← 返回

2026.09.22 | TB | Software | AI SDRs & AI Employee Market

原标题：PE Sector Update: AI Sales Development Representatives & the Broader AI Employee Market

#software #agents

Third Bridge · 2026-09-22

Analyst-led • 8 Sep 2026

Harry Bryars (HB) Founder at Ojara

Poorvi Banka (PB) — Third Bridge Sector Analyst

Agenda

Key Insights

Introduction

PB [00:01:52]: Welcome to Third Bridge's Interview entitled PE Sector Update: AI Sales Development Representatives & the Broader AI Employee Market. I'm Poorvi Banka, and I'll be hosting today's Interview with Harry Bryars, former Chief of Staff at 11x.ai.

Harry, before we start today's Interview, could you please state I agree or I disagree to the following statement: You understand the definition of material non-public information and agree not to disclose any such information, or any other information which is confidential, during the Interview today?

HB [00:02:23]: I agree.

PB [00:02:24]: Could you please provide a brief introduction to your background in this industry?

HB [00:02:28]: Yes. I managed sales and product teams at multiple start-ups before joining 11x right at the beginning. We then grew from zero to where we are today, and I did that through leading growth and then becoming the Chief of Staff.

Questions

[00:02:53] How would you describe the AI SDR [sales development representative] and the broader AI employee market today? What has changed most meaningfully over the last 12-18 months?

HB [00:03:07]: It's a great question. I would say what has changed most meaningfully is that if you look back, say, two years ago, every company would launch a highly verticalised solution, so there would be an AI SDR, for example, there would be an AI marketer, there would be an AI accountant or whatever. Each of these would be different industries with different companies. I would say the difference now is people are moving more towards horizontal and just general AI agent platforms and using those to then in one place, create many different types of digital employees or agents that ultimately do part of or all of a role. I would say the biggest difference has been where if you go back two years, it was very much, "This platform is for this thing, this platform is for this thing, this platform is for that thing." It's now the case that people are using the same platform to automate the vast majority of their work.

[00:04:15] Where do you think we stand today in terms of market maturity and adoption levels? We can talk about adoption at the horizontal level and across different verticals. Which ones have been using AI agents more than others?

HB [00:04:35]: I would say that as a whole, I think the growth space has seen the largest adoption outside of coding agents, of course. In the growth space, you can break it down into many different parts, of which AI SDRs are a grouping of those. Really, I would say that most companies now are using AI agents for everything from identifying and researching the right ICP, producing the right content for them, researching individuals, reaching out to them, ie, the outreach part of an AI SDR, call prep and notes and research, call note recordings, follow-ups after that, deal hygiene, etc. I would say that growth is still one of the largest non-coding use cases in the B2B side of things.

PB [00:05:36]: What are your estimates regarding adoption levels for this bucket?

HB [00:05:44]: In terms of the proportion of companies that are using these, I personally don't know a company that isn't using any form of AI. When I'm talking about that, I'm talking about funded start-ups and scale-ups. I'm talking very specifically about venture-backed companies. I don't know a single one that isn't using AI for part of, if not a significant proportion of that.

[00:06:10] What do the potential use cases vs the actual use cases for which AI agents or employees are being used look like?

HB [00:06:25]: Basically what are people not using AI for right now? Is that the question?

PB [00:06:34]: Yes. What they could use AI for potentially, logically, but are not using as of now?

HB [00:06:43]: I would say that many companies still have a human element to their sales process. I would say the largest unused part of that is replacing part or all of the AE role. Right now, it's very much the case that you use some sort of AI agent to automate or improve the efficiency of, let's say, your outreach, but you're still having a human take that sales call.

[00:07:16] When we think about the evolution of sales development, specifically AI SDR, where are the AI employees actually replacing human work vs simply making SDR more productive? How much of the market is genuinely AI agents replacing employees vs AI agents augmenting employee productivity?

HB [00:07:43]: I would say that most likely the case that these tools are still being powered by a user or a set of users. I would say that where your growth lever before would be, "Hire more SDRs," that's very much not the case anymore. It's not the case that you can quote-unquote hire one of these employees, press a button and never think about outbound again. It's very much the case that you still need a human to be driving and optimising them, but you no longer need 10 of those humans, 15 of those humans, 50 of those humans. You still need someone to own it and to optimise it and to be tweaking it. It is realistically a large portion of someone's day doing that, if they want to be great at doing that, but it's no longer the case that you need to be hiring as many SDRs. Now there are parts of this that I guess will come on to in questions later around what can't humans do. For example, if your sales process has a large amount of outbound cold calling, that's something where you might still have an SDR team, a human SDR team doing that on your behalf.

[00:09:00] Specifically for the SDR bucket, which are the use cases where customers are seeing real ROI from using AI agents? You mentioned cold calling is something AI agents cannot do as of now, but are there aspects where a lot of cost savings or other ROI realisations have been made?

HB [00:09:29]: Yes. In terms of ROI, that comes from two places, the first is a cost saving and the second is a revenue driver. In terms of cost saving, AI, right now is just great at the hard part of doing outreach. When I say the hard part, I mean just the labour-intensive part, which is really finding the right people, finding everything that you know about them from their contact information through to who they are, where they work, what they've been up to, basically everything that you need to be able to create a personalised message, crafting that message, ie, writing the e-mail, the outreach, the different touch points and then sending it on a schedule without having to load it into some sequencing tool, etc. That's a very real cost-saving part, and you could argue that it's also a revenue-driving part in the pipeline that it generates, the dollar figure associated with that. That's still the largest ROI piece, but it's not yet closing deals, which would make it an even higher ROI piece that you could just very clearly point back to.

[00:10:44] If we look at the economics of an AI SDR vs a human SDR, what would that look like? How are customers looking at that angle today?

HB [00:11:01]: During the sales process, it's very common for these vertical AI SDR companies to position this as a cost saving, and so what they will look at is the average all-in cost of a human SDR, which is the salary cost, the technology cost on top of that, the taxes, the benefits that you give them, whatever, just think of it as like an all-in cost for an SDR vs the cost of their equivalent output with an AI SDR. Now the reality is that (1) you might still need a human to drive this, so it's not necessarily a true cost replacement. It might be a cost reduction because you need far fewer of them. (2) The point here is that an AI SDR can output significantly more than a human. What you might pay overall might be more, but you're getting proportionately way more output, if that makes sense. Generally, as part of the sales process, you would get from any one of these companies an overview of you're paying USD 500,000 per year on your SDR team, you can get the same output for USD 100,000 per year. That's, whatever, proportion discount.

[00:12:22] What are the most common pricing models that are emerging in the AI SDR market? Would that significantly vary between SDR and AI accountant or others?

HB [00:12:35]: Yes. Like I said, the industry as a whole is moving to be more horizontal. In that case, consumption credits are the clearest pricing model, because you then have a platform where you can build any part of the process. It could be that you build an AI SDR. It could also be that you build an AI prospect researcher or an AI note taker or something, whatever. When you're moving to these horizontal platforms, the only pricing model really that they can offer is a consumption credit. Now there were many AI SDRs that have tried outcome-based pricing, ie, you pay us per meeting or per deal or something. Now I see you have per lead on your list. In my mind, that is essentially a consumption credit. It's just that what you're consuming is defined slightly differently. There are a few companies that did try outcome-based pricing. No one that I know of really stuck to that. The reason being the same reason why an employee ultimately wants a base salary. There is an unknown element of generally how well does your business do? Is it competitive in its space? Many people turn to, let's say, an AI SDR because they haven't figured out how to sell as a human themselves, and they think of it as a golden ticket.

If you purely rely on outcome-based pricing, not only is it a very unsustainable business model, it's just unpredictable and hard to build a business around, but it's much harder to actually achieve that success.

[00:14:23] Are there particular characteristics that make a company well-suited to use an AI SDR? Are there companies that are probably not using AI SDRs or not as well-suited to use AI SDRs at this point?

HB [00:14:46]: Yes. Let's take a two-year-old view of the industry. Ultimately, the original use case for AI SDRs was anyone with a large TAM. If you had a super large total addressable market and there were many people to go after the quality of your outreach, you could afford for it to be lower because you're blasting a wider space, if that makes sense. There are thousands of people you could reach, you would never be able to reach them as a human, and so if you deliver, let's say, a slightly less optimal message to a larger market, your outcome is still better overall. Where it didn't work at the time was where your messaging had to be super, super specific and very, very tailored in a low TAM market. Now that's gotten a lot better, but there are still many industries where at least the end-to-end part of an AI SDR, so going all the way through from identifying the person to reaching out to them isn't really appropriate. There are many industries that run off trust. They run off building relationships over a long period of time with people, meeting people, in-person events, things that generally aren't appropriate or possible for AI right now. I'd still say it very much is still true that if we're talking about an end-to-end AI SDR, then this works much better for larger TAMs.

[00:16:16] If we look at the broader AI employees market beyond the SDR segment, which other functional areas do you think could see higher adoption over the medium term, but aren't as well-penetrated as they could be?

HB [00:16:50]: A few things that spring to mind, or I guess the way that I think about this question is what areas of AI are significantly improving at, say, the model level. In my mind, that is content. The content is a huge use case for AI agents already and people using AI in their workflows. The quality of content is improving at still a dramatic rapid rate. I'm talking specifically about AI video, AI image here. It will very quickly get to the point where it doesn't make sense to hire a human to produce any form of content. There is no reason that you would have a photo shoot or a video team to set up a video. The quality of that is improving so significantly. Now I think that that's an area that's penetrated quite highly already, but there is still a significant amount of spend going toward human content production that will divert a large portion of that into AI content production.

[00:18:04] There seems to be a fairly important distinction emerging between AI copilots and genuinely autonomous AI employees. How is the market shifting in terms of having a human in the loop vs being fully autonomous, and how meaningful is that distinction commercially today?

HB [00:18:22]: I would say, it really depends on two things. The first is, do you trust the quality of the output. The original reason for having a copilot aspect was you didn't trust the output. You knew that this was good, but you knew that there were a certain number of cases that you just wouldn't be comfortable putting your business name associated to. You wanted to see the output, you wanted to be able to read the messages, make a small tweak here and there, put it into your own words, whatever it was, you wanted some visibility before it would send. I would say that people are now trusting agents to a much greater extent where the businesses that I know that have, let's say, a much broader set-up of agents across their business, they have very little visibility into exactly what is happening. They're just looking at the outcomes, and they go through an iteration process to get it to a point they're happy with. I would say the volume of work being done end-to-end autonomously has increased significantly.

[00:19:31] When you say the volume of work being done end-to-end autonomously has increased, how much more autonomous have these agents become in the last one year vs before?

HB [00:19:49]: Significantly. I can't put a number on it, specifically, but I know that the reason… Firstly, the quality of everything has got significantly better. The ability for these platforms to reason has got significantly better. I would say the UX of these platforms has also got significantly better as well. All three of those things contribute to… At the beginning of the AI SDR market, in particular, it was very much the case that the recommendation would be launched with some sort of copilot approach, review the output and then slowly and gradually move everything to an end-to-end autonomous output. Now I would say it's very much the case that people just, for lack of a better word, YOLO, and just click and see what happens because they know that the starting point is significantly better.

[00:20:46] Are AI SDRs being used to reduce human SDR headcount and hence, saving on cost, or are they also being used as a revenue driver where a larger part of the addressable market can now be targeted by the company with the same number of people, but with this technology?

HB [00:21:14]: It's very much both, and it very much depends on the stage of the company and who they are. For a company to have an existing SDR team of multiple SDRs, that's generally a larger, more mature company in the first place. For them, it's quite obvious that they can get cost saving. Now that's more of a defensive position as a company. We have this team, they're already producing attributable revenue, and so if we're looking to lower the cost there, there's a reason for that, ie, they're not being successful. Most companies, I would say, especially who are earlier stage, are looking to drive revenue. They don't already have that existing cost base. They might not know what SDRs do or that that's even a thing that people used to have. They're simply looking for how do we drive more pipeline, how do we get more prospective customers further down the funnel without having to figure out how to hire a team and how to manage the team, etc. I'd say it's very much a revenue driver for larger companies that might be struggling or might have over hired, then it's cost saver.

[00:22:31] The 2026 AI SDR industry report estimates roughly USD 3.1bn market growing at around 42% annually. Do these market size and growth estimates broadly resonate with what you are seeing on the ground?

HB [00:22:47]: When you say the market size, what is that? Is that the revenue, or is that the total value of these companies?

PB [00:22:57]: I think it is revenue basis, but whichever way you resonate with the figure, you can talk about it accordingly.

HB [00:23:05]: I would say from a revenue perspective, that seems plausible, especially if we're talking about usage-based across more horizontal players, but the proportion of that usage that is, let's say, growth-related agents. From a market cap perspective, so the valuation of these companies, that's significantly undervaluing, there are many of these companies individually who are close to those figures already themselves.

PB [00:23:34]: Which is why I would assume it's revenue basis, but as far as revenue is concerned, that is what you think they aggregate at right now?

HB [00:23:43]: Yes. I don't look at that number and think, wow, that's tiny or wow that's too high.

[00:23:50] How are you looking at growth for the AI SDR market in the medium term? What growth estimate do you have for this market as a whole?

HB [00:23:59]: I would say that growth comes in two places. Like I said earlier, every company that I know is using a platform like this, but I very much group that within funded companies, funded B2B SaaS companies. They are a very small percentage of the overall business landscape, and so I think there's significant growth potential outside of that. I also think there's significant growth potential in these becoming significantly more effective, therefore, having more usage, therefore, having much more revenue associated with them. I think that this market, the figure that you mentioned said growing 42% annually. I think that's a low estimate. I think most of these companies are significantly ahead of that. I think their ability to grow will only increase as ultimately this technology becomes more effective. When it genuinely gets to the point where it's indistinguishable between hiring a human SDR and an AI SDR or any AI employee, then things are just going to go absolutely insane in terms of growth.

[00:25:16] How would you segment the competitive landscape today, focusing on the AI SDR market, the broader AI horizontal platforms that are operating in the space and the different categories of vendors in the space?

HB [00:25:38]: There are a few different ways to look at it. There is, let's say, the end-to-end AI SDR, the AI-native autonomous SDR as you put it. I would say that there are fewer new entries into that space that are operating at a great scale. You have the same set of companies, let's say, as of a few years ago, but they're just doing it at a bigger scale. I would say where we've seen the biggest growth in new players, it would really be, I see you have on your list here like data or enrichment companies adding agents and then broader AI employee platforms. I'm just going to call those horizontal workflow builders where you're now able to have a platform that can address whatever need that you have. I would say that that is the single biggest point of growth, especially in the tool stack that I noticed people using.

They're often opting for a platform that can plug into a bunch of data sources and then be guided by them and have many different, let's say, agents doing part of their role vs an all-in-one platform that's quite rigid and means that you have to have one set-up for your AI SDR, one set-up for your AI content analysis, one set-up for your AI personal assistant, whatever it is. I think that, that is an area that we'll see more resistance, but I think the biggest growth in terms of a new segment is really the broader AI employee platform space.

[00:27:22] Is there any capability that the end-to-end AI SDR vertical-specific players offer, which the broader players probably can't offer and could be a reason why some companies stick to the vertical ones?

HB [00:27:42]: I would say it's more about friction, so a native AI SDR is going to have built-in lead sourcing. It's going to have built-in enrichment. Those things are going to be very specifically built around sales. It's going to have deliverability built in, so you don't need to think about how do I actually send these messages? How do I do it in a safe way? How do I make sure that I'm landing an inbox? It's going to have a much more resilient CRM integration. It's going to integrate out of the box with whatever CRM you're using, it's going to have a really simple UI to configure all of the different fields. There's no barrier to set it up. It's very much, "Hey, you have a quick onboarding call, you configure a few things and you're good to go." The broader AI employee space has a much higher barrier to actually creating something useful. I would say people's appetite is much more doing that themselves. They're already more generally using AI to replace existing platforms that they have. I would say that conceptually, psychologically, using a horizontal platform to build your own custom AI agent feels like a win. It feels like you're creating this thing super unique to your business. It's true that when you invest the time, you can get something that's significantly more tailored to your business and the different ways.

I'll give you a real example. Let's just take a random example of a company that use drones to survey insurance. Let's say there's a wildfire, you can send up a drone and the drone can analyse the damage and then whoever, I don't know who they might sell to, but that's an example where you're building out your own workflow. You could get really creative, you could be looking at data for wildfires or earthquakes, and you could be identifying of any type of company or any type of location, which is the most likely to need our services. Out-of-the-box native AI SDR will never be able to do that. They have some sort of data source and that data source has some set of fields related to businesses, and it might be that those fields are just not useful for you. The point there is the reason for using an AI-native platform is it's out of the box. It's already set up. It's already great or as good as it can be at being an AI SDR, but you can create something much, much more specific to your business with an AI employee platform or just agent workflow builder overall.

[00:30:36] In terms of customer segments, are there distinct customer profiles? Are there customer segments that would definitely prefer a vertical AI SDR despite some benefits that the horizontal platform offers, or are there customer segments that, despite the friction, would go to a horizontal platform instead of a vertical one?

HB [00:31:01]: I would say there might be some characteristics that you could say from a company profile perspective, ie, if you have an existing SDR team, then simply being able to replicate initially everything that they're doing is already a significant advantage. We already know that they have a workstream that generally probably end-to-end falls into the role of an AI SDR. Just like if you're a recently funded earlier-stage start-up, then building out something from a workflow perspective that's maybe more bespoke to you and is maybe more fitting of the stage of company that you're at, maybe you don't even have a CRM, you're still using spreadsheets and things like that. I think that's an obvious distinction, but I would also say it's just the profile of the person that you're selling to. Is it someone who is relatively new to AI adoption and let's say, they're a CRO who's burdened with a human SDR team and they simply want one platform that they can easily replace their team with or they can easily adopt within their team. The benefit there of being a platform is it's an individual contract that you can point your finger towards a company and say, "You're responsible for driving the growth." You're responsible for building a platform that does this well.

That's very different than, let's say, an individual contributor who's trying to build something to significantly improve their performance, where they might use a horizontal platform. There, they would be building out much more granular specific things based on their actual individual work stream and all of the different tactics that they have. They don't want to onboard a vendor that has an annual contract, charges them hundreds of thousands of dollars. They want something that they can easily get started with, with maybe no money or USD 10 or USD 100, etc.

[00:33:00] What do you think are the most important competitive differentiators in the market today, focusing on competitive differentiators for vertically focused AI SDR players and also the broader AI employee market level?

HB [00:33:24]: I would say, it comes down to a few things. I would say data is still a big part of this. Early on, most companies would just pull from the same data sources, which means they're all reaching out to the same set of people with the same things that they can talk about. A great example of this would be, if you go on LinkedIn and you move companies or announce that you've moved up in a company, you'll immediately start receiving the same messages from everyone. A clear example would be fundraising. If you've ever announced that you've raised funding, you'll immediately have a wave of low-quality automated outreach messages, let's say, "Hey, I saw you raised your series A. You must be thinking about hiring now." Now it's very much got to the point where each of these companies has acquired significantly better data, and that is maybe more proprietary. It might be that they have a data source that is much better for funding data or much better for wildfire data in that example I used before. I think data is still a big part of that. I would say that workflow and orchestration is a huge part of that. Which part of the workflow can you do? How much observability do you have? How customisable is it and how friendly is the UX? I would say things like deliverability infrastructure, they're commoditised.

There is only so much you can do related to deliverability, and if it's an AI-native SDR, you just absolutely have to have that. It's a need-to-have vs a differentiation. When it comes to integrations, that is also important. Do you integrate with the CRMs that they use? Do you have a good quality bidirectional CRM integration? At this point, every company should have all of these things. Model quality, absolutely not a differentiator. There's very little difference now between the models, and you would just expect these companies to use the best or give you the configurability to pick whichever one if it changes how they price, etc. I would say, ultimately, it comes down overall, though, to execution. You can have all of those things and have a terrible output. The output is by far the most important thing, and that's the thing that will drive recommendations and things like that.

[00:35:59] Does scale create a meaningful data network effect in the AI SDR category? Does an AI SDR that has handled millions of prospect interactions become materially better than a newer entrant, or is that advantage relatively easy to overcome? How would a horizontal AI employee compare to an AI SDR in this regard, given that the prospect interaction might be more for AI SDR-specific agents?

HB [00:36:30]: I would say that scale is a significant advantage. If I go to just, I don't know, ChatGPT and say, "Hey, ChatGPT, this is everything I know about a prospect, write me an e-mail." It's going to create a general e-mail based on what its understanding of how to write an e-mail is. That's a very different thing compared to what you can create when you have millions of data points to say, create me the most optimal high-performing e-mail. The biggest thing when it comes to any form of outreach is that things change over time. For example, if you're the first company that starts to reach out to people after fundraising announcements, that's going to work really well. If you're one of 1,000 companies that are in their inbox, that will work terribly. If you have a system that can optimise itself and learn from millions of different data points, it will be significantly better. How does that change with horizontal platforms? You lose the starting advantage of being in a place where out of the box, this platform has already engaged millions of people and has better researching, better prospecting and better message quality. What you gain is you get that data for yourself, and you can then build something that is much more driven around your own business and your own output. I would say you lose the out-of-the-box starting ability.

You gain the ability to do it maybe better yourself after some trial and error. If you want something out of the box that you just don't have to think about and it's more of the fully autonomous, "Let's just set it live and see how it goes," then your out-of-the-box AI SDR will be better for that reason.

[00:38:26] Where do you think existing sales engagement incumbents such as Salesforce, Outreach, Sales Cloud, etc, that already have customer relationships, workflow data and distribution have an advantage? How difficult would it be for an AI-native vendor to compete with them? Would it be easy for sales engagement incumbents to get an AI layer on top of their existing systems, and would that work better for them?

HB [00:38:57]: I would say there's a few ways to think about it. They obviously have a significant scale advantage. If you look at Salesforce and Agentforce, the revenue from that is just genuinely insane, you can't even comprehend how much revenue they drive from it. The quality of the product isn't great, and it can do a very small portion of that work, and that's the reason that people will use an AI SDR alongside Salesforce or HubSpot, etc. The advantage is that AI SDR companies can do significantly more of the flow than an incumbent might be willing to do. The general prospecting mass outreach side of things is very much a grey area in terms of service and things like that. For example, some SDRs will let you automate LinkedIn outreach. Salesforce would never let you do that. It's very difficult to do that compliantly with LinkedIn's terms of service. I would say there are a few ways to look at it. The first is these incumbents like Salesforce and HubSpot have a significant scale, and so proportionately, they can drive a very large percentage of the revenue that is attributed to the space. Quality-wise, they'll struggle to reach that. That's why Salesforce Ventures, HubSpot Ventures invest in many of the companies in this space as well.

[00:40:29] AiSDR, Agent Frank, Artisan, and11x.ai are some of the well-known AI-native companies in the SDR space. What position do these companies currently hold in the market? Are there any other vertically focused AI SDR players that we should add to this list?

HB [00:40:50]: Yes. Maybe taking them one by one. I would say that AiSDR generally is known for having a great product, but a terrible brand. At least at the beginning, they had something that was actually quite high quality at a lower scale. I would say that they were seen as being less enterprise ready. It was very much a product-led growth, you have a product you can onboard to and use, but the quality of the product in terms of the design, the usability, the branding, they very much narrowed themselves into a much smaller bucket by calling themselves AiSDR. An AI SDR can only ever be an AI SDR, whereas any other of these platforms can offer significantly more outside of that. Agent Frank, in particular, ultimately, that company is not known for the agent. That company, Salesforge is known for their e-mail infrastructure. I'm pretty sure they started as an e-mail infrastructure company, and I don't know any like private information about the company, but I know that a significant portion of things related to them were discussed relating to deliverability. They basically offered a way to create thousands of inboxes and have them warmed up and have all of the infrastructure underneath. They were one of a few companies that did that, and so they had a significant proportion of the market share. They have their own AI SDR called Agent Frank.

I don't personally know a single company that's ever used that, or I was about to say used it and had good things to say. I just don't know anyone that's used it basically. The next person on the list, Artisan. Artisan have the biggest brand by far in the space. Their out-of-home marketing is insane. I have been in many different cities and seen their advert, living in San Francisco and London, their adverts everywhere. Now they have become a very controversial brand. They had a lot of controversy with replace humans, don't hire humans messaging. I think that that got a lot of eyeballs. It's rage marketing, but it also makes a lot of people just hate the brand. They very much have flip-flopped between trying to do something more enterprise focused and trying to do something that's more like product-led growth where you can self-onboard, you get a free trial, the entry point and cost is lower. I would say they're still in that bucket and very much the same with AiSDR as well, where their ACV is lower, the types of companies they're working with are at a much smaller, less mature scale, basically. I would say 11x is a counterpoint to that, which is their customer base is genuinely enterprise focused. There is no self-serve motion. There is no low-cost entry point.

It's very much an expensive product that is tailored towards larger enterprises. You can see from the logos simply if you compare the logos on the websites. Each of these platforms generally has one or two enterprise logos that they throw on there, but those are mostly a small team or an individual at those companies using this as opposed to the entire company adopting it. I would say other players in the space, Monaco is a newer company that has come out recently that has firstly, just done very well from a growth perspective, but more importantly, been very well-recommended by a significant portion of the people that have used it. I've seen many people say great things about Monaco. Rox.ai is one of the larger players in the space from a market cap perspective. You mentioned earlier the USD 3bn number. As of their last fundraising, I think Rox's valuation was USD 1bn, USD 1.2bn, USD 1.5bn or something like that. From a valuation perspective, they are significant but they seem to be talked about less and less over time. I'd probably add Monaco to the list in terms of a newer AI-native platform. Yes, the rest are mostly the original group that haven't really seen super significant growth.

You then also have Unify GTM. They pivoted a little bit. They pivoted more into a prompt-based, go and find me all of the people that relate to X, Y and Z. I know that their platform, again, is seen as very high quality. It has great UX, but they rebranded a little bit and their positioning is a little bit less clear, I would say, right now.

[00:46:40] You mentioned Rox.ai as one of the highly valued players, but it is not talked about as much now. Is there any reason for that? Could you provide more context there?

HB [00:46:54]: All I know of Rox is that they built a nice product from a design perspective that no one ever talked about. In terms of the sales process, they didn't come up once. In terms of companies that would move from one AI-native platform to another, it was never the case. They were one of those companies that I would always read about simply because of their fundraising. I think they made some acquisitions, but I never had any experience with people actually using them or say anything about them other than I tried out these. I think they had a self-service motion. I didn't hear anyone say anything other than I tried it. It was nicely designed, but it just didn't work very well. Now that was a while ago, and I'm sure they have improved significantly. They have some interesting logos on here. They obviously raise a lot of money at a high valuation, but I haven't seen actively anything about them anywhere that I would be seeing that sort of thing.

[00:48:01] In general, is it easy to switch vendors in the space? Is it easy for companies to stop using the AI SDR vendor that they are using and use another one? What is the average churn rate in the industry currently?

HB [00:48:18]: I would say there is some friction to doing that. Generally, the platforms would own, say, the e-mail infrastructure. It might be hard unless you're syncing back the data to your CRM to move to a new company and have all of the existing context. It's not overly difficult other than the fact that generally, these companies, at least on the enterprise side, there are obviously annual agreements, and so you would either lose the rest of your contract or you would have to negotiate some buyout with a new vendor. That's not the case. It's like an AI SDR or something that's more product-led growth, and it's just a monthly subscription. I would say, really, it's just asking the question, if you had a human team, you can just fire the team and hire a new team. It's just the friction of doing so, but it's not impossible.

PB [00:49:12]: What does the churn rate look like for this industry currently, on an average?

HB [00:49:19]: I would say it's gotten significantly better for companies that are already performing well as a company. If you're a company that is new or small and you can't even afford to hire an individual SDR and you're using a company like an Artisan or an AiSDR or an Agent Frank, where the price entry point is quite low, a few hundred dollars, a few thousand dollars a month, you'd expect churn to be much higher there. I don't know the actual churn numbers of those companies, of course.

[00:49:51] How much of the addressable market for AI SDR companies is with the horizontal AI employee companies as of now?

HB [00:50:08]: I would say that there are multiple horizontal companies that have launched within the last year that have individually more revenue than most of these AI SDR companies combined from a growth use case part of their usage. I would say that this is now a much more topical, much more relevant segment of the market is people using companies like those vertical companies to build out growth-based workflows.

[00:50:41] 11x is increasingly talking about a broader revenue workforce. Do you think the vertically focused AI SDR vendors can move beyond SDR use cases and become broader AI employee companies themselves, or is that too ambitious?

HB [00:51:02]: I would say when you start in a vertical, the point at which it makes sense to expand out that vertical is you have a much higher barrier for that. If you're still not dominating your vertical, if your product is still not as good as it can be, then it's very difficult to make that point to investors and your team that you should divert mind share and focus away from winning your existing category. When you're a horizontal company, it's very much just a case of you add more integrations and a little bit of refinement and that brings aboard a whole new vertical that you can move into. "Oh, we added some finance-based integrations. Now we can build finance-based workflows on the platform." If you're an existing company that has raised a lot of money, for example, if Rox went to [Firm] and said, I know we've raised like loads of money, but we're going to move into finance agents, that would just be blocked as a decision. It wouldn't make any logical sense.

[00:52:12] Can vertically focused AI SDR companies potentially go to some adjacent use cases beyond just AI SDR because the market is inclining towards horizontal platforms? Is there any company, in particular, that is well-placed to do that? How much market share of the SDR segment do you think horizontal platforms hold now vs the vertically integrated ones?

HB [00:52:51]: I would say there is obviously an ability to move out within the growth space. I think there's more of that, that these companies can do. For example, AiSDR is a company, they're fenced in by their name. Everyone else has the ability to move out into the broader growth space. I wouldn't be surprised if some of these platforms start offering ad integrations, if they help you produce content outside of written content. I think that is definitely possible. In terms of the market share between incumbent or let's just say, AI-native SDRs and horizontal workflow builders, I would say that anecdotally, I would estimate it's 50/50 at least right now. That's only because I know multiple companies that have seen significant revenue growth far beyond any of those AI SDR companies. I think that, that will only grow significantly as the quality of these platforms becomes greater.

[00:54:00] How do you expect M&A to play out in the industry? Do you expect some of the horizontal workflow builders to perhaps acquire some vertically-focused AI SDR company, or can the AI SDR companies survive on their own in the longer term?

HB [00:54:19]: I think in terms of M&A, that will come on the AI-native SDR side of things, that will come from CRM platforms, larger data platforms, larger sales platforms that will not come from horizontal agent builders. The whole point of horizontal agent builder is it's a thing that you can quickly click a couple of buttons and have in one place. An AI SDR platform is the exact opposite of that. I think their acquisition potential is still very much Salesforce, HubSpot. You can see that in terms of where similar companies have been acquired before. It's sales incumbents who want access to the customer base ultimately and the revenue, and it's not horizontal competitors.

[00:55:10] Could you call out a couple of horizontal AI employee companies that are doing particularly well in the SDR segment?

HB [00:55:24]: There are smaller companies like Viktor is one that I've seen have significant growth. I've seen a lot of people using that for many things, but growth being a big part of that. I would even say tools like Grok Bot, so xAI's horizontal platform. These types of tools are the tools that people are using to say, how do I optimise my entire business? How do I optimise my customer support and my personal assistant and my growth? I think we'll see more of those, but I think they've had significant growth, and I've seen a lot of people using them, particularly for growth-related things.

[00:56:04] Which vendor category do you think has the strongest chance of being the long-term category leader? You did mention that you expect faster growth for the horizontal category. Between the CRM incumbents – who might acquire the AI-native SDRs or develop an AI layer of their own – and horizontal workflow builders, which category do you think will win over the longer term within the SDR segment?

HB [00:56:37]: I think, ultimately, AI-native SDR companies have the ability to get to a point where their product is genuinely better than a human SDR and is as easy to adopt and train as an SDR. I think horizontal platforms will never be better, but they might be better because they win more customers because they're easier to use and they have lower entry points and they don't lock you into contracts and things. I think in terms of a success perspective, then an AI-native SDR has the best chance of getting to the point where, oh, we need to hire an SDR, let's just click this button on this company. They're not there yet. No one is even close, I would say, and so they have the best ability to do that.

[00:57:25] Do you have any closing comments you would want to leave us with?

HB [00:57:34]: No, I would say, high-level summary of my thoughts on the market. No one yet has realised on the vision of building an AI employee that is as good as hiring an employee. Right now, there are still trade-offs. If it gets to the point where this thing is as good, if not better than a human. It doesn't even need to be better. It doesn't even need to be as good, but if it's close to being as good as a human, then the growth will just be absolutely insane. Every single company will just default to hiring that platform. That platform doesn't exist yet, but I think when it does, then that will change everything significantly.

PB [00:58:12]: Is there any player in particular that you think is closer to becoming that sort of a platform than others?

HB [00:58:20]: It's hard to say right now. I would just bet based on capital. I think that some companies they have more ability to take more swings than others, and I think that's where I would place my bets.

PB [00:58:35]: All right. Great. Thank you so much for that. We can wrap the Interview here. It was lovely speaking to you. Thank you so much for your time and all the insights that you've shared today.

HB [00:58:43]: Thank you so much. See you later. Bye-bye.

Transcription ends at 00:58:48 of the recorded material