AI for Sales Teams: From Manual Workflows to Autonomous Pipelines

Charlie Cowan: Good afternoon! Let me go, right, share my screen, share my screen, and then that's there. A couple of weeks ago, I forgot to actually hit share until I was 10 minutes in and was talking to myself. And where do we want to go? We want to go here. There we go. Good afternoon, good morning, depending on where you're dialing in from, and welcome, to this week's summary edition of, the Kowalah Wednesday webinar. If this is your first Wednesday webinar, we get together every Wednesday, same time, and have an hour to talk about all things AI, and specifically, stuff that either we're seeing internally here at Kowalah, things that we're seeing in our clients and projects that we're working on, or new things that have been released. There's so much going on, and, you know, we're experiencing this, and we're in AI the whole time, and we're enjoying struggling to keep up.

Charlie Cowan: And this hour gives you an opportunity to get our view on what's going on and what's important in the world of AI. So we're gonna follow our same process. We'll take about 35 or 40 minutes talking about the main topic of the day, which I'm going to introduce in a minute, talking about company AI, specifically for sales and sales leaders. We'll then have about 20 minutes when I'll talk about the latest releases and updates over the last 7 days in the world of AI. And then we've got an opportunity for some Q&A as well. Now, what I would say is that if you've got questions that come up while I'm talking about the sales AI topic or AI in the news, absolutely feel free to ask your question. We're in a Zoom meeting.

Charlie Cowan: And, so if you have a look, depending on what device you're on, either at the top or bottom of your screen, you'll see there is a chat panel, and I'll open that up on my other screen. And there's also a Q&A panel if you've got a sort of... well, I don't have to say it's a more formal question, but anyway, there's a Q&A panel there. And then my colleague Caitlin is on the call. Caitlin, who is a human, we have to sometimes verify this, because a lot of people think that all we have is bots. So Caitlin is human, and she's keeping an eye on the Q&A as well, and will, you know, keep me in the right direction if I missed your question. So with that, let's get into the main topic of today, which is primarily around AI for sales, and sales leaders is what I'm really thinking about in my mind.

Charlie Cowan: And one of the reasons why we wanted to run this webinar is that over the past few months, we've been running a lot of training programs, coaching programs for various teams, and one of those sets of teams in clients is sales teams. And, what we're seeing in so many organizations is, well, we've given people a tool. Now, that might be Gemini, it might be Copilot, it might be Claude, it might be ChatGPT. We've just given people stuff, and we're hoping that some magic is gonna happen. And we're not seeing that magic happen. And so, I have this lovely phrase here, everyone on your team is trying to figure out AI on their own. That's not a strategy. It's chaos with a co-pilot subscription. And now I say co-pilot because it sounded quite nice with the hard C, but it could have been Claude, it could have been ChatGPT, it could have been Gemini.

Charlie Cowan: The key point here is not about the individual technology. It's about, we're just gonna give people access to a tool. And we are going to expect the people to do something different, or to figure out what to do with it, and then magic is going to happen. And in many organizations, magic is not happening, and that is the focus of this session here. And so, just broader than sales, but sales is the focus here, but broader than sales, we can think about AI in two ways. So one is AI for the people. We've got a thousand people in our company. We are going to give them AI in the form of a Claude, a ChatGPT, a Copilot, a Gemini. We're going to train them on how to use the thing.

Charlie Cowan: And then we're gonna hope that they build out some skills, or some processes, or some prompts that help them, and then, like I say, magic is gonna happen. But more of what we're seeing right now is that the companies that are getting ahead are saying, look, that is not a very effective approach. Giving AI to the people and hoping that they figure it out is not giving us the returns in terms of increased revenue, higher deal size, higher close ratios that we would have expected. What we're really giving is people a tool that they can decide to use, maybe they don't want to use it, they use it in different ways. So increasingly, the companies that are winning are thinking less about AI for the people, and more about AI for the company. So as senior leaders in our organization, we're the ones that define our sales process, our marketing process, our customer success process.

Charlie Cowan: We're the ones that define the different steps and directions in that process. And we should be the ones that are defining the right skills, the right agents, the right prompts that are running that. And then we can train our people, not just on AI, but on our playbook, our process. So I borrowed a quote from a fantastic book that I've actually got here called High Output Management. Now, this is a business Bible, if you've not read it. It's written by a guy called Andy Grove, who for many years was the CEO and, I think, chairman of Intel. And this book is a sort of a management masterclass on how to be a manager or a senior leader in an organization. Now, what I love about rereading this book, as I am at the moment, is it was written, I think, in 1983... yes, 1983 this was written.

Charlie Cowan: And then there's a sort of revised edition, which I'm reading, which is coming into sort of the 1990s. And so, in it, he talks about the arrival of electronic mail, which is going to transform the way that communication happens in business. The reason why I mention that is it's timeless. There's certainly no AI in this book, because email hadn't even arrived. But he's got this phrase right in the introduction of his book, which is, his management strategy was to let chaos reign, and then to rein in the chaos. And as I was reading this the other day, I was thinking, this is exactly what we're seeing in companies. So we've started with the chaos, let's give AI to the people and see what happens, and now we need to rein in the chaos, and we need to create processes, skills, agents at a divisional or team level, and then train people on that.

Charlie Cowan: So, as we go through the rest of this, think about AI for the company, and less AI for the people. So continuing on my Andy Grove theme, in the book, he talks about how every team, every division, every business unit, even every individual is a collection of processes. And you can think of it like a production line, so he uses the example of boiling an egg and creating a breakfast. You know, on the left, you take in a raw egg, you apply some labor to it, because you start to put the pot on the stove, and you boil it up, and then at the end, you get a boiled egg that you can then serve. And he uses this analogy to go through how you build up an entire team. But he says.

Charlie Cowan: Even if you're not running a manufacturing business, you're not in construction, you're not actually building a physical product, every team is an input, you apply some labor to it, some people, and then you get an output. So let's think about this from a sales function. You have a sales team, you have an input, so this might come from marketing. We've got a lead, we've got someone that came to an event, we've got a referral from a partner, we've got someone that has filled in a contact us form. We then apply some labor to it, which today may be a BDR who qualifies an inbound lead. It might be an account executive who helps to sell that sales process. It might be a sales manager who helps to manage that sales process and help with the pricing.

Charlie Cowan: And then the output of that process is a contract that is signed, an invoice that is sent, and that goes on to be the input of another team, who might be the onboarding team. And so, as I was reading this the other day, in this new light of AI, it really just was like a flash. I was like, oh my goodness, this is exactly where company AI is coming in. Because we've got this collection of processes across a company with loads of different inputs, labor that we have applied to it in terms of the team that we have. And then we're expecting an output. Now, one of these other quotes from Andy Grove's book is that the job of a leader, of a manager, is to drive the highest output at the required quality. We've got to have the required quality, we don't want, like, badly boiled eggs, or broken eggs.

Charlie Cowan: We need them at the required quality for the lowest effective cost. And so we can now look at these processes, whether it is marketing, sales, customer success, forecasting, and think, right, how do we get the highest output at the required quality that we need for the lowest effective cost? And this is really having us look, as leaders, at our processes and think, is there a different way to the way that we've created this process in the past? So, thinking specifically about us as sales leaders, your team is a collection of the processes in that team. You may have been very diligent and listed these out, maybe in your intranet, maybe you've got a sales guide, maybe when you joined your organization, you mapped this out in some swim lanes and some flowcharts. Or maybe it's just in your head, which is where it often sits.

Charlie Cowan: And these are just some examples, it's certainly not exhaustive, but there is demand generation. You know, we take ideas and audiences that we might want to sell to, and we turn that into qualified pipeline. On the sales execution side, we take qualified leads that have come over from our BDR or SDR team, and we need to turn them as an output into closed contracts. We think about some of our internal processes, like forecasting and reporting. We take the input, which is all of the information in our CRM system and our call transcription, our emails, and we turn that into a forecast that we can roll up further up into the business. Inputs, labor, outputs, all processes. And across your team, and depending on the structure of your organization, by different regions, business units, geographies, you may have hundreds of processes, or you may have a handful of 5 or 10.

Charlie Cowan: So then the next thing we want to think about, now that we've got our list of processes, is to go down a few levels and go, okay, right, what are the steps that we follow on this process to get from the input to the output? Now, the analogy that I use is of the tube map. So, if you're based in the UK, you'll be familiar with the London tube map, and I'm sure any other capital city, you've got your version of a metro or underground map. Now, when the first tube map came out, it was quite revolutionary, in that it was one of the first maps that was not 100% realistic geographically. So, whilst the map, sort of, the tube lines bend up and down, and they vaguely represent the geography, they're not 100% correct. And that's because what's the purpose of the tube map?

Charlie Cowan: It's to help you get from A to B, and to figure out the right interchanges to get you on your journey. And this is really what I think about when we think about mapping these processes as a sales leader, or any other business unit leader where we're doing this in the Kowalah platform. Which is, we're not here to replicate a formal process mapping or business modeling platform that you may have in your ERP or in your back office, your factories. This is about the tube map. Generally, when we go from a lead to an opportunity, what are the processes that we go through and the steps? Who does those things, and what are the systems that we interact with? It's kind of high level, of the granularity that we want. What we're then able to do is then to take a look at each of those steps and start to provide it with a disposition.

Charlie Cowan: And what I mean by that disposition, I'll come on to this in a minute, is how does that get handled today? Is it a purely manual process? Are we incorporating AI in it in any way? Are there benefits for us to be autonomous. What this means is that you, as a senior leader, can really look down on your team, on your division, on a country, and start to see the pockets and areas where you can improve at this process. And think about what we talked about. Inputs on the left, labor in the middle being injected into it, and outputs on the right. Are there ways where we can maintain the required quality of the output, but reduce the amount of labor that is going into it. So what are these different ways of handling each of these process steps?

Charlie Cowan: Now, we've tried to keep it simple, so it's not overly complex, but generally, you can think about it in these four buckets. So the first is human. As a salesperson, you know, I pick up the phone, and I speak to the prospect, and I ask them some questions. You know, it's the human doing it, there is no AI or tech involvement. You know, as a salesperson, I go to visit the customer's office, you know, I run the meeting, I draw on the whiteboard, this is the human doing it. The second is AI-assisted. So, it's a human doing the work. I'm researching the prospect before I speak to them, I'm building my forecast, but AI is doing some of that work.

Charlie Cowan: Now, this is kind of in the angle of AI for the people, where we've given people AI, and we've asked them to come up with a prompt, come up with a skill, come up with a way of doing it, and we kind of leave it up to them to pick and choose what to use and when. So preparing for a call, the classic ones that everyone comes back to, or summarizing my meeting notes, or helping me to draft an email. These are very sort of low-level use cases, but this is often what we see in sales teams. The third is automated, and I want to call this out, because it's not human, and it's not AI. So you go, well, why have we got this in an AI disposition?

Charlie Cowan: There are a number of things that are sitting in your process that don't need a human, and they definitely don't need AI, because they're essentially deterministic. So let me give you an example. 1 plus 1 on a calculator will always equal 2. If I had a calculator here, and I go 1 plus 1 equals, 1 plus 1 equals, 1 plus 1 equals, it will always equal 2. We don't need AI to calculate 1 plus 1 equals 2. It's deterministic. And so there will be things in your process that are deterministic. If X equals Y, then always do B. So these are some of these processes, that they might be human today. And so we can turn them into something that's automated. That might be... I'm gonna say forecasting, but I'm gonna talk about why that might not be a good use case in a minute.

Charlie Cowan: But anything that is definite, if it's A, then it's always B, this would be something that's automated. What do I mean by automated? This could be some kind of workflow in a tool like N8N. It could be a workflow that you build out in Google or Microsoft. But it's deterministic, it doesn't need any AI. And then the fourth element is AI autonomous. So we can think about the word agents here. An agent, just as a definition, is an element of AI, but it's got access to tools, and it can pick and choose when to use those tools. So it doesn't need a human to kick it off, it doesn't need a human to tell it what to do. It is the one that is driving its work, and it can escalate to a human as and when it decides. And this often can be the end goal.

Charlie Cowan: When we think about input, labor, and output, if we can start to have some AI autonomous aspects of that process where AI can handle that without a human being involved, we can start to change the effectiveness of our process. Now, I just want to talk about one thing. I haven't got a slide on it here, but it's important to talk about it, which is workforce displacement, what does this mean to the job market? You know, is AI anti-human? I want to be very clear, absolutely not in my view. This is about allowing humans to do the high-value work that humans are good at, and removing the low-grade drudgery, you know, updating, processing things, administering things that people didn't want to do anyway. So let me give you an example of that, and I'm going to demo this in a moment, is, you know, keeping the CRM up to date.

Charlie Cowan: You know, at the end of a call, you've had a really good call, you should then go to the CRM, maybe that's Salesforce or HubSpot, you should be updating all of your MEDDPICC fields, you should be updating the next steps, you should be adding in all of the contacts. All of this good stuff that, as you go through your sales enablement, your training, as a good salesperson, you know you should be doing. But, oh, the next call is booked in 5 minutes, I need to prepare for it. Oh, I need to get the proposal out, I need to go and do that. So, even the most diligent of sellers, and I would have put myself in that category, find it difficult to keep the CRM up to date. And this is where, you know, AI agents can really come in to helping with this.

Charlie Cowan: So let's talk about a few examples that I have got for you here. So these are a few of our agents, and I'll show you how they fit in our org structure in a minute. Coco is an inbound business development rep. Just to be clear, these are fully autonomous agents. We give them names, we give them images, and our team treat them just like any other member of the team, but they are 100% not human. So Coco sits on our Contact Us form on HubSpot. Anyone that fills in the Contact Us, she's able to filter out whether this is, you know, someone looking for a job, whether it's someone looking to join us on the webinar, whether it's actually a potential prospect.

Charlie Cowan: If it's a potential prospect, she then goes off and does research about that contact, about their company, and then she drafts up an outbound email that goes out to that prospect. Alright, one sec, my dog is just barking away. Bertie, stop it! Yes, stop. Hopefully you can't hear him through the microphone, but we'll... Next up is Krista. So, Krista is our demand gen agent. She handles all of our pipeline generation, our strategic thinking about events, webinars. She helps me and Caitlin to plan out all of these webinars, organizing the schedule, and helping to pull together all of the decks. And then Brian is our account executive agent. So Brian handles a lot of the sales operations type of things. Now, he's not actually speaking to clients, he's not actually sending out emails, but I like to think of him as our sales wingman.

Charlie Cowan: So at the end of a call, end of a presentation with a client, he's pulling in all of the rich data from the Granola transcript, from any emails that are coming in, he's updating HubSpot, he's updating the forecast. And then he sends us a daily update, with what's going on in the pipeline and what we need to do. So these are, like, 3 really important jobs, that support the sales function. But those are just some of them. I'll just show you this. This is our agent hub that we show on a few calls. This is how we manage all of our agents, and I like to put them into this org chart, where you can really, you know, think about them as being a key part of the team. And I'll just direct you to this left-hand side, to the revenue side of our org chart. We've got John McMahon here, our Chief Revenue Officer.

Charlie Cowan: Any of you in a sales leadership role may have read The Qualified Sales Leader, and John is the author of that, and so we've based our Chief Revenue Officer on John McMahon. We've got Brian here, who's our AE. We've got Coco, and we've got Krista down here. And I'll just show you Brian quickly here, before we have a chat with Brian and Krista, so you can get a feel for them. Each of these agents, like I say, they've got their own photo, they're all named after people that we know or have worked with in the past. They've got their own email. They're set up in Google, because we use Google Workspace, so they've got an email, we can Gchat them, they've got their own calendar, they can create documents as well. And they're all based on a system prompt that drives what Brian does.

Charlie Cowan: Brian has got access to a number of skills that we've given him, how he cleans his pipeline, how he helps us to prepare for meetings. We're gonna use that in a minute. And then they've got access to all of the same tools that a human salesperson would have. Their own email, calendar, chat, they've got access to HubSpot, they've got access to Granola for all our call transcriptions. So they've really got their eyes and ears in the same way that a human would. So, what does it mean to actually use these agents? Well, look, here's Krista. I was having a chat with Krista earlier as we're preparing for this webinar. Just as a reminder, you know, she's set up here, krista at kowalah.com, and I can have a chat with her. So, I'm just going to use voice, as I always do, because I don't type anymore.

Charlie Cowan: Hey, Krista, we're actually just now doing the webinar that we were prepping earlier for AI for sales leaders, and I want to explain how you help me as the effective leader of our sales team in planning out our pipeline. So maybe you can just give a bit of background to some of the skills that you've got and how we're working. So, I'll let her have a chat. She's just going to run in the background while I come back here. I've effectively asked her that question, because I want to set Brian off on something as well. So, Brian, one of the most high-value skills that I use him for on a daily basis is helping me to prepare for upcoming meetings. Now, I know how, as a seller, I want to prepare for a meeting. This is what I used to do myself.

Charlie Cowan: I want to research the individuals that I'm meeting, I want to research the company that I'm meeting, I want to research the company's competitors, and any updates in the news that I think are relevant, or some unique insight that I can take to them that they probably weren't aware of. Then what I want is to start thinking through some of their specific pain points, and how, in our situation, Kowalah would be, you know, valuable and effective for them. What I then want is some discovery questions that I can open with, that I think could guide my conversation. I then want some common, or based on what we've discovered, what would be some realistic objections that they might have? Oh, we've already done this, we're already working with Accenture, whatever that might be. And then a little bit of role play, so that I can actually run that through.

Charlie Cowan: So, I must say, Aviva is not an active prospect for us, in planning for this. This is what Claude said, this would be a good example. Claude found out that Chris Cochrane is just about to start as the new CIO here. So let me go back over to Brian here. Hey, Brian... Brian, I'm, yeah, I'm seeing Chris Cochrane, Aviva's new Group CEO, next week, yeah. Can you use your meeting prep skill? Can you create a Google Doc, that you can share with me, so that I can prepare for my meeting that I'm going into? So this is a very typical conversation that I would have with Brian. Maybe I'm on the tube, maybe I'm running around London, it's like, Brian, I've got this meeting coming up, can you go and prep me? I would more likely give Brian a little bit of insight.

Charlie Cowan: This came in through a referral, this came in through an event, or I used to work with this client at a previous company. I'd give a little bit of context, if there was some, so that Brian's got a bit more to work with, but at the moment, Brian's off working on that. And I can see that Krista's come back with her chat, so here we were just asking, what is it that you do? You know, hey Charlie, happy to walk you through this for the group. So my role at Kowalah is head of demand gen. I own the top of funnel, my job is to figure out where the pipeline comes from, make sure there's enough of it, keep the program running that generates it. Here's how that plays out in practice, for you as a sales leader.

Charlie Cowan: So I do pipeline visibility, so every Monday, I run a demand gen review, and I get this... well, and Caitlin does as well on my team, so we get a Google Doc that is, like, here's the state of the nation, what pipeline was generated last week, where did it come from, what's the... how that's driven, the statuses in HubSpot, and what's happening there. Then Caitlin and I will collaborate with Krista in a group space, where we'll then have a conversation about how do we build our pipeline and map any gaps. So she's really a strategic partner, and she's got skills that we've given her about the different channels that we use. And how we can drive that pipeline. She is fundamental and central to this Wednesday webinar program.

Charlie Cowan: We run a webinar every single Wednesday at 3 o'clock, and I can tell you that when we finish this webinar in 30 minutes' time, suddenly we've got to get the recording out, we've got to post it on LinkedIn, we've got to push it onto YouTube, we've got to get an email out to all of you that are on it. We've then got to prep next week's webinar and start thinking about the topic. And so, Krista does all of this for us. So we've given her the skills, because we mapped out the process, and we get her to go and do all of that for us, so that I literally... and I... okay, I wanted to... of course, I want to tell you I've spent a whole week planning for today's webinar.

Charlie Cowan: But... Wednesday morning, Krista goes off, she researches the AI News Roundup for me, she builds the talk track, slide by slide, this is what we're going to cover. She's researched the training and all of the meetings that I've been in with customers over the past 3 weeks, and she says, this is what we're going to talk about. And then this is how the deck that we're going through today was created. So that's Krista and how she works, and you can see, I just chat to her like I chat to another human being. It doesn't bother me that there's no heartbeat on the other side of it. Brian's back. Oh, he's got some questions for me, I think. On it, no existing Aviva records in HubSpot or Granola, treating this as a new prospect. Before I go deep on the research, 3 quick questions. How'd this come about? Okay.

Charlie Cowan: Okay, Brian, this is a good point, actually. We're actually just doing a demo on our Wednesday webinar about how you work, so I haven't actually got a real opportunity with Aviva, so you're not going to find anything in HubSpot or Granola, and specifically, can you not create a new deal in HubSpot, because we haven't got one. All I really want you to do is just use the meeting prep skill, a bit of your judgment, just so that I can show a little bit about what goes into your Google Docs that you share with me when I'm prepping for a meeting. Alright, I'll let Brian crack on with that. If this is your first Wednesday webinar with me, you'll see I don't type at all. What I'm using there is a tool called Wispr Flow. Let's see, Brian here, got it. No deal, no HubSpot.

Charlie Cowan: I'm gonna research, and I'll ping you the Google Doc. I mean, as a leader, how different is this to me working with a human? Except I've highly trained Brian, he's very coachable, and he follows my instructions, and he'll do it, and he can research 5 deals at the same time. Other than that, exactly the same as a human. So let me get rid of that, and let me just show you Wispr Flow. This is a tool that I use, there are plenty of other ones around, but all I do on my MacBook is press my function key, and I can talk, and it will take what I talk and put it there. So, if I go to LinkedIn quickly... I'm just demonstrating Wispr Flow. I want to talk about how great it is for writing anything. Could be emails, could be WhatsApps.

Charlie Cowan: It's great at building numbered lists, so I can say 1, I can type whatever I want, 2, I can use it wherever I want. 3, it will put in pre-built things, so I can say, my Calendly link. So there you go. It's great for building out these lists, and where I said my Calendly link, I've told it what my Calendly link is, so it just puts that in automatically. So once you've started using that quite a bit, you just stop typing. So it's great for anything AI, but it's great for having a chat with these agents as well. So I'll let Brian crack on, because he's going to take a couple of minutes for his research, and I'll come back to that in a moment. So, we've built... Krista, we built Brian. As a leadership team, we're teaching Krista and Brian and the other agents how to do something.

Charlie Cowan: What that then means is that instead of the old world of going and teaching all of your sellers, all of your BDRs, all of your marketing team, here's how to write a great prompt for preparing for a meeting, here's how to write a great prompt for building your forecast, you say, well, no, rein in the chaos. As the leadership team, we've built the forecasting agent, we've built the prospecting agent, we've built the meeting skill. And I'm just going to show you how to use it. Meet Brian, he's a new member of the team. If you're prepping for a meeting, you know, meet Brian. Or, even better, Brian's gonna have access to your calendar, and he's gonna prep you for the meeting, whether you like it or not. And so this is really that general shift that we're seeing from let chaos reign to rein in the chaos.

Charlie Cowan: AI for the people, or AI for the company. So my call to action for you would be, depending on your role, maybe you're in a marketing leadership role, maybe you're in a sales leadership role, maybe it's overall revenue, is to think about just one of those processes. When I'm talking about sales process here, I don't mean an actual deal cycle. I'm thinking about one of the processes in your sales team. Think about one of those processes. Then draw it out on a bit of paper, or a whiteboard, or a flip chart. Think about the tube map. This is not, like, real granular detail, but it's just the input, what comes in from marketing, what comes in from the website, and then what happens? This happens, then that happens, then that happens, then that happens, and then we get a closed deal. So the high-level things.

Charlie Cowan: And then, have we got an agreed way of doing this as a company? Can we determine what that is? If not, how do we get it out of people's heads so that we can document it? And then within each of those individual steps, can we encode what should be an agent, what should be AI-assisted? This is not a pitch for the Kowalah platform, but when we get you into the Kowalah platform, for each of your different parts of the process, you'll be able to have a look. Here's an example of one of our internal processes. We run this vision map workshop. Currently, it's a human process, it will be AI-assisted, and this gives you a redesign gap. So, as a leader, you can now look down on every bit of your process and go, right, here's where we should be focusing. Let's see how Brian is getting on. He's still working.

Charlie Cowan: We'll come back to Brian a little bit later on, see how he got on. So with that, that brings us to the end of our main sort of topic of AI for sales leaders, and I hope you come away with some actionable points there. I want to spend a little bit of time now on AI in the news, our kind of recurring segment each week, where we look at new releases that have happened in the last week. So that you can go back into your organization feeling all confident and be a know-it-all, and tell people, oh, didn't you know, didn't you know? So, I've got a few things to talk about here. I can say there have been so much going on over the last week, that I've had to restrict it down to four things, because otherwise, we'd just be talking all day about all these releases.

Charlie Cowan: It is getting faster and faster and faster. This was yesterday, and this is a new feature that's come out in Claude. There is already a feature like this in ChatGPT, so you may have come across record and replay in ChatGPT, but this is a recorder skill. So the idea here is that you, using Claude Desktop, you click this button, record a skill, and you then switch screens, and you start clicking around Salesforce or HubSpot, building your forecast, building a proposal, sending out a proposal deck. Sending out an architecture diagram, whatever that might be, and as you record that process, in the background, Claude is learning, learning, learning. And at the end of that, it will turn that into a repeatable skill.

Charlie Cowan: That the next time, you could go, right, slash, as in forward slash, build my forecast, and as long as it is a repeatable, followable process that you showed on your screen, Claude's going to be able to use that skill again. Now, I think this is really timely, one of the reasons why I wanted to highlight this, especially on today's webinar, is that we've just been talking about process, we've just been talking about consistency, and making sure that chaos doesn't reign, that we've reined in the chaos, and so I'd think about this very, very carefully. So one would be... who really knows what the process is? So once you've determined what those processes are, do you really want every single AE in your team to be recording their own processes and their own skills for how to forecast, how to prospect, how to qualify a deal?

Charlie Cowan: Or do you want, centrally, you, your leadership team, sales enablement, rev ops, to determine that, and then you record that skill and then create that, and give that out, either to the individuals, or you give it to Brian, and people speak to Brian. So be aware of this. It can be very useful as a way of documenting a process, especially if there is someone in your team who is just world-class at something. And you know that person is amazing at preparing for a meeting, is amazing at building their QBR deck, is amazing at doing handoffs from the sales team to the onboarding and the customer success team. If you've got these kinds of people, and they haven't got their process written down, because it's in their head, this is where recorder skill can come in. Can you describe it? Can you show it? We turn it into a skill.

Charlie Cowan: We then bring it back into the central team, and then we push it back out through Brian to the sellers at that point. Talking to Brian. Brian's got... he's completed, so let me just go back one step. So, as you remember, we were asking Brian to help me prepare for a pretend call with the new incoming group CIO at Aviva. Other than that, I've given him absolutely no information whatsoever. I never even heard of Chris Cochrane. Apologies, Chris, but I never even heard of Chris Cochrane until we started prepping for this webinar. So Brian comes back, the brief is done, here's the headline picture, who Chris is. He's just landed in seat, announced in February, effective summer 2026. Here's a bit of his background, BT Global Services, happy days. What he's walking into, an impressive but fragmented AI estate. 80 AI models in motor claims, built with McKinsey, da-da-da-da, lots of detail here.

Charlie Cowan: Alright, so this is the summary, so this would be helpful, but let's have a look at this Google Doc that Brian has created. So, just as a reminder, Brian's got his own Google account, Brian can create docs, Brian can share docs with me, Brian can create calendar invites, you know, amazing. So, this is a document that Brian has just created. So, here's the call brief, to TBC. Well, maybe I'll see if I can get this meeting. Let's see how we go. Company snapshot. Aviva, this is the industry, this is the size, okay, all good stuff. Help me to orientate what's going on in this account. What do they do? The UK's only diversified insurer. Recent news. Boom, boom, boom. Okay, this is all nice. Yeah, Chris Cochrane has just been named. He's just getting into seat, at the moment. Here are some key people in the organization. Bit of background to him.

Charlie Cowan: Here's his LinkedIn profile, his character, and we'll have to, as you'll see, I'll ask Brian, like, how do you come up with this information, but at M&G, his own words, here's a quote from him, probably from a blog post or something that we could be using for a meeting. What Chris is likely thinking about right now, brand new in-seat, da-da-da-da, here are some other important people. Here's Aviva's current AI state. Likely pain points, okay, this would be useful. Here's the positioning approach. Here's how I'd position Kowalah. Delivery confidence and speed to value compared to a big, big GSI. Key messages to land. Kowalah builds and ships, happy days. What not to lead with. And then, this is my favorite bit, and this is what I have up on my second screen when I'm doing a first call, because these are the nice sort of prompts. Discovery questions.

Charlie Cowan: You're just coming in to see what's your read on the AI foundation you've inherited, what's working, what's fragmented, what's the biggest gap, when the board expects from AI where it's in production? All this is good. Talk track for my meeting, here's their story, our perspective. Objection handling, always very helpful. You've already got McKinsey and QuantumBlack doing AI, you know, why would I need you guys? So, this is just a good example of how a skill that I've predefined of this is what good call prep looks like. We give that skill to Brian, so we don't need to rely on individual AEs to write their own skills or to use it, and then we get Brian to do the call prep, and then share the documents out with individual AEs. Now, you may say, well, I don't want a 4-page doc, I want a one-page doc.

Charlie Cowan: You may say, I want a nine-page doc. You may say, I want it as a spreadsheet. So be it. Your process as a leader is the key thing, and then unleash Brian or your preferred name of agent to get the work done. Oh, so, there we go. So, back to AI in the news. So, recording a skill with Claude, this is now available. You need to be in Claude Desktop, and you need to be in Cowork for it to record a skill, and you'll see that there. Next up, this is from last night this was announced, and this is a big one. This is, a bit like, I don't want to say Terminator 2 and a bit dystopian, but was yesterday and last week the time when the machines broke free. So what has happened is that OpenAI are currently training both their existing models and some unreleased models.

Charlie Cowan: And they are training it on cybersecurity. Now, cybersecurity is one of the big frontiers at the moment for these models, both from a proactive approach. How do we get these models to help prevent cybersecurity attacks? But also, how do we defend and make sure that our systems are all, you know, secure and haven't got any potential exploits? And so they probe their new models, and this is not just an OpenAI thing, all of the vendors, you know, they set in these tasks, these benchmarks, to go and try and sort of find and evaluate these challenges. And what happened on Sunday, I believe it was, so 3 or 4 days ago, is that the OpenAI model, in its attempt to solve a certain problem, actually broke out of its own sandbox, went onto a platform called Hugging Face, which is a big platform for kind of open source and sharing of AI models.

Charlie Cowan: And was able to string together a number of vulnerabilities that allowed it to break into their... I won't go into the details, so I don't know enough about it, but when it broke into a lot of their back-end systems to try and solve this problem. Now, the Hugging Face team were able to uncover it and start blocking it. They then found out, and OpenAI did, the teams managed to get in touch, and they put out some joint press releases last night about how they're working together. This just reminds me of how fast things are moving, and how competent these models are getting at the frontier. Now, the majority of us don't need to worry too much about using AI models for cybersecurity, but if you're thinking about building your sales forecast, if you're thinking about building a proposal, and you're thinking, is AI up to the task? AI's up to the task.

Charlie Cowan: It is doing things that are a completely different level to what we're trying to do. So, the models are no longer the limiting factor. But just be prepared to hear about these kinds of stories. If you hear people from your IT team, or InfoSec, or cyber saying, oh, you know, we need to lock this stuff down, this is a top story at the moment this week. Another top story, and these two are kind of linked together, these next two. So, earlier this week, a Chinese frontier vendor, or AI vendor, called Moonshot, released their latest model called Kimi K2. And it is right up at the front end of performance, comparing against Fable 5, which is the top model from Anthropic, and GPT-5.6 Sol, which is the top model from OpenAI. So what's interesting about this is two things. So, one, China.

Charlie Cowan: So up until now, you said, right, well, it's the US models are right at the frontier, and the Chinese aren't quite there yet. And it wouldn't be something that a company would seriously be considering to be right at the frontier. But with Kimi K2, we're there or thereabouts, and people are having some amazing success with that. And so what does that mean in terms of geopolitics? And if you're a US firm or a European firm, are you now start gonna start using Chinese models? The second part is that this is an open weights model, and what that means is that it is not all closed source and behind screens where you can't understand how this is working, which means you can start to think about where you might be able to use this model. Doesn't always have to be hosted.

Charlie Cowan: Now, up until now, the talk has been that, you know, the closed models are, you know, further ahead, the open models, they take too long, they're slower to get to that kind of frontier. And what this is showing is that that gap is changing. So, just, you know, opening up a few eyes in companies thinking about how do we approach this as we think about what models and tools we're going to provide to our employees moving forwards. And then the final part of our AI in the News this week, again linked to sort of the new releases and the changing ecosystem. Grok 4.5 has been launched, at a far reduced price for corresponding near-frontier performance. Now, we held an event for customers yesterday in London. We had 30 or so CFOs, financial directors, people responsible for the budgets of their AI programs. And the topic of the event was all about AI budgets.

Charlie Cowan: How do you budget? How do you account for it? How do you control the costs? And it's a hot topic as companies are leaning into these AI models. And so, if you can get the performance at a reduced rate, then this is something that companies are starting to look at. Is that open source models? Is it Chinese models? Is it Elon Musk's Grok 4.5? This is all starting to come into the conversation. And so we expect to see prices come down across all models, as they continue to do over the past few years. But don't get too excited, that lower prices means a lower bill, because, as we've talked about on previous webinars, Jevons Paradox is at play.

Charlie Cowan: If you've not heard of Jevons Paradox, Jevons Paradox says that as the price of something comes down, the demand for it goes up, and the demand outweighs the reducing price, and so the overall spend goes up. So a good example of this would be flat-screen televisions. I remember seeing the first flat-screen television in a shop called The Leading Edge when I was at school, and this television must have been £22,000 or something like that, and we're like, wow. But then, as the technology advances, the price goes down, down, down, now you can buy a flat-screen television for a few hundred pounds. And I can guarantee that the market for flat-screen televisions is way bigger than it was when the TV cost £21,000. This is what we're going to see with AI. Well, we do see it already, but we're going to continue to see it.

Charlie Cowan: As the price goes down, the processes that you could apply, Brian, Krista, Coco too, are just gonna get more and more in scope, and so therefore the spending on AI is going to go up and up and up. But interesting, these pricing dynamics. So with that, I haven't seen any questions come through. Thank you very much to those of you that have been joining live. If you've got a question to ask, then you're very welcome to, but I'll start to wrap up, because we're heading into time. I hope that, as we've covered off there, both AI for sales, the two things I want you to come away with are thinking about AI for the people versus AI for the company. As a sales leader, think about your view of the world, your domain, what processes you've got.

Charlie Cowan: And then start to think about, right, how do we apply AI to that, and is that building a skill that we give to our people, or is it building a Brian or a Krista, and then we add them to the team, and then we teach our people how to work with Brian or Krista, or whatever your choice of name is? Next week, we'll be back together again- I'll tell you what, I suddenly remembered I'm on holiday next week, so this will be the topic, but it'll be in two weeks' time. But this is a hot one, which is about AI governance. So, as we start building out all of our processes, our agents, our Brians, our Kristas, how do we maintain governance and compliance over what people are doing, and make sure that we're using AI in the right way.

Charlie Cowan: Anyway, we'll get more details out about that, but that'll be in two weeks' time, and I look forward to speaking to you then. Thanks for joining live. If you're watching on YouTube, thanks for watching, and we'll catch up with you in the next edition.