesses by just getting rid of all this other stuff. Which can be automated. So yeah, but how do we, yeah, was like, was always with these new technologies, how do we really get started with those, with chatbots, how to setting this up? So yeah, it's always indeed, yeah, we need to buy in from everyone and experiment a lot to get it right. But as it is very easy with those templates, I think it is, yeah, you can just try it out very, very quickly and see if it works for you. So, but you said something very good earlier, Arno, was don't make it too human. So yes, and I think you missed that one.

I think, you know, in people's perceptions of chatbots is that, you know, I think when chatbots started off, people did try, people that deploy them did try and make them too human, almost hiding the fact that it's a bot. And I, I think that was the wrong approach and the wrong strategy. I think once that was flipped on its head and people knew it was a chatbot, providing that this bot actually resolved the query, I think most people didn't really mind them. I think in, in the first generations of chatbots that we saw that wasn't AI-driven, and they were more sort of if-then-else, if-then-else driven. You know, it really turned off a lot of people from using them because it didn't take long for people to realize they're not actually speaking to Alex here, they're just speaking to a piece of computer software because bringing the bot out and making it intelligent was actually the right strategy. And I think that, you know, we're seeing that, you know, by deploying these smart chatbots and also declaring the fact that they are actual bots, and there is a possibility if you don't get your query resolved by the bot that you're going to be transferred to a human. People got a lot more comfortable sort of engaging with the chatbot.

Yeah, I guess they understand as well. Okay, here's some mechanism for me which sort of works in most cases or 80% of the cases 24/7. So that's already a huge benefit for me as a customer. So then they accept that, that there's sort of a mechanism, a bot for them there available. And then yeah, I think that's okay. I see it in the same way. I thought first as well from our business perspective, we have a bot, we sort of have to hide that. But what kind of piece of code do you need to have to hide the fact? Yeah, it's too complex. And if you have a very simple chatbot, it's very clear immediately if someone puts something in which doesn't really answer sort of the question and the structure the bot was expecting. It is immediately clear and it's just a dumb piece of code. It already falls apart if someone else asking a a simple bot would ask, please give me my email address, please give us the email address, and then someone answers no, and then the bot is sort of in the loop. Oh, this is not a valid email address, please give me your email address. So this, yeah, you can't really hide that anyway. So only if you have like a super— yeah, and we are not at that stage. So everyone accepts that absolutely. And I think that also brings me to the next point. What's the audience, really how people are interacting? Interacting with the bot and that really that data and that behavior, if that flows back into sort of the learning of your system. So then an AI-powered bot is then even getting more intelligent and can suggest even more things to the customer potentially at some point on its own to come up with the sort of the next best sort of recommendation on solving problems. But in general, yeah, it tells us so much about what's going on currently on the customer side. So if we have suddenly lots of conversations going on about a specific problem, yeah, so we definitely need to, we need to see how our customers are experiencing it and behaving so that that all can, can be very valuable.

Yeah, and that is true. I think That brings us to the next point, is really to start small and using an iterative approach, grow bigger. You know, you can get started with your first chatbot, you know, within a day. There's so many platforms out there that you can use. It's all low-code driven. The better ones has got natural language processing engines, which are very sophisticated, so the intent of the conversation and the context could be extracted very quickly. My recommendation would be to perhaps look at an internal-facing bot first. HR is a very, very good place to start. Simple use cases like, you know, leave management. How many leave days do I have left? What's the policy on sick leave? You know, employee well-being. You know, these are simple things that you can Have a very soft landing for your chatbot with inside your business. And as people become more accustomed to these chatbots, make them more capable, broaden their reach, start rolling them out in other areas, other departments in your business, finance, you know, keep on iterating, keep on learning the lessons. You know, you talked about watching your audience, see how people interact with the bots, adjust these bots. And I think kind of build that confidence, um, that the chatbots are actually a very, very, um, seamless way to get problems resolved and to go get help, um, you know, without having to send out requests and waiting for responses from somebody that's going to be late. And also, um, talking about privacy earlier in, in, in, in, in, in our podcast, I I think, you know, privacy is a big deal. And, you know, we need to make sure that compliance is a priority. You know, we need to make sure that the data is secure. We need to make sure that when we talk to a person, that person is actually the person that they say they are. So all of these protocols needs to be built into your bot. But luckily, With the chatbot platforms out there, that's simply just another construct that you drag and drop into your chatbot configuration to, to tell it how to authenticate a user. While the user is logged on to their local workstation, that user is obviously, you know, their passwords are by definition, you know, it is the person that you think you're talking to, or there could be another verification that you can build in just, just to be on, on, on the safe side. But again, you know, all of the compliancy around data protection, luckily, you know, most of the chatbot platforms we, we, we work with, all of that is solved for, all of that is catered for, and it really just gives you the opportunity to focus on, on these journeys that you want to create with these bots and You know, just, just kind of touching back on where do you start? Start small, land something small. HR, again, really good space, something very simple, get people used to it. And then from there on, you know, grow, grow the reach of the bot.

That's great. Yeah, thank you very much. I think this is a really good, good point to, to, to sort of come, come to in the end with this episode. I hope you, you like this episode, the first one on chatbots. More will follow. It's a very, very interesting topic. And yeah, if you, if you're interested about that more, send us your feedback and then we will plan in more chatbots episodes earlier. And yeah, if you like to have a chat about chatbots, reach out to Arno or myself via our social profiles or visit our website theautomationguys.net for more details. Yeah, thank you very much for listening to the episode today. And yeah, we We hope you, you will tune in next time and see you soon. Unfortunately, that's it again with this episode of the Process and Automation Podcast. If you like this episode, please give us a 5-star rating and don't forget to subscribe to this podcast so you don't miss any upcoming episode. We hope you will tune in next time and until then, let's automate it.