The HubSpot ChatGPT deep research connector: What you need to know

Read the full transcript here:

So there was a significant announcement yesterday by HubSpot. Announcing an integration between HubSpot and ChatGPT. 

Do you remember the name of it? The official name...

The ChatGPT Deep Research Connector.

Okay. Good.

00:17
My name is Craig Taylor and you're listening to Truth, Lies and B2B Tech, a podcast for anyone engaged in complex B2B sales. Along with my co-host, James Ingham, we discuss the latest best practices in marketing, sales and service. In each episode, we take a go-to-market strategy, look at it through a technology lens and figure out if it's worth your time. Let's cut through the bullshit—no fluff, self-serving agendas or veiled pitches. Just honest insights from people who work in the trenches.

As always with HubSpot, this integration has been lauded and celebrated. In this particular instance, I think it's something that people really do need to take seriously because it is pretty game-changing in terms of the potential for the technology. However, it also comes with quite a number of considerations and concerns.

01:13  
Today we want to talk about what it is, discuss some of the use cases, and explore some of the major considerations around security—particularly from a data perspective. We also want to talk about data readiness, which is such a big thing. We want to give listeners a bit of an idea of what steps you can take to move forward on that. We see it as fairly problematic in a lot of the organisations that we work with. They tend to be medium-sized businesses aspiring to be really effective from a go-to-market perspective, and data is one of the bottlenecks holding them back.

01:56  
That problem exists from a go-to-market perspective anyway, but AI—particularly with the advancement of AI we've seen from HubSpot, and that's been a pretty rapid series of introductions around AI—is just making that problem even worse.

02:13  
Data has always been a bottleneck for many organisations. Because the technology wasn't at the point where it is now and where it's heading, it wasn't a major issue—you couldn't really do sophisticated things. But now we're getting to the point where AI is making you really think about the art of the possible and getting people very excited about what can be done from a go-to-market perspective. But again, if the data isn't there, you're not going to be able to do anything meaningful.

02:49  
In practical terms, what it does is enable you to use ChatGPT, but informed by your HubSpot CRM. HubSpot are the first CRM provider to allow that to happen. I think partly because others may have concerns around data privacy, and partly because HubSpot have really, from the start, been on the ball when it comes to AI and are really driving that agenda.

If you think about how that might improve how ChatGPT can be useful as a tool: until now we've had great CRMs or customer platforms like HubSpot, and we've had great consumer AI platforms like Claude and ChatGPT. They're both powerful on their own, but together, the art of the possible is much, much greater.

Previously, using the likes of ChatGPT and Claude, people would be able to complete certain tasks based upon the internet—basically based upon the overall knowledge it can gather from its training and user input. But it's not been specific to an individual person's own data. That's where this is really different and why it makes it so much more game-changing. It can make the output of AI much more relevant to your own individual circumstances and therefore much more useful.

04:07  
HubSpot's developments around AI have been embedding AI into its own platform. It's been developing agents such as the prospecting agent and customer agent to perform certain tasks and to support certain functions. So this takes it to another level where HubSpot is now integrated fully with what is arguably the leading generative AI platform: ChatGPT.

04:38  
That's also a downside because ultimately HubSpot's efforts to integrate AI into their platform were really useful—they enable you to use AI in your normal workflow in a way that you're familiar with, without having to go to another platform. The AI that has been brought into HubSpot as a platform is powerful and useful, but it's ring-fenced to a degree. You can't exactly use it in the myriad ways that you can use ChatGPT, which is the major benefit of building the connector. 

But the downside is that it isn't integrated into your normal workflow. For people listening to this, just imagine basically writing a prompt and asking it to query your CRM data.

05:29  
Even with the Connector, you're not operating the activity or making the prompts within HubSpot directly—you're doing it in ChatGPT and it's pulling in the information from HubSpot.

05:39  
Yes, which is a massive game-changer in terms of the usefulness of ChatGPT. But what it can't do, like the tools in HubSpot can, is perform actions based upon your prompts. With the HubSpot copilot, you can ask it to build workflows, query your existing reports, write emails and build those emails. ChatGPT can do a lot of that generation of content and querying data, but it can't then take the action afterwards.

If we're being critical, that's a downside. But we've already got our hands on it and been able to play around with it, and some of the use cases are really powerful.

06:18  
Pretty much you can take any data that sits within CRM, within HubSpot CRM, and you can use ChatGPT to basically run queries on it. It's not just the company and contact records—it's deal information, user information, pretty much any bit of data that sits within HubSpot you can manipulate through ChatGPT and turn that into insights.

06:43  
To a degree. The way that it works is through an API, so anything effectively available through the APIs is available to ChatGPT to query. There are some limitations, and we haven't read all the finer detail yet—it was only announced yesterday—around what personal data is shared. We'll do some research on that. I know there's probably another topic for the podcast around the data privacy element of using AI more broadly, including with HubSpot.

But yeah, in theory, any of the data that sits within your CRM is readable by ChatGPT and therefore queryable. It sounds a little bit basic when we say it like that, but if we talk through some of the use cases and how it can practically be applied, I think some of that usefulness will become clear. AI is at this stage so developed that it's only limited by your imagination to a degree—and of course the data that is fed into it, including both the data and the prompt. Some of those use cases are really powerful in terms of what it can achieve.

07:46  
For some time, we've been working on our own version of AI to be used to help score leads and help support the prospecting agents to be more effective by targeting and prioritising accounts. ChatGPT takes that to a whole other level. The first experiment was really trying to get the tool to identify some really standout accounts that are probably a better fit for what we do.

08:19  
Once the connection's in place, I could ask ChatGPT for the top five accounts in our CRM that are most likely to be responsive to a request for a call. It sounds basic on the surface, but when you think about what ChatGPT can do to help surface that information, that's where it becomes really clever.

In the CRM already, we can interrogate that data by maybe running searches around demographic data—the number of employees, industry, etc.—to determine whether a prospect or a company is an ideal fit for us. But where ChatGPT can take it to another level is looking at all the other things about an organisation that might make them a good fit, such as how they position themselves as an organisation on their own website and whether that would align with our values as a business, whether they've received recent investments, whether they've engaged with any of our previous activities, whether they have responded to outbound phone calls made to them, and the news around recent hires or any particularly big wins.

It can aggregate all of that data that within the CRM alone would be either really time-consuming or really difficult to do accurately. It can aggregate all of that and analyse it amongst every company within the CRM. Within maybe five minutes—it takes a little bit longer than a normal prompt might—it can return five accounts or companies which, with all of that data fed in, look like they're the top of the pile in terms of the people that we should be reaching out to.

10:06  
It's a mix of data that is pulled in from HubSpot, but also what it can find on the internet.

10:11  
You can certainly configure it in that way. The point is, it's looking at a much more granular, much more nuanced view of why someone might be a good fit prospect for us, rather than just top-level data, demographic data, response rates, etc. It's looking at the much bigger picture. We haven't reached out to those accounts yet or companies to see if they would be more likely, but we will obviously be experimenting with that in time.

10:36  
You sent me the output and it threw up five companies that probably aren't really on our radar. The rationale was still quite compelling for us to focus on them. It certainly ticks those boxes. I think there's always that element of doubt around accuracy, which I'm sure we're going to come to shortly. I think that's not just in your own data set, but how accurate is the data that ChatGPT is pulling in from the internet? Because we're still not completely sure how that's actually working in the OpenAI world. Obviously, if you're going to be relying on that data to start to prioritise your outreach, then you want to be pretty sure that it's accurate.

11:22  
There is one disclaimer that I should have made at the start, which is that in the UK and the EU specifically, people need to have a ChatGPT plan that's greater than plus, which are quite expensive. In the US and I think other countries, you can have the plus plan, which is about 20 quid a month, and access this tool. And loads of others, by the way. It's not just HubSpot that you can create a connector with ChatGPT. You can connect it to your Google Drive. You can connect it to GitHub and various others. And I haven't tried this yet, but I'm assuming you can combine those sources of data from all those tools to get even richer analysis.

11:59  
So before we move on to data, other use cases. The big one is how it can support SDRs. So they're planning to make an outreach to an account and they can get all of that information on that account in terms of demographic data. Like you say, new hires could be investments. That's all now at their fingertips potentially. That is pretty game-changing because we know that in most organisations, in most mid-market organisations, SDRs aren't doing as much prospecting as they could be because they're held back by the amount of research they need to do at an account level.

12:36  
Or they're banging their phones and doing loads without the necessary research to make it compelling in the first place.

12:41  
The big problem for them is that they, unless they've got a really good CRM with lots of integration and they're pulling in data sources and signals and intent data from reliable platforms and it's all pushed into their sort of workflow that they're using to actually make the calls, make the emails, that is going to really slow them down if they have to research and look at different platforms like LinkedIn Sales Navigator as well as HubSpot as well as 6sense or whatever it is that they're using. So having it all in one place now and a quick prompt with ChatGPT again is just going to really speed up the process.

13:18  
Yeah, and if we look a little bit outside of the sales, I think sales is one area where it's going to be most impactful. But I think if we look beyond that into other stages of the customer journey, we could look at, say for example, the clients that maybe are most likely to churn. If a customer is using HubSpot for the customer success workspace, they're tracking all the deals going through. Ideally, they're tracking things like if it's like a SaaS provider or something like that, usage and login data and can get that into HubSpot. Again, and this is really what it all boils down to, being able to aggregate large amounts of data from multiple sources and surfacing true insight from it is what it all boils down to.

But if they can aggregate all that data on customers and they can inform the model on what makes someone most likely to churn, then again, they can surface that data and maybe reach out to them and give them a bit of TLC so they don't. Ultimately, I think it does boil down to that. And that's why the data point that we made earlier is so important. Because ultimately, if we're aggregating data and crystallising it into insight, the data that it receives is the primary point that is either going to be successful or fail.

14:39  
The art of the possible now is limitless. Things that we used to dream of doing, we can now do, but it all relies on the quality of the data. I mean, arguably you can move from basic reporting and dashboards to just say, you wanted to try and understand attribution, for example, you just create prompts that's going to ask for attribution. So how many of these marketing leads in the last six months have actually turned into customers?

15:01  
Well, that's another point actually, because I think it does rely on two things—it relies on the data, but it also relies on your ability to produce a good prompt. What we can't expect ChatGPT or any other AI tool to do is to kind of magic up what you're looking for in terms of a particular thing.

15:20  
And you can't afford hallucinations from ChatGPT if you're going to the CFO.

15:24  
If you went to ChatGPT, plugged into HubSpot and said, okay, well, what were my most successful marketing activities this year? In my experiment earlier, it said, well, define successful. And that's kind of the point. I think you need to spend enough time feeding it the information that it needs in order for the output to be successful or valuable. There's another example there, which would be, you know, show me which 10 customers are the best fit for me. Well, it needs to know which you consider to be a good fit for yourself.

So it's not going to be able to determine what industry is best suited to you or what particular pain points are best suited to your own kind of, well, if you know, in time, it probably could have been the right prompt to maybe it could too. But the input basically in the prompt is just as important as the data that goes in.

16:17  
It's that shift again, isn't it, towards prompt engineering. Previously, you're trying to create a report that's quite customised, you'd use the custom report builder. But that requires a lot of skill and understanding and expertise to do that properly. And again, it relies on data. You're still going to rely on data, but now it's a very different skill set to actually get the output that you want.

16:37  
Before we kind of transition into the conversation about data, there is another use case actually that's outside of the aggregation of data point that I made earlier. And that's around data itself. We're going to come onto data readiness in a second, but lots of people are asking questions around, well, I know my data needs to be in fit shape for me to make the most of AI, but how do I go about doing that? And again, AI can help with that process because it can read your CRM. AI, or ChatGPT plus HubSpot can help with that.

Because it can read your CRM, it can see where those gaps are. So I asked it to do an audit of our CRM earlier where it picked out some key gaps in the data that we have, which are going to limit the effectiveness of any AI usage that we embark upon. And it came back with an audit that two years ago we might have done ourselves—and we've done many of them—that might take two days.

17:31  
Yeah. So it brings in the next dimension, which is security and how that data is going to be governed. So we've got a lot of customers that have previously voiced concerns around AI within HubSpot. Particularly if they're in certain industries or maybe they've got real focus on ISO or GDPR, they're very stringent on how their data is governed, then this is a whole new level really. It's going to certainly potentially scare the shit out of CTOs, CISOs that are really trying to control the use of AI in a sensible way that doesn't put the company at risk. What's your take on that?

18:11  
Well, it's a balance, isn't it? Because ultimately, we speak to customers from a sales and marketing perspective that feel really frustrated that their CTOs, CISOs are holding them back from using AI or only allowing them to use it in a really kind of ring-fenced manner. The one we get most is I'm only allowed to use Copilot, which obviously is a great tool, but is it connected with your HubSpot CRM? Well, no.

On the flip side, it's completely understandable that an organisation, particularly in a really heavily regulated industry or a security-conscious industry like cybersecurity would want to protect themselves. I think the challenge is you kind of want to find the right balance, like I said, because it's either you open the floodgates and you're going to very quickly end up potentially with a data breach and the problems that come with that.

18:59  
If you've gone all in with HubSpot and marketing sales, service data, there's a hell of a lot of data that you would really be very concerned about.

19:08  
There's a lot of personal data in there and customers, etc. And for some even that have pushed the boundaries of HubSpot with custom objects and stuff that maybe have sensitive data in there as well. And if they're properly implemented, sensitive data should all be encrypted so it shouldn't be an issue. But if they haven't, then there's potentially very sensitive data in there. There's one side of the coin, which is opening the floodgates, allowing people to use AI unhindered, and potentially very quickly ending up in a position where you're in a data breach scenario.

19:40  
I think it's, I mean, security protocols around ChatGPT are pretty secure, but you just don't know where that data's going.

19:48  
Well, no, it is a breach anyway. Some of the concerns around the initial launch of Breeze were that Breeze itself uses OpenAI, whose servers are in the US. Now, obviously with GDPR, EU-based countries don't want their data stored in the US. It has to be stored in the EU. So that would be a compliance breach straight away if it was used. Whether the data was leaked or not, it would be an issue.

But there is a flip side to that problem, which is that, of course, AI is happening whether people like it or not. Ultimately, if it's not well utilised, then your marketing sales teams are going to be left behind. And as a result, so is your business. So finding that balance between enabling the use of AI for competitive advantage versus opening the floodgates and kind of allowing an unhindered use of it is really, really tricky.

And HubSpot, to be fair to them, do a really good job of kind of advising people around how that data is used and allow them to make decisions on the way that they use AI within the platform itself. But of course, we're talking about the ChatGPT HubSpot connector, and that does use OpenAI. So people that have that concern around using data, data centres in the US, are going to need to cross some bridges, I think, to satisfy themselves that that's in compliance.

21:20  
But like you say, for a long time, only really the top performing larger organisations that have got the resources and the budgets to be able to afford a fully fledged marketing operations team and specialists in data. It's a full-time job to make sure your CRM is as optimal as it could be. The vast majority of organisations aren't in that privileged position. And they've probably been able to get by now.

Those organisations that are unable to take it to this next level, the gap between those that are still operating sort of as they were a year or two ago without any real advancements in their data quality, they're going to be so far behind. The gap is going to be significant, not like it used to be where you might just have a slight advantage. Now the advantage is going to be ridiculous. And I think clearly there are going to be major concerns to get over from a security perspective to satisfy the CISOs that this is all working in a way that's safe for the company. But that will happen at some point. The rate of the technology advancements in this area now is only the data that's holding it back, really.

22:33  
And that's why I think, unfortunately, yet again, assuming large companies or, you know, upper middle, upper mid-market companies can get over the compliance challenge and can therefore invest much more heavily in making sure their data is up to date will be at a greater advantage from a sales and marketing perspective, like the old days of when they could afford massive advertising budgets or big ABM programmes, etc.

But there is a flip side to that as well, which is the smaller organisations are going to get over that compliance challenge perhaps more quickly, maybe foolishly, but more quickly and can probably steal a bit of a march as long as they can get the data in order.

23:12  
So let's move on to data quality then, which has been a common thread throughout the conversation so far. So maybe using the example that we gave earlier with the first sort of experiment with Deep Research Connector. What did that throw up? Because clearly, whilst it looked quite impressive, we also felt there's some gaps there.

23:35  
Yeah, and I think you've got to be sceptical generally when using AI full stop. And I think the first thing to reinforce, we mentioned it at the very start of the podcast, you do get this kind of shit in, shit out problem with data. And a good example of that, based upon the example I gave earlier on the prompts we did for our own business around identifying five accounts which will be most likely to be responsive to an outbound call from us.

Of course, it did give me five, but it could have given me the five where it had a good understanding of the data rather than the actual best five within our CRM. So ultimately, if you think about it, ChatGPT only knows what it can read within the CRM. So if there's certain data points that it's looking for around what might make someone relevant and more likely to engage with us, then it's going to choose accounts which have that data there.

There could very well be five other accounts in the CRM that don't have as rich data, but could in all likelihood be a much better fit for us.

24:42  
I suppose there's just the basic eye test isn't there? I mean, we looked at those five accounts and intuitively, anecdotally you're thinking that's not the five I would have picked. So you're thinking on the one hand, that's quite impressive. But also you're thinking, well, I'm a bit concerned that it didn't flag any of the five that we would have probably picked.

25:01  
That's because there's information in your head that isn't in the CRM. So if you imagine a world where you're, and this is an impossible thing, it's Nirvana, but you want to get as close to it as you can. Imagine a world where your company, your contact information in HubSpot is universally as rich as each other. You had as much information, and ideally a lot of it, on company A as you do on company B and the rest of the companies in your CRM.

That means that ChatGPT can perform the kind of analysis which is going to result ultimately in a much better outcome because it's not leaving people out purely because the data isn't there for it to read. So that summarises the problem with data and why it's so important to enrich it. But I think it's also important at this point to tell people how they can do that. We're not going to do like a really detailed step-by-step guide—it's probably for another podcast. But ultimately, the first job is to identify the problems or the gaps in your data, which again, I used ChatGPT for our own business as well.

26:04  
Yeah. And I think all we ever hope from doing this podcast, we're not going to go into huge detail with how-to guides and practical implementation steps. I think what we're trying to get people to think about is the important considerations and try and change the mindset, particularly with data. I mean, we know that probably many organisations don't take it as seriously as they should. So when I say some, that some would be those real top performers that we mentioned earlier that have dedicated marketing and operations teams that live and breathe data.

If you haven't got that dedicated resource focused on ensuring you've got the right quality of data in CRM, it can go out of date so quickly. You can buy the richest, most accurate data set and it starts to erode in value the next day. Whilst marketers will obviously be working more with data every day, much like the SDR team, maybe less so the sales team, if you've got a sort of divided view of the importance, sorry, the importance of data quality, then you're always going to be on the back foot and you're always going to have a CRM that's subpar.

Yeah, if you have no dedicated resource, all the data, even the basic demographic company contact data is going to go out of date. We see gaps in CRMs all the time. But then if you've got a wide organisation that isn't taking every opportunity to add data to CRM.

27:34  
That can be anything, can't it? That can be notes, that can be recordings of conversations I've had on calls. Ultimately, a lot of that can be done by being disciplined around using HubSpot to track all of that activity.

27:49  
Form capture, automation, all that. I think, you know, just basic things like you've got sales team at events speaking to people on the stand or at a conference and we'll just scan their badge.

28:00  
Or even make a note in a notebook and then...

28:02  
You can have a conversation where you're saying, we're in this contract until X, Y, and Z. We've got this technology, all this rich insight that is happening through a conversation is never captured. There's such a lost opportunity. And I think maybe the value has always been, or maybe it hasn't been seen as valuable because there's never really been an easy way of actually taking all that data and turning it into insights that can be acted on. Now there is.

For those really disciplined organisations and teams that value the quality of data, they're really going to have a big advantage over those that don't. And I think that's the first place to start.

28:43  
I think what you're saying there is that, and it's a good point, there needs to be cultural change. There needs to be an understanding of the value of data and therefore protection of it and good collection of it and entering into the CRM, obviously.

28:58  
I think if everyone gets carried away with the new tools, there's so many different individual tools now that the barrier to creating a new piece of software, a new agent is so low. You've got AI tools that can help you create AI tools. Every day you see more and more people on LinkedIn trying to extol the virtue of their impressive stack and all the different tools that they're using. And I think it probably comes on to the other point really, which is what's so good about HubSpot's strategy—they really believe in having one unified data set. Therefore to do that, you need to really rationalise the amount of technology that you've got. I don't believe in a huge tech stack. I think the more you can do in one platform, the easier it is to manage that data.

29:41  
That's always been HubSpot's manifesto, hasn't it? That's why they exist and largely why you can do so much within the HubSpot platform. I think that's really paying dividends now AI has come around because ultimately, if you're running your marketing through HubSpot, you've got your website in HubSpot, sales, service, operations, etc.

30:04  
Revenue operations.

30:07  
Then of course, all of that richness of data can be read in one place. And the interconnections between that data can also be understood and be used to drive insight. When you say a marketing lead goes over to sales and ultimately ends up being a customer, being able to analyse that across the full customer journey and the interactions between those different business functions is really insightful.

HubSpot's in a really, and this may well be the reason that they've gone so hot on AI, but they're in a really good position to help their customers make the most of AI as a tool.

30:46  
Data quality then obviously culturally that's a big problem that's driving poor quality. I think just the sheer pressure on marketers to move fast means quite often they'll just be bringing in new data sets as and when. If they've got external tools like ZoomInfo for example, they want to run a campaign, let's get a shitload of data really quickly. That happens consistently over time without that data being revalidated and just checked.

You just end up with a Frankenstein's mess of a CRM with lots of different lists and data and lots of gaps between it because there's not been a strong, robust set of governance processes around how you import and export data.

31:30  
Yeah. And I think if we think practically, just to finish around how people can make sure that that problem is first of all, remediated and resolved, but then kept in check. Obviously the first step in a, you know, if we're looking at this practically in order to get your data in a fit shape for you to use AI is to understand where the gaps are in your data. So missing demographic information, incomplete contact records, duplicates, etc. And like I said earlier, you can use AI to understand where the gaps are in your data so you can better utilise AI in the future much more quickly than you might be able to manually.

There are obviously other tools that you can use, the data quality dashboard in HubSpot and others as well. But naturally, the first step is to understand what you need to remediate. The second, of course, is to remediate the problems with that data. And that means data enrichment, basically. So once you understand the gaps, you've got to fill them up.

And there's multiple ways you can do that. Of course, with HubSpot, data intelligence is a good way to do that, assuming you've got a plan for credits. You can use, obviously, the traditional data providers, Cognism, ZoomInfo, etc. And you can use other means. Lusha is a good example. Most SDRs will be used to using Lusha, and they can collect data as they go, etc. You've just got to plug the gaps in whatever way.

33:00  
Then once you've got the basics, you can start to supplement that with intent data signals. You can maybe integrate things like data and Built With it. If you want an understanding of the technology platforms that your target market is running, you can get that en masse with the integrations. You can bring in that data. There's tons of information through LinkedIn that you can bring in using a variety of different tools. You don't necessarily have to have the very expensive Sales Navigator licence to integrate with HubSpot. Those are ways around that. And there's all sorts of things that can supplement the very basic firmographic data.

33:40  
But it's important to get the basics done first before you start to look at things like signals. The third thing really is keeping it in that position. So there's two things related to that. First of all, putting controls in place around how data can be entered into the system. Now ideally, you don't allow a kind of a free for all on downloading lists or getting them from events and anyone chucking them in. You want to have a central point. And like you said, in a bigger organisation, that might be the marketing operations team.

They are custodians of that data and treat it very seriously. So putting policies in place that protect the quality of data when it's input into the system. And ideally a standard import sheet that has all the columns that need to be filled before the data goes into the system to ensure richness helps. But there is a lot of automation out there that can help keep data continuously enriched. Some of them are through integrations with things like ZoomInfo, Lusha, Cognism, that can fill gaps in your data as and when they appear automatically. HubSpot can do that too with continuous enrichment.

But ultimately the job there is to ensure that you don't get back to the poor quality data that you had before. And that brings me on to the last and final step that we'll finish with, which is ongoing analysis of data quality. The data quality dashboard in HubSpot is a really good way to do that. You can track constantly, as frequently as possible, ideally, the quality of your data. And when you can see it start to dip, or you can see there start to be a problem introduced, then that needs to be remediated as quickly as possible.

So it's all around, first of all, identifying and fixing the problems, and then putting automation in place to ensure that the CRM stays clean. And I think perhaps most importantly, having a mindset that means that you're continually monitoring the data quality within your system so that any issues can be remediated quickly.

35:38  
I think just to add and to close—that sounds very fucking formal, doesn't it? Close. Well, we've touched on it, which is why you probably didn't include it in the summary, but I think the cultural element is probably the next stage, the fifth stage after the... But I think actually now there's a much bigger opportunity to do that because I think in the past trying to get the sales team to recognise the value of admin—and we don't want salespeople doing loads of admin and there are ways to automate the process—but I think it's getting the mindset established with them that it will really have some value. And maybe in the past there hasn't been the real kind of use case for it. And now there is with AI.

I think how I would approach it, it would be to really get the CEO, the CFO, the CRO on board, show them the art of the possible with AI. Show them what a game changer it could be.

36:34  
That's easy, isn't it? You do this, you can use AI, you can make loads more money.

36:39  
Well, yeah, but just show them some of the examples that we've already talked about. What an impact it could have on their success, just by having that really rich CRM would take them forward leaps and bounds. So I think there's never been a better opportunity to try and drive that change and convince people that it is worth the effort.

Thanks for joining us. If you haven't already go ahead and subscribe to our podcast