From Conversation to Case Action: Automating ServiceNow Workflows from Live Voice with Vonage

Rob Wilkinson speaks with Jonathan Kershaw, Director of Product Management at Vonage, about how live voice can bring real-time voice transcription and context directly into ServiceNow to trigger workflows during the call, not after it.

Kershaw explains where the pain shows up most clearly: agents toggling between up to nine systems, inconsistent case notes, delayed case creation, and AI outcomes that fall short because they are missing the most important customer context. He shares practical scenarios, from IT outages to payment and fraud-related calls, showing how native voice in ServiceNow can support intent-based routing, mid-call escalations, warm handoffs powered by AI summaries, and faster wrap-up with automated dispositions.

The conversation also covers what to automate first, how to prove ROI using credible baselines, and which KPIs tend to move earliest, including average handle time, after-call work, data quality, and reductions in follow-up contacts. The core message is straightforward: treat AI like any other business tool. Start with measurable operational gains, then scale.

For more Customer Experience tech news visit CX Today.

Transcript

SPEAKER_02 0:09

Hello and welcome. I'm Rob Wilkinson, and today we're taking a closer look at how live voice can trigger real action inside ServiceNow. So if you're challenged with slow case handling, inconsistent updates, or teams switching between too many tools, stay with us because you'll get some practical takeaways that you can implement in your organisation. I'm joined by Jonathan Kershaw, Director of Product Management at VonAudge. He's an expert in this space. He's got a very clear view of the decisions, the risks, and the opportunities that leaders are facing today. So it's great to have him with us. Welcome, Jonathan. Thanks very much for joining me. Thank you very much for having me and great to be here. Fantastic. Now, before we jump into the detail, let's just for the audience's benefit, can you kind of set the scene a little bit? Give us a kind of what's the background to this topic? Why does it kind of matter right now for service organizations?

SPEAKER_00 1:05

Yeah, no, I think that's a good idea, just to level set. So I think without stating the obvious, you know, customer expectations have basically outpaced existing systems. That isn't new, but I think that is accelerating. So contact volumes, handle times are up, contact center managers multiple manage multiple channels, sorry, you know, without unified infrastructure. And although that isn't new, and people have been saying that for you know quite some time, it becomes more urgent now because platforms like Service now becoming customer service backbones. So everybody's experiencing this. And also with AI Matured to act on real-time data. And I think also, you know, with Gartner predicting a gentic AI resolving 80% of issues by what, 2029, I think they said, but that requires reliable data, you know, which is missing from uh often missing from voice, and the channels, you know, which is the most important channel and complex channel, but without that real-time voice and data, you know, workflows are operating on a delay. So to have those put together, you know, that's when we really need to, what we need to address to try and give benefits to, you know, organizations looking at you know this sort of topic.

SPEAKER_02 2:20

So where are the teams right now feeling the pressure most directly? Where does it kind of turn up? Kind of in layman's terms or playing English, what's that real cost if people keep things manual?

SPEAKER_00 2:35

Yeah, I think that's a good question. And I guess there's no harm to keeping things manual. But as we said earlier, customer expectations are rising, and therefore the pressure on customer experience kind of manifests in three areas. So I think the first level is kind of the agent level. Agents waste time and risk errors, toggling between, say, up to nine systems, you know, that's a recognized sort of number, you know, which harms both their productivity and also the customer uh and agent experience. You know, the frustration gets to the agent level. Then you get to the workflow level, you know, without these real-time voice data, case creation occurs post-call, you know, which can lead to incomplete, inconsistent, and mistyped notes, which then in turn leads to poor categorization, missing context in escalations, longer resolution cycles, and more repeated contacts, you know, and that all negates the benefits of tools like ServiceNow. And then lastly, you know, AI, you know, people have expectations. This is the future competitive differentiator. So it's recognized that siloed systems hinder seamless AI-driven experiences, you know, and generative AI like Now Assist is only as good as the context it receives. So missing voice data results in incomplete summaries, inaccurate next best actions, and poor recommendations. You know, and manual processes where it's not just current inefficiencies but broken future AI strategy. So hopefully that's you know, plain English, you know, because it's agent workflow and AI level, that's where the pain is, you know, hitting people in the customer experience and their pocket, really.

SPEAKER_02 4:21

It's so true. And when you look at the kind of opportunities that AI and and the new uh automation capabilities that we now have, when you look at those, it's kind of like the answer to a lot of historical issues that we've experienced in the operations, as you as you point out there, and and it's really kind of exciting about kind of the possible. Um, so before we kind of jump into what that looks like and what the capabilities are, I'd like to understand from you, because you you obviously deal with companies every day. Um, in addition to kind of um what what are the costs of the businesses, how does it show up for customers and the frontline agents when things stay manual versus the opportunities to improve stuff like today's presented to us?

SPEAKER_00 5:10

So I suppose in a way, you know, what we need to consider is you know those uh you know, how voice triggers, you know, workflow in ServiceNow. So what does that, you know, in a way, what does that actually mean in practice? So to try and bring that to life, you know, if I give you just a simple use case type scenario, so if we imagine you know, business critical IT out outage, you know, uh call highlights the contrast between old and new workflows. So let's you know try and imagine that that difference. So historically, agents would manually search for accounts, locate assets, create and categorize incidents in Service Now, and update records. Um, you know, and that causes problems and frucks friction, you know, and then they'll summarize this. But now when you start to put voice natively inside Service Now, you know, with intelligent routing, instantly queries, you know, things like a CMDB, you know, using the caller's identity to match them to incidents and assets. So now the agent can see the full customer context immediately, eliminating manual search. With that real-time transcription, we can now feed into you know, now assist. And also the incident auto-updates during the call. So we can take advantage of things like flow designer, can trigger mid-call escalation. So you were sort of saying what can happen during this, you know, so updates to field uh engineering, and you might look at you know, wrap-up time, AI generating the summary and suggested dispositions, you know, so it just makes a big, big difference across the entire interaction, you know, including that voice conversation. And it's all captured in the ServiceNow record. It's not multiple records and things like that. So really you get the result when you look at the old way of doing it and the new way of doing it, you're trying to remove that manual case creation, increasing higher record accuracy, and much richer data for AI, which starts to feed um back on itself. So hopefully, you know, you can see those benefits to the business from that point of view when you start to bring the technology and workflow together.

SPEAKER_02 7:22

So I guess when you trigger those workflows in service now, what can you walk through um kind of call-to-close scenario? What does that look like? How does it play out in practice?

SPEAKER_00 7:36

Um yeah, of course. So in that one, again, sort of a similar sort of scenario. Imagine somebody is contacting you with you know a problem with payments on their credit card, or there's an issue with their bank account, you know, it's locked, whatever it might be. A human agent would capture all that information historically, then they would sort of maybe take notes, then they'd have to wait for a human agent in the back office, somebody who was, you know, rely or creating all those fraud rules. Are people trying to, you know, create fraud? Are they trying to get into somebody else's account? They'd then have to transfer the call to that agent or back office expert or consult. You'd have to spend time uh explaining the situation, and then that person can pick it all up. Now, instead of doing that, and that can take sort of 10 minutes to solve it, now you can capture everything in what would have historically been the IVR, but now you'd call it an IVA, capture all that information there. The human agent at the front understands all the intent, brings everything together in one go. And now, should they need to transfer, because obviously with AI that's not necessary, but you transfer, then actually you can do mid-call summary, do a warm handoff, but using AI to drive the warm handoff, not the human. So, what that then gives you is the first agent can drop and get ready to handle the next call, and the second agent, the expert, gets all the context, much better experience, and all the and the customer feels they're looked after and handed over because the second agent has picked it all up. So hopefully you can see a difference between all the clear blocks that people had to face historically, and now how AI can join everything up, which is why we talk about the context, the voice triggering workflows, all those sorts of things. So instead of being isolated elements of technology, everything comes together, thus you know, generating um, you know, business benefits, which I think we'll be touching on a little bit later.

SPEAKER_02 9:50

Oh, we've got we've got to, absolutely, yes. Um before we do, I think um what that we've it's really interesting at the moment with ServiceNow's recent announcements. Um there's now so many options and so many different things that you can do with the automations and the workflows. I guess it's almost like we're spoiled for choice now. From your expert opinion, someone who's worked with companies doing this, what are the kind of things that you would prioritize? So the first two or three automations you recommend that people start with, um, something that's realistic but that will have a good impact to kind of prove the case, if you like.

SPEAKER_00 10:32

So, in effect, you know, I guess key um uh you know automations, those sorts of things. So I suppose, you know, one would be you know, could you automatically create that incident or case, you know, within there rather than having to click on the create case, pre-populate it? You have all the context, you know, with voice driving that. You know what it's about, you know what the intent is. So pre-populate the database records. It's a yeah, it's a it's a quick win. Second, we know who it is. One of the key pain points is being transferred to the wrong person again and again. So let's understand that intent at the start and get it to the right person. And that's the beauty of voice. It understands what I'm saying, what my intent is, not just one for this, two for that. Where there's always a one and a half or a two and a half. This is intent. This is what I want to solve. And then I think the other one is um, you know, post-call summarization or auto wrap-up, where you can actually manually measure, you know, classic KPIs like how long did that average handle time take? How long was the after call work? These are things that you can measure and get clear success from. So, can we do things like summarize the call? You know, and after all, after call work is typically what, 20% of a call, roughly. You know, so anything where you can reduce that amount of time and increase improved data, like what was the call about, is going to drive benefits. You know, just a disposition code. If that can be automated automatically, then we all know. Um, and maybe I've been to the wrong contact centers or the right contact centers in life, agents will always just choose the easiest one. You know, whereas now we're getting the right one without any agents making the time. So, yes, with you know, announcements, what now assist is doing, this is where it all starts to come together as a much more holistic experience.

SPEAKER_02 12:37

You're you're absolutely right. I I think the audience will absolutely appreciate the uh the disposition issue. Uh it's just one of those classics. You just cannot rely on on the data coming out of the disposition codes, unfortunately. As much as the problems haven't been done. No, no, no. So let's let's automate that. Let's let's let let's let AI do that bit. Take it as it's a job that doesn't get done, move it away. It's it's a it's a prime example of really kind of low-hanging fruit for really easy sort of wins, isn't it? So no, I love that. Um you touched there on being able to kind of uh measure some improvements and through through kind of uh making some of these changes, and you touched on like AHD as an example of something that can get reduced. How should someone who is rolling something like this out in their operation, how should what should they do to prove that they're actually getting results? So are this are there the better KPIs that they should be watching, or is it more around signals that that that show up? What's your experience in in kind of in the real world?

SPEAKER_00 13:42

Uh I think you've got touched on some great points there in the real world, you know, not just the dream that is AI. We know that you know a huge percentage of AI projects fail. Um so we're great advocators of you know doing it, properly solving it, should we call it old school problems, you know, like you know, CSAT, AHT, those sorts of things, and take AI as something that can actually make a tangible difference. So, for example, you know, where we want to prove it's working in key KPIs, as you say, you know, is instead of lagging indicators, you know, we want to focus on things like CSAT, you know, which is a much more leading indicator uh of success. Um, and one of the things that will drive that is your classic average handle time, you know, is the most immediate signal. You know, so anything that can reduce 5 to 10% is a credible target. So as I said earlier, after call work is a key percentage of the entire call, 20%. So if you can just reduce after call work time down by 5 to 10%, and we typically find it can be 30%, you know, from some of our customer deployments, then that should drop sharply with you know, like automated summarization and disposition codes. And we talked about data quality going up. Um, you know, and also that reduces the agent admin and also softer KPIs like attrition, you know, you do get fed up of doing the same things. So actually, you need to think of, you know, churn. I mean, I've worked with contact centers where the staff churn rate was over 100%, which blew my mind that something could be over 100% because people are going through that fast. So if you can reduce your staff attrition rates and therefore the onboarding, all those sorts of things, that's got to be great. And then other things like, you know, if you can populate case notes more comprehensively with summary notes, then that starts to get a qualitative metric that becomes quantitative, you know, instead of just the they called about billing, you know, well, really, but actually you can start to measure the length of the notes. Um, and as a result of that, you know, the first contact resolution might not change, but what you might start to see is uh reduction in you know escalations to second line, things like that, because agents are being forced to think about what the call was because they know what they say is being summarized, so they consider it and they can't just go, oh, let's transcribe, uh, sorry, let's um transfer, because it will it be in the notes. So they'll focus on first contact resolution. Also, should you need to, and no one wants to have a follow-up call, you know, end-to-end, then actually should there be a second follow-up call, all the history is there from the previous one and it's good notes. People would have said, oh, well, the history was there before, but it's a case with a disposition code. Now the agent has, or the second agent has more details. So if first contact resolution hasn't gone up, then actually, should we call it second time resolution or follow-up calls should start to reduce and get better because there's more details for the second agent to deal with it. So, you know, I think you know, just by looking at that, you can start to see improvements within the first quarter. AHT goes down, after call work goes down, and potentially, you know, second calls start to drop as well, which ultimately leads to capacity. That is, you know, the gold standard within a contact center. So they would be some of the areas and KPIs that I would certainly be uh looking at. But they are achievable, attainable. I'm not telling you that you know, suddenly you can deflect 80% of all your calls with our bot. You know, these are the things that prove AI work. You've done that, then you can go into your next AI project knowing that there is ROI and success can be achieved.

SPEAKER_02 17:53

Yeah, I think you're from from some of the stats I've seen, I think you're um you it's it's very credible, but I think you're you are kind of being reserved in your in your figures there. I think there are some really kind of big numbers out there that you can that that are true and are having really big impacts. And just that whole kind of being able to uh transcribe and and copy and paste uh it saves so much time in in the real world. So thank you for thank you for that. Um I guess I'd like to ask around what shifts you're noticing uh in the enterprises that you're talking to now, um, about what they're thinking about service now, because there's a lot of announcements like we've we've talked around, they are kind of going deeper into sort of customer servicing, sorry, sorry, customer facing service, and obviously um agent AKI is a big part of that, um, those announcements. So, what are your customers, what are your prospects saying uh and thinking about uh they might, you know, opportunities they might see?

SPEAKER_00 18:57

Yeah, no, I think that that's a good one. It is a big shift because obviously ServiceNow are taking advantage of their dominance in that, you know, the old classic ITSM space and pushing into the CRM space or CSM space. And I think that's interesting customer service management, you know, not relationship. You know, they're trying to differentiate themselves that way, which is good, um, you know, and challenging and reshaping those platform choices. And I think because of that, you know, they're starting to see economic uh benefits. You'll have fewer vendors, everything's joined up, there's less integration to do, um, both at the CRM level and the telephony level, and it now starts to drive that unified system of record, which is great. Um, also, the other bit, I think, is because people have seen uh ServiceNow is successful as a system of action, not just as a system of record, then why not apply that sort of logic and mentality to your customers' problems? Yeah, but outside facing customer problems, not internal customers, because it's proven those sorts of things. And actually, with the adoption of AI and taking advantage of workflows, why not just expose some of those things that you built internally, externally, because the at a high level they are the same thing. My password doesn't work, you know, where's my order? But as an employee, I'd be going, where's my laptop?

SPEAKER_01 20:27

Yeah.

SPEAKER_00 20:28

So build on that. And I think from a, I talked about the unification, you know, Gartner as an example, you know, notes the shift from service teams to AI leadership teams, you know, where the human roles govern that automation, you know, which is you know, requiring that complete context, which we were talking about earlier. And again, it's the native voice within it. So there is that shift, not only just, shall we call it, cross-cell ITSM into CSM, but vertically, how can the channels make a you know a big, big difference? So I think ServiceNow also, you know, can demonstrate you know, ROI and embedding that voice becomes an easier way to do it, you know. And within that, you know, I what we're seeing is certainly the metrics are starting to improve as a result of that shift because they've taken what they've done before with ITSM, applying it to CSM. And also there's an element of IT are um obviously involved in AI projects, they're clearly involved in ITSM. Um, but because of the AI bit, they're starting to influence how and what the success looks like through AI for a CSM market. So it's building on some of those strengths that people already have within those businesses and knowledge.

SPEAKER_02 21:48

It it certainly is evolving. Um it was always going to with the you know the advent of AI, and it's been so fast and it's so interesting to be in this, you know, this watch. It happened, which is great. We're running out of time. I could talk to you about this all day. But we're running out of time. So I guess one last one then. What if if you had to give one practical step to our audience, to someone who's um you know looking to use a platform and um take you know this month to move from what are we doing, should we do this or not, to actually doing something about it? What what's a tangible way to do that?

SPEAKER_00 22:27

Yeah, okay. I think that's a fair question. And I think build a credible ROI. And the key word here is credible because people have just said, put in ROI, sorry, put in AI, and it will be great. But it has to be credible, but you have to baseline it and prioritize it. So dare I say it, you know, I'm old enough to know when TQM was a good thing. Um, but actually simple things like time and motion studies on your highest volume, routine calls, and I'm not saying every agent and every person, but just take, you know, maybe a cohort and actually see what they're doing within those calls, um, see what they're um spending their time on. So basically map and quantify every manual step from answer call to quote close, you know, and then when you realize all those different steps, as you said earlier, copy and paste, well, that's just two seconds, you know. But if you then have a contact center, um, you know, of for I can't say 100 agents over a month, those two seconds equate to one FTE or two FTEs, which ties back into your earlier question of you know, what differences do we need to make, be competitive? Well, actually, it's the cost to serve, the time for, you know, to wait to be answered. That's what this means. So back to your question of the ROI, you do your time and motion study of where are the long bits, understand that if we got rid of these, you know, copy and paste moments, the handover, all that sort of stuff, you can see where ROI actually kicks in, because you get rid of that. And as such, you know, you might find 40 to 60 percent of routine handle time is you know handled through AI, you know, and that becomes your automation target, really. So then as a result of that, you know, implementation starts to become much faster because you know what you're doing, um, rather than trying to find a result. So I know it sounds old school, but ROI is old school, and one of the things people are finally realizing is AI is a tool like all other business tools, and you just need to justify it and all those sorts of things. So, yeah, just go old school, ROI, but baseline where you are so you can prove it.

SPEAKER_02 24:46

I think it's a really good point because we we we change things all the time in this space. So we we're used to change, we're used to rolling out new technology, cranky. If any industry does it, it's ours, right? So I think you're absolutely right. We we just because you know the technology changing doesn't mean we need to reinvent how we deploy and we should just trust you know the old school ways of doing things. So I love that. Uh unfortunately though, uh Jonathan, that is all we've got time for. Um, thanks again for joining me uh and answering all my questions. Before we do properly close, though, for anyone watching this who wants to explore this in more detail or you know, find out how to get in contact with your teams. What's that what's the best way to do that?

SPEAKER_00 25:28

Uh in that case, I would go on to the website. We've got contact us uh there. That will go right through to the appropriate teams who will get back to you automatically. We also have, dare I say, our own AI on there which will answer stuff. And also follow us on things like LinkedIn where you'll find out, you know, what we're actually up to, you know, individual people. But yeah, the website will give you all the answers that um you should need. That's great.

SPEAKER_02 25:58

Um, and don't forget, um, you can also find a wealth of related resources, uh, stories, videos like this one um at cxtoday.com. Uh but for now, that does wrap things up. Uh, I'm Rob Wilkinson from CX Today. Uh, thanks very much for joining us.