TRASCRIZIONE 🔗
Sandy: Season two of AI Powered RevOps is proudly sponsored by Attention, the AI platform that helps revenue teams turn every conversation into action. Attention records and analyzes your calls, updates your CRM for you, and gives reps back the hours they used to spend on busy work so they can spend them selling. Visit Attention at Attention.com. welcome to season two of AI-powered RevOps. This season we're cutting through the noise and going straight to the people doing the real work. I'm Sandy Robinson, and this is AI-powered RevOps. Less hype, more real life. Let's get into it. Welcome back to AI Powered RevOps. Today I'm joined by Natalie Furness, CEO of RevOps Automated and Revenue Operations Leader known for connecting people, process, systems, and data to measurable growth. Today we're digging into what it takes to make AI useful inside RevOps. If you're trying to turn AI into measurable impact, then keep listening. Natalie, welcome to the show.
Natalie Furness: Thank you so much for having me, Sandy. It's an absolute pleasure.
Sandy: I'm so excited to talk to you today. first, why don't you tell everybody a little bit about you and your company?
Natalie Furness: Yeah, well, a little bit about me and my background. I actually come from an operations background, like pure clinical operations actually. A lot of people say, How did you get there? Well, I mean, if anyone has seen healthcare systems, they're pretty damn broken. And trying to align people, processes, and data systems across healthcare, that's kind of where I started working on big projects in Her Majesty's Prison Service and the NHS, the big one in the UK. And I kind of found myself hitting like that kind of like ceiling these people aren't ready to change yet. So I
Sandy: Mm-hmm.
Natalie Furness: started looking at where I could add true value. And I found that if I took my love of operating systems, alignment of systems and processes and my love for commercial growth, I ended up in revenue operations. And I started working first within businesses before going fractional myself. And I suppose I accidentally started a revenue operations consultancy because during lockdown we just got so busy. And before I knew it,
Sandy: Mm-hmm.
Natalie Furness: I'd hired one, two, three, four. more people and started applying the RevOps systems that we know well to be able to scale Revots Automated. And today we are, I think, considered the largest rebots consulting partner in Europe.
Sandy: Wow, that is amazing. And and that was really on the on the outset of COVID, how you got started with everything?
Natalie Furness: Yeah, that's how RevOps Automated really grew that
Sandy: Mm-hmm.
Natalie Furness: first wave. It was it's quite interesting 'cause when I was researching revenue operations, and the kind of like term and of course doing my SEO and and how we're gonna launch the business. This was back in the day in which when you search Google for revenue operations, there was about two pages on Google before
Sandy: Mm-hmm.
Natalie Furness: it turned into like DevOps. so I I right place, right time as we say with product market fit. but from then we harnessed that and and grew from there. And now we're serving businesses in Fortune 500, in the US, in the UK, from Christian-based parish businesses to charities to reinsurance businesses. And it's been absolutely fascinating.
Sandy: Wow. Wow. that's so amazing. And just for you to have accomplished that and grow your business the way you have, it is just, it's really awesome. So I'm excited to hear from you too, because you get to work with all these businesses, all these different people, people in RevOps and different roles in companies. So I think you bring really in the line of work that you do a different perspective, right? So I work at a company. So I have the perspective of our, enterprise systems and the things that we work with and and maybe some of my prior orgs. but you really bring that broader perspective. So I'm excited to dive in today.
Natalie Furness: Excited too?
Sandy: This podcast is AI powered RevOps, but I don't think you can do anything in RevOps without saying the word AI. It's just really part of RevOps now. what are you seeing out there right now that feels genuinely useful versus just, you know, noise?
Natalie Furness: Gosh. I can start with noise. That's probably the easiest one. This whole idea of
Sandy: Ha ha.
Natalie Furness: AI SDRs, somebody please stop them. I mean
Sandy: Ha ha ha.
Natalie Furness: but to be fair, they are do you know what though? I say that and then I actually look and be like, they are starting to get sophisticated. I actually got emailed by somebody, well something, about a board position, of which I replied to and then the AI, I realized after a few emails it was AI, referred me to someone else that was then involved in the conversation and they were C C
Sandy: Mm-hmm.
Natalie Furness: they didn't manage to CC each other in though. They could they could only do separate
Sandy: Mm.
Natalie Furness: threads. So that was quite interesting. I would say like please stop doing anything with AI without the correct context. I think that's the big
Sandy: Mm.
Natalie Furness: issue we're seeing in general. I would say allowing your Teams to just query your CRM database using whatever terms they like to try and find the answers to a report in which they don't know how to prompt. Please let's try and figure out a way to stop people from doing this. And you will
Sandy: Mm.
Natalie Furness: know if you're in Ravots how frustrating it is when you've worked for millions of hours on a beautiful dashboard and then someone queries the underlying data and tells you that your dashboards are wrong.
Sandy: my God. my God. Can we just pause there so I can vent for a minute? Because that
Natalie Furness: Please don't.
Sandy: is so freaking real. Okay. dashboards are like nobody uses them because they can just ask Claude. It's like, I mean, I spent the
Load full transcript
better part of my first year in my role building like pretty solid kick-ass dashboards that are like the source of truth reference. And, you know, everybody's pulling it from Claude and and It is it is a problematic, and without that really strong context layer, you're going to get shit. that's just the bottom line. You're going to get shit. And then you've got people who have no business pulling numbers. you want them out selling, right? And they're sitting here pulling numbers and you're like, hey, can you, can you just come to me for that? Like, I know you can pull it from your little Claude, but Yeah, it's a huge problem and I'm glad to know it's you know, it's not it's not just a a me problem, but you know, we're working on steps to solve it, but it's it's just really hard unless you totally lock people down.
Natalie Furness: Yeah, and I suppose that's where your skills or agents come in. If if if you understand
Sandy: Mm-hmm.
Natalie Furness: what people want regularly and the reports that they regularly want insights on, you can build skills to help them query the data correctly. So I would say that's one of the
Sandy: But but they still have to tag the skill. they still have to do a thing, right? you
Natalie Furness: Yeah.
Sandy: can you can build skills, but if somebody doesn't know what it's called or know how to enact it or choose to, it's like my kid has a bicycle helmet that I tell him to wear when
Natalie Furness: Mm-hmm.
Sandy: he rides his bike. I give it to him, I got the right size, I got all this stuff, but you know, he consistently rides off and doesn't put the helmet on. I mean, like, you know, can't make you can't make him do it.
Natalie Furness: Yeah, a hundred percent. Yeah, yeah. And behavioral that's cool. Yeah, behavioral change is always a big thing. I'd say from a RevOps
Sandy: But anyway, all right. I'm done venting. Yep.
Natalie Furness: perspective, like where I've personally, most recently in the last few w weeks, I'm thinking back to where I've experienced the most like the great thing with AI. I think being able to process really large data sets, providing it's given the right context layer and be able to integr interrogate them. And help you with data enrichment, data deduplication, all of those things where we might have had some data dedupe tools where you had to match a URL or match a an exact company name, which we might have had to code in one at a time. Whereas now you can actually like get the AI to go and like really research and really question like what's the likelihood that these are the the same companies or the same contacts and like what's the reason for it. And it can actually reason and explain and give you like a confidence level, which then you can automatically, improve the the the health of your data sets. I think as well most recently I'm working as VP of RevOps for quite a large property I A sorry. I'm working at RevOps. my gosh, let me pause a second. I'm currently working as the VP of revenue operations for a prop tech company who are their AI first in terms of their product as AI and they all use AI. Every single person in that business is very active on Claude. But I will say they have bought an AI consulting firm in who isn't they're not from a Roblox background, but they have helped enable them better than I would say I've seen at a lot of other businesses. But one thing we were doing is we were really trying to categorize ICPs into target markets using a very specific nuanced set of rules, which,
Sandy: Mm-hmm.
Natalie Furness: you know, is not matched to you can't find it on Apollo, you can't you can't find it on like LEM list, you can't find it on any sort of like any single data. It's their definitions and the way they want to categorize it.
Sandy: Mm-hmm.
Natalie Furness: So we spent a long time working with the team who are experts in the property space to be able to create those different market types. So we're able to create segments, sub segments, asset classes, rule out PE firms, things like that, which have allowed us to create like not only a great TAM, but also map kind of target account tiers. So then we can build better territories and then actually rotate the target. accounts that matter to the reps and the BDRs to work, which hopefully will result in a lot more pipeline. So it's it's kind of thinking about like in the back of that. We've also done
Sandy: Mm-hmm.
Natalie Furness: quite a a lot of use of agents with clay as well. I I'm just sure many of you listeners are are familiar with clay. I actually was like not super into clay until I actually found a GTM engineer which I now love working with. I would say it it's like any tool, it's not the tool, it's how you yield it. and together
Sandy: Right.
Natalie Furness: with one of our na new team members, we built out this GTM engineering function, which we're now working with businesses to do a lot of searching for very specific profiles of people within companies using agents that can look for very specific things. So it's not a case of running a LinkedIn sales navigator list where you say, I mean, if any of you have done that, you'll know it's frustrating. Certain seniorities
Sandy: Mm-hmm.
Natalie Furness: job title a certain department and and it's always the filters are just rubbish. So it's
Sandy: Yes, ugh.
Natalie Furness: much nicer when you can do a a people search based on very specific AI rules and use a a bot to kind of like well, an agent to go and look at all the different data sources you have. Yes, they do call them clean. Yes. Yeah, yeah.
Sandy: Clagin. It's a clagiant. The Clagents. Yes. They yeah, they call them clagians. No, I love it. I want to hit on something. You you made a really good point. So the first part about the ICP work. that's just ground floor critical RevOps. RevOps needs to own that and truly diving down into your ICP, into that. into the details, not just the basics, really and building it for AI, right? So building
those descriptions. you know, we went through an exercise like that recently and everything that I built, I built this brain basically, it's like a matrix of all of our different segments, but not not there's not just descriptions. Like there's a column that has the, you know, the AI prompts To go find it. And then that's what we use to power clay. So then I gave that to my clay engineer and said, like, hey, okay, I'm not gonna build out clay, but here's, you know, here's the brain and here's how, here's how how to build it. So I think that's important to really do that work and understand your ICP and then your RevOps people to actually understand and use their brain. while they're doing it and learn it, not just do the work and let AI spit it out, but you need to make these decisions as you're building the stuff out. So I just I think that's just such a a good reminder to do that work. and the de dupe that you mentioned and the customer and prospect database project essentially and just getting everything like super dialed in.
Natalie Furness: Yeah, there there are a few of the things I've seen. another one that we're working on at the minute, which I'm interesting to see how this goes. I many of you will probably work at Slack based businesses. of which people no longer do anything in the CRM. They just like to have Slack channels and and talk about things on Slack or, you know, teams, whatever you're using. A big one though is I have been thinking about this a lot. There's a like gong channels, you know where people celebrate like a deal is closed and they give loads
Sandy: Yeah.
Natalie Furness: of information on a Slack channel but they don't really
Sandy: Yeah. Yeah, ring the bell.
Natalie Furness: Yeah but this is this is not something we want to stop people from doing because they all celebrate each other's success. So what we've been looking at is how we can get some of that contextual information that they've just not put into HubSpot into HubSpot using agents to be able to scrape data from specific LinkedIn sorry not LinkedIn channels from specific Slack channels and bring them back into HubSpot. there are of course some challenges with that because of course if people just say, I spoke to Jenny and she said yes, woo-hoo, we won,
Sandy: Mm-hmm.
Natalie Furness: then the the AI has got to ask, like, okay, well, who is Jenny? What's is Jenny associated to a specific deal? Like, what is that? And so we're kind of working on the logic of of how do we get that information back into the CRM? So that's that's a project we're working on that I think will provide us with some really great value if we can figure it out. Of course, you can you can backwrite stuff from Slack into your CRMs if you've got specific channels for deals, but not if there's an aggregate channel.
Sandy: Right. Nobody does that. They they do a messy, they do messy chat, trust me. yeah. I mean, I think that is a yeah, they just start a chat, then they add somebody, add somebody, add somebody, add somebody. And then if you're not in it, you don't you don't know about it either. So how do you how do you connect the dots? Yeah, that is that is definitely a I think a very common problem, whether it's Slack or Teams and you know I I don't know if it's better because we have teams in Salesforce, which suck together. They're not a match made in heaven. you know, but I would imagine Slack and HubSpot isn't a great match either because Slack's owned by Salesforce. Like how is that?
Natalie Furness: We weirdly, there is a lot of integrations between the two. and they actually weirdly aren't that bad together. Although I actually I
Sandy: Okay.
Natalie Furness: have to say I've never worked with Slack and Salesforce together since Salesforce have
Sandy: Mm.
Natalie Furness: owned Slack. So I I can't tell you what the state is of of that integration versus it with HubSpot. But there's a lot I would say there's a lot of developers that are like love creating apps on the HubSpot side, so they they tend to build whatever we need. So that's all good. Another one, and I think this is an oldie but still a goodie, is deal intelligence and automated enrichment of medpic or ban or spiced or whatever you're using based on cool transcripts to be able to improve
Sandy: Mm-hmm.
Natalie Furness: forecasting accuracy on deals. I cannot believe the number of businesses who still have not subscribed to any form of cool intelligence. It blows my
Sandy: No.
Natalie Furness: mind. Like I think we as a business have had call it's the first tool I think I purchased as a business owner after my CRM was like, okay, we need to record every one of our calls, not only for the intelligence for deals, but for the intelligence for marketing to be able to pull case studies from. we also record our client staff calls as well. So we can look for common trends of common complaints, common churn risks,
Sandy: Mm-hmm.
Natalie Furness: common common learning. So and then we can create all of our education around that. people who aren't recording today, it blows my mind. So I think AI transcripts are probably like the one of the most under under kind of mentioned like value prop that we we've got out there.
Sandy: I a hundred percent agree with that. it's same same for me in terms of the importance of that and across outside of just the go-to-market teams, as you mentioned, because if somebody's on a support call, you want to get that data, that intelligence data. I've got mine locked down so it just records external. So it's it does not record any internal. kimono is open, so anyone can access any call across the company. And anybody identified as even maybe once in a while having a meeting with a client by themselves has to have it. and so we're able to capture, you know, I mean, I I'm assuming we're getting probably around 90, 95% of like all external like calls because and then you you're right. I mean, there's so much you can do with that. How are your
how are you seeing that used? updating fields in HubSpot or Salesforce, product stuff?
Natalie Furness: yeah.
Sandy: what are you seeing there?
Natalie Furness: Ugh, I mean the main the main stuff that we're seeing is for those who are using it is in deal enrichment and potentially
Sandy: Mm.
Natalie Furness: like being able to move companies in qualification sta like deal qualification stages. Cause you know like sometimes like a rep will like do a call They won't update the CRM. It will sit there for a week. It's still in stage. When actually, pretty much all of MedPick was known. It should be qualified. Its forecast stages should
Sandy: Mm.
Natalie Furness: have been updated. And by the next forecast meeting, everybody's questioning what happened. They're like, yeah, yeah, that's progressed. I just haven't updated it. Like you know
Sandy: Mm. Yeah.
Natalie Furness: what I mean? You can do all of that before the next forecast meeting. I think that is potentially where I'm seeing it used the most in RevOps. There are so many other places I would like to see it used. of course, if if our sales teams were coached, ideally by their previous calls and like getting scorecards and things like that, I am seeing use of
Sandy: Mm.
Natalie Furness: scorecards with AI, which should be helping coach the the reps, because if the reps perform better, we all perform better. so that stuff. I actually the business I'm working for at the minute has recently invested in some AI sales coaching software where
Sandy: Mm-hmm.
Natalie Furness: we feed all of the previous history of that particular rep and also all of our playbooks into the AI coach. And the idea being is that the sales rep will have a more frank and honest open practice with AI because it's not a human. So actually in a way it kind of makes them feel more comfortable. Well that's the theory anyway. I maybe we'll have to do a second session so I can update you on how that goes.
Sandy: I know, I know, I wanna hear. what's the tool?
Natalie Furness: Do you know what? I'm gonna have to share it with you after this or you're gonna have to drop the link. So
Sandy: Okay. Okay. Yeah, just curious. going back to that context layer, you mentioned some skills and things like that, but you know, how how do you really how do really help them solve that problem? because it seems to be like the real the biggest thing right now that people are talking about is is just kind of, you know, all the great ICP work. Like are you having them, you know, put it in SharePoints and, you know, referencing there? Like what what kind of things are you, you know, are you thinking about?
Natalie Furness: So for yeah, I I'm I've always been like a keen documenter, like document everything. Like I just
Sandy: Mm-hmm.
Natalie Furness: think it's so critical. I mean, typically we will try and work with the tools that our clients are most familiar with. When I'm talking about this
Sandy: Right.
Natalie Furness: recent engagement I'm on, they're a big notion company. And of course, Notion
Sandy: Okay.
Natalie Furness: like straight into Claude and we you know that they are big on Claude and Notion. So anytime that a definition of a segment or a sub segment or an asset class changes, we are updating it in Notion so that that can be the context layer. We'll also even like I'm big on any workflows that have been built in the CRMs. If we say like this is a tiering model, this is the logic of the tiering model, and this is the link to the place in which it is built. Of course, we don't make it necessarily viewable and editable by anyone. But at least having the all the information there, not just this is the rules, but this is where it sits in the CRM.
Sandy: Mm.
Natalie Furness: And here's the link to it if you need to look it up. So, like thinking about the fact that if you were wanting to run any form of agent on it, what is the agent going to need to know? And and I suppose as a Revox person, you need to think, well, You want to give it the same level of context that if you were to have a junior member of staff and you were to say, look, I want you to go and update this tiering model. Here's a notion page. Go do it. I
Sandy: Mm.
Natalie Furness: think if we all think like that and think like that's the level of detail of documentation we need to do, we can start building up more of a context layer so AI can do some of those things. But I suppose that also brings me onto the kind of the topic of when we start building out all these context layers and gives AI the job to do that we'd usually give our junior teams jobs to do, where does that leave us? But I think that's that's a whole different conversation.
Sandy: Right, right. But you make a great point. I mean you have to think about that. You have to think about it as a a a new a new person at the company or a junior person that you're you're really s mapping it out for step by step by step as if they don't know nothing. No don't know nothing. Don't know anything, if I can speak properly.
Natalie Furness: Yeah, well we we're both having one today. You got me with LinkedIn channels,
Sandy: They don't know nothin'.
Natalie Furness: no Slack channels. But yeah,
Sandy: Mm-hmm.
Natalie Furness: this also brings me back to a conversation I was having with it wasn't I can't remember his name off the top of my head, but he was an AI thought leader. This was a couple of years ago, and he was talking about the importance of the context layer and like literally giving AI access to everything so it can have a better context layer, even to the point Of if you want your AI to behave a human, you should be bringing them into meetings and talking to them like they are. which which has freaked out a few people that I've spoken to, but I I got it. Like, if if you're expecting AI to give you answers or reason based on information, surely it should have access. To that same information. And while we're working,
Sandy: Mm.
Natalie Furness: actually a
lot of us in a hybrid world in which we are jumping on calls and and recording our meetings and having clawed AI, that's great. But actually I think the companies that will struggle are the ones that are all face to face and having do you know those corridor chats where everyone shares context
Sandy: Mm.
Natalie Furness: without it being recorded? I I think that Yeah.
Sandy: But ex and then we'll all be wearing wearables around, right? Like, I've seen
Natalie Furness: Yeah, yeah.
Sandy: what was it, plod or something? They had a booth at a conference. I was like, there's gotta be so many laws you're violating, like walking around just like recording everything. But but yeah, no, I I I agree. I mean, I think it's Having the transcripts, it's not even the video recording, it's just can I get the transcript and dump the transcript into you know, essentially a knowledge base type of a situation. our head of marketing, when he started at the company, he did these founder interviews and he just recorded them hours and hours, and we got all the transcripts, and that's kind of the basis for. some of this training that we're building right now. we just are repurposing these transcripts and it's like the founders of the company just their brain dump and their 30 years of experience in the industry. and it's building out, you know, our our onboarding and our training and, you know, some of these different I've got a new head of go to market enablement who is like, you know, making podcasts out of it. She's doing all kinds of crazy shit. And But yeah, getting those transcripts even for internal meetings, different use case, right? You're using those to train, to train agents or people. but you know, you bring a a really, a really good point there is getting as much of that information and the the relevant new stuff, not the old stuff that's you know, like three years old and confluence.
Natalie Furness: Exactly. And I I've actually been exploring and I suppose I'm putting my CEO hat on now. we talked about the fact
Sandy: Mm-hmm.
Natalie Furness: that I'm I'm working two days a week fractionally as a VP of RevOps whilst running my business, whilst trying to give everyone my attention, whilst bringing up a a young child. so and I'm working trying to trying to fit it into four
Sandy: You do it all, you do it all, and very well.
Natalie Furness: days a week. bless you, that's really kind. But I am exploring using Delphi to build a digital Duplicate of myself so that if I allow
Sandy: Mm.
Natalie Furness: it to ingest everything in terms of my writing, I mean blog a lot, I post on LinkedIn a lot, I'm I love to share information freely. I've been on podcasts. If I can put everything into there and and perhaps create a Slack version of myself, that when I'm not there and someone wants to know what I have would say or how how I would respond to something, I'm curious to see if it's possible. so yeah. Yeah.
Sandy: I freaking love this. I'm gonna have to have your digital twin come back to the podcast, maybe.
Natalie Furness: my gosh, yeah. Well I I have been do you know what? That's one thing I would really personally love to learn myself is to how to make some better AI video versions of myself. Cause
Sandy: Mm.
Natalie Furness: to be honest, I've tried a few times and they all suck. But I've seen some good ones after. So like hopefully they'll be better by then. But at least we'll start
Sandy: Yeah.
Natalie Furness: with the the Slack stuff, which then makes me think, well, if I can train a digital version of myself which has all of the context layer. And I then linked it into a CRM, would my digital version of myself actually be able to update the CRM the exactly the same way that I would've? that's
Sandy: Ha ha.
Natalie Furness: that's a crazy deep dive, but who knows?
Sandy: That's freaking I love it. I think you should do it. I'm very excited for this. I think it'll be great. I mean, so I mean, think about people that are in seat and they're trying to figure out some sort of a small AI project that can have a big impact or big optimization. Can you give some recommendations there or talk about what that looks like?
Natalie Furness: Yeah, for sure. I've developed a system called the Revenue Outcome Qualification to help people identify what's worth working on. It is a case of understanding the reality and the outcome that you want to achieve from the change and quantify if the change is even like worth doing it. So, like what's the actual value of it? Is it urgent? Who actually needs to own it? So are there other stakeholders that how many stakeholders actually need to be involved with this? Because let's face it, if someone can block it, just go with something that's easy. And then at the end, validate. Like make sure that like everything that you're thinking about doing is worth doing. And I'll give a very specific example of of something that I worked on with a a finance team recently. So we were kind of looking across the business and looking at like To be honest, it was a hiring need. And maybe this is a place to start. We were the business was looking at hiring another accounts associate or finance associate. That's fine. I also start again. bless you. get some water, Sandy.
Sandy: I was trying to find the mute button so I didn't interrupt you.
Natalie Furness: That's fine. I remember where I remember where it was. That's fine.
Sandy: my god. Hold on. I was doing good. Okay.
Natalie Furness: You were doing so good.
Sandy: And my voice just gave out on me. It's like the itchy a little itchy in the back of your See if I can get through this.
Natalie Furness: Cool. So
Sandy: Whew. All right, go for it.
Natalie Furness: It was actually based on a hiring need. This business was looking to hire another accounting associate or accounts administrator. And I suppose the first question was, is before spending I don't know how much it would equivalent to in the US, but before spending 30,000 pounds, which
might be $50,000 or whatever, on another person, like, let's break down like why we need this additional head. So actually, maybe it's worth thinking about like what hires you're gonna be making, and then thinking, like, okay, well, what is it that is becoming laborious and a problem? Because they were looking for quite a junior person. And we identified that the biggest like time-consuming thing was actually just Categorizing expenses into very specific categories, which was being done manually using their financial tool. And one of the biggest issues was staff weren't sending on their receipts for expenses. And of course, there's loads of platforms out there that can help you manage this, but this business wasn't large enough to be able to afford one of these platforms. So what we looked at doing instead is building out a categorization of all of the different categories. creating definitions, getting access to all the different emails using Google Workspace, getting agents to go and search for any information that looks like it might be a receipt, pushing that, like pushing that forwards into a DEX platform using the API. And then then from there, like making sure it was categorized based on the definitions. And of course it still needed the the accounts person to finally say yes. But rather than going back and doing all the manual looking and searching and asking people and chasing up people for like their receipts and then having to then think about how to categorize them by researching perhaps what they spent with, that was all done for them. And actually it reduced the need for that head at all. So the outcome was is a saving if you I mean, love to times things over ten years. So you look at a cost saving of five hundred thousand over ten years because you had basically reduce the need for that hire at that time. So that's how I like to look at things. find something that's like a change or or a spend that's about to happen or or when people feel like they're spending too much. And then try and look at well, how much money or time can I save by using AI agent? and it was more that kind of context layer that research it was able to go and look at the companies which the receipts were from and then be able to categorize based on that. Now of course it comes up with confidence levels as well. So, we can't always say it's gonna be a hundred percent, but it it minimized the workload by about eighty percent for the accounts person.
Sandy: Wow. Wow. That's huge. And nobody's gonna say no to that. So if if somebody just starts working on that, nobody's going to say no to that sort of impact. And you made a very important point that you're starting with the outcome that you're trying to achieve, right? And what problem you're trying to solve and you know ultimately where you're wanting this to go, just like any other project, any other, small or large project that you would approach. I think that's an important thing to remember because I think people get caught up in, well, the AI agent is the, it's that's like the focus versus the underlying thing you're trying to do. That's just really one of the, you know, avenues to get there, right? Or a piece of the overall equation to get the outcome you're trying to achieve.
Natalie Furness: Yeah. And so start with the revenue outcome and work backwards from there because by identifying and the reven revenue outcome doesn't have to be revenue, it could be time because time is money. Because let's face it, if you reduce workload, then you reduce the risk of needing another head. but also, you know, money as well is always good. And you can put a financial amount on time once you know what someone's time is worth. and then get the reality and then get try and figure out the smallest thing that could be changed to make the biggest impact. Because I think I did at least two hours of interviewing with the financial team to really go step by step. Tell me what you do. What do you what's this typical day look like for you? What does the processes look like for you? I think we even went as far as going like, what okay, who in the business. has the most receipts that's causing you the most pain. So we could actually start even on testing on one specific email with one set of receipts, go really deep and fix
Sandy: Right.
Natalie Furness: the smallest change for the biggest impact is always best in RevOps.
Sandy: I like the way you're thinking about that. And that leads me to ROI in AI now. And I know this is kind of a cliche
Natalie Furness: Mm-hmm.
Sandy: thing, like what's the ROI on AI? And but what metrics tell you an agent or automation is actually creating value, not just, you know, well, it was a few you know, a few less clicks to get a job done type of thing.
Natalie Furness: Yeah. This is so challenging. And I think this might be the biggest thing. I I was at an event with quite a lot of private equity investors and and partners the other week and they were talking to me about like, we've got all this AI, but how do we actually like measure the impact? The biggest thing is did you collect a baseline? And I think this is like the biggest thing that we're like all so busy building that we forget To collect the baseline. And that's the first R of the revenue outcome qualification is reality. what is the quantifiable reality today? Because if if you don't have a quantifiable reality, then you can't measure the change. It's like any scientific experiment. If you just, you know, run an experiment without collecting a baseline, how do you know if there's statistical significance? You don't. So
Sandy: Right, right.
Natalie Furness: so before you build your next agent, start with. A hypothesis. I am building this agent because the current state is this. We hypothesize that if this is successful, the outcome will be why. You know, like do that. And like feel free to guesstimate. I think this is also a problem with Revolts people. We love to be so black
and white and matter of a fact and have like an
Sandy: Mm-hmm.
Natalie Furness: answer that's like so on point that we don't like to forecast something that we can't guarantee. But like sales are doing it, like RevOps, you need to be doing it. Like they're happy to like hang their hat on a number, which sometimes they hit, sometimes they don't. And it's okay for you to do that too. So find something, tell them what you're gonna change. Say like this is what we're gonna measure at the beginning and the end. And no one's gonna ask you to do this. We have to if we don't collect the baseline, it's our own fault. And I think we do
Sandy: Right.
Natalie Furness: need to own that. so yeah, next time you think about building something, collect a baseline, look at the difference, and then do you know what? You can then write an impact study to say what you've done and present it to the board and then get lovely lots more budget to do your next AI project.
Sandy: like that. Lovely lots more budget. That is always
Natalie Furness: Yeah.
Sandy: the key, especially as we're thinking about budgets for next year at this point. I think a lot of us, but really, really great points. And it's just back to the basics, right? Measure, take a baseline. We're measuring change. I mean that is that's what we do in revenue operations, right? We measure change. So just as a as a foundational best practice, always take a baseline and understand what what the difference is going to be. In terms of when we think about as as we're thinking about AEO. So AEO is is you know obviously what we should all be thinking about. how how are you incorporating that as part of the fabric of the revenue operating system, not necessarily just a content tactic.
Natalie Furness: gosh, that's such a quick question. from From the AAO point of view, we're talking about AI agents being able to find businesses the best way they possibly can via the internet. and we are we have been running some experiments at RevOps Automated ourselves. of which please everybody, if you haven't got a plain text, how did you hear about us, please create it because or at least update your dropdowns to include both Clawed and you know anthropic or or whatever you want to add. But The plain text hills are are the best. we've actually started seeing people who are writing. We found you via Chat BGPT, we found you via Claude, which has been really interesting. We have ourselves been working with a AEO agency specifically on that. I would highly recommend Content Mender. and in terms of like the technical side as well, like you just need to be making sure that you're preparing all of those landing pages to be able to to deal with it. There's definitely some custom
Sandy: Mm-hmm.
Natalie Furness: scripts that they're writing in the background. But in terms of a content stuff and visibility, I don't know how many people know that HubSpot has an AEO module within it where it
Sandy: Okay.
Natalie Furness: actually allows you
to like write what you want to be found for and the prompts you want to be found for. And it actually leverages AI to spit out not only what you should write but who your biggest competitors are. in your industry
Sandy: Mm.
Natalie Furness: and you can actually see their visibility over your visibility in there, which is really interesting because I ran it recently and it it
Sandy: Interesting.
Natalie Furness: actually got our competitors spot on.
Sandy: Wow.
Natalie Furness: so yeah, that's a tool I'd recommend. But I'd be curious to hear from your point of view. Have you approached anything specifically?
Sandy: Yeah, I mean, I think I I think our you know, we work with the marketing team to make sure that everything that we're doing is optimized for AEO. I think right now we're I mean we're we're in the rebuilding of the complete website like phase. So everything is kind of being done from scratch. but it's It's obviously, going to continue to grow and going to continue to be a bl place to focus. you know, in terms of those leads coming back in, right? how are you how are you handling that? So there, you know, mentioned the different, you know, web pages and landing pages and things like that. Is that what you're is that what you're kind of focusing on?
Natalie Furness: Yeah, it's primarily on on the on the web pages and the content optimization. We've got like
Sandy: Mm-hmm.
Natalie Furness: I'm seeing a lot of like blogs that are are written. You also need to make sure you've DAIF your own blogs. not only that, but like looking at third party platforms. So of course LinkedIn
Sandy: Mm-hmm.
Natalie Furness: is a big one. I mean, I'm blessed in the sense of I've been writing on LinkedIn since I don't know, two thousand and nine or something. So the history history
Sandy: I know, right.
Natalie Furness: goes like way back. And of course we look back that it is considered a platform of authority. So if
Sandy: Mm-hmm.
Natalie Furness: people are talking about your business on it, Reddit scores really highly as well. less blessed in that sense. I wasn't a Reddit person growing up. So it means that I do not have the the the kudos or karma I think they call it to be able to to do lots of posting on there. But I mean there's a whole business model around people happy to post about your business on there.
Sandy: Right.
Natalie Furness: so I think it's It's less about backlinks and more about reputable content across third party platforms, from what I understand. But I wouldn't say I'm an expert in marketing. This is what marketers have very much told me and and
Sandy: Right.
Natalie Furness: we will support them with that. but yeah, so attribution to it, like making sure you have that plain text field. So then of course, when you're looking at attribution to revenue, you can then source that this was found by AEO. I think
Sandy: Mm.
Natalie Furness: A RevOps perspective, making sure you're collecting that source at the beginning is just so key because
if you want to know if it's worth investing more in AEO, that's how you have to do it. We are seeing though a lot of spam coming from it. A lot. I don't know if you're seeing the same, but
Sandy: Mm, mm. Interesting.
Natalie Furness: like I would say we've never had a point, and this is speaking about my business personally, we've actually never had a point in which we've had so many people fill out our contact us forms. to try and sell us rather than inquire. We're getting more inquiries,
Sandy: Mm. Mm.
Natalie Furness: but then we're getting more of the that stuff as well. So
Sandy: Yeah. That's interesting. Yeah, we're building or or I'm calling it our inbound 2.0. So we have everything set up in clay, but we're optimizing it to flush all of that out, whether it's a competitor, because we're getting a lot more competitors that are that are coming in and hitting our DQQ a lot. more vendors and things like that than I think ever before that are that are coming through. And
Natalie Furness: Mm mm mm.
Sandy: so our our new routing and workflow in clay is just contemplating all of that. Not just like here's our, you know, ICP and put that into the right buckets, but what do we do with all of these other, you know, 10 buckets now that we're seeing come in, right? and get them. And some of them could be legitimate. partners potentially. So you have to be real careful when you're sifting through. this could actually be a partner for us and not a vendor. So we're really refining that that that logic and that you know using the clay clayents to sift through that with that, you know, kind of matrix I was telling you about. So but yeah, we did we we've been seeing that too, I think. And and I didn't even think about that potentially being from AEO, from people finding it in,
Natalie Furness: Yeah.
Sandy: chat GPT or something. But interesting.
Natalie Furness: Yeah, exactly. Because because people that are looking to sell to you are also doing the same thing as you looking to sell to other people. So
Sandy: Right.
Natalie Furness: and the question is is is whether these people even filling out this information are human as well. So I think yeah, I love that idea of of like the auto enrichment following form fill out because of course you can get an agent to do the research your S Dr
Sandy: Mm-hmm.
Natalie Furness: would usually do following a form complete. Like you don't need a human to do that. So I love that you
Sandy: yeah.
Natalie Furness: yeah.
Sandy: yeah. It's coming with the full brief and everything and the task and and yeah, it's a it's a pretty good workflow. But but yeah, so well this is this is great. I would love to hear maybe some parting words of wisdom for folks sitting in a RevOps seat like myself, from you and in your experience and working with RevOps leaders.
Natalie Furness: Yeah, I'd say There's there's so many words. I'm just trying to make sure that these resonate. Don't forget your fundamentals. remember, your purpose is to align people processes and data systems. So don't forget your fundamentals. Make sure you build things that produce a metric that matters. Because if you're building for the sake of building, you're going to be seen as an expense, not as a a a team that returns on investment. And Make sure anything you're building and anything that you are working on, you are socializing. I think keeping an open project map, socializing what you're working on, looping in stakeholders and keeping them well informed and updated on how it will benefit them. Remember, like, we have to sell the RevOps function to other departments. You know, we do add value. And make sure as you're doing things, you are celebrating your success. Just as much as sales are celebrating the gong channel in your Slash or Teams
Sandy: Yeah.
Natalie Furness: or whatever they're doing when they close a deal. Remember to celebrate. We reduce the need for another head. We improve forecast like accuracy by. We have filtered out this many spam leads because we've done an auto enrichment, which has saved the business this much time in SDR research, which is an equivalent of this dollar value. let's tie everything
Sandy: Mm.
Natalie Furness: back to revenue impact because there is so many stats that suggest that the RevOps function should produce an a ROI of 100%. And that's out there. Mark Roberg said it's ex CRO of HubSpot. So like We need to be evidencing that we are producing a hundred percent ROI.
Sandy: That is such a great way to frame it and to think about Well, Natalie, where can everybody find you?
Natalie Furness: Well, to be honest, I pretty much live on LinkedIn. It yes, it's me. It's not AI. I perhaps have a problem, but I do love to hear from people. So please do reach out, like comment on some stuff. I'll happily comment back. Maybe we can have a chat on LinkedIn. I can be found in the UK. If you're ever visiting from over the seas, please do get in touch. I have some great like restaurant recommendations, bit of a foodie. otherwise, yeah, follow me on LinkedIn.
Sandy: Natalie, thank you so much for joining me and sharing such practical perspective on AI and RevOps. And thanks to everyone listening to AI Powered RevOps. If you found this conversation useful, follow the show and share the episode. We'll see you next time.
Natalie Furness: Thanks so.
Sandy: I'm Sandy Robinson, and this is AI-powered RevOps. Less hype, more real life. The views and opinions expressed on AI powered RevOps are those of the host and guests and do not necessarily reflect those of their employers, affiliated organizations, or sponsors. All editorial content remains independent. AI powered RevOps is a Sandy Elevates LLC production.