#450 Peter Bonney on How AI Is Disrupting the RFP and B2B Sales Process

[00:00:00] 
 

Mehmet: Hello and welcome back to a new episode of the CTO Show with Mead today. I'm very pleased joining me, Peter Bonney, CEO, and founder of Fastbreak. Peter, the way I love to do it, as I was explaining to you before we start the recording, is I keep it to my guest to introduce themselves. [00:01:00] Tell us more about you, your journey, your background, and what you're currently up to.

Mehmet: So. The floor is yours. 

Peter: Sure. Yeah. Hey, uh, thank you very much. Me pleasure to be here. Uh, so Fastbreak is my second startup. Uh, first startup was, uh, a procurement solution, uh, called Vendorful um, and in that bus, in the course of that business, I saw that a lot of companies would struggle to respond to RFPs because we were helping buyers issue RFPs and manage their suppliers.

Peter: I also had to respond to a lot of RFPs and security questionnaires as part of the sales process. 'cause we were selling to large companies by and large. Uh, so yeah, so along the way had the idea that we should try to automate the response process using ai. And once the tooling, you know, in a post chat GPT world kind of made the natural language parts of that challenge feasible.

Peter: Um, I got really excited to do it, and once I'd exited ful, uh, it was time to, to [00:02:00] focus on that product. I, I already knew exactly what we wanted to do. Uh, and so that's, uh, that's what we're doing. Fastbreak helps large organizations and medium sized organizations as well, respond to RFPs, complete security questionnaires, uh, using their own content, uh, and, and automating the process with ai.

Mehmet: Great. Answering a lot of RFPs also as well. Mm-hmm. So, so. Let's start, you know, by focusing on, you know, the way you are trying to simplify this using ai. 'cause you mentioned like the chat GPT. So where do you see exactly the AI playing the role in, in streamlining, you know, these interactions, uh, between, you know, clients and, and vendors?

Peter: Oh, that's a really interesting question. 'cause I we're taking, you know, kind of a little piece of it here, which is, you know, at the, after the engagement has already happened, right? And you need to. Exchange these lengthy questionnaires where buyers have a lot of [00:03:00] questions they want answered, whether it's about, um, the capabilities of a solution if we're talking about a software company, but it's not just software companies, right?

Peter: Like, um, health insurers answer RFPs, manufacturers. It's really. Any industry services marketing, um, it, it really runs the gamut. And so people wanna know about the service offering, the buyers wanna know about service offering, and especially if it's a, a very expensive purchase. So they put out these lengthy questionnaires, um, and they're gonna go out to multiple suppliers, right?

Peter: And, and they want to compare them. They want the best information possible, uh, to do the best comparison possible. And that is a whole separate challenge. Like that's actually a hard process to do for, for buyers. Um, but it's a hard process for sellers to complete these questionnaires, especially when you start getting into the hundreds or even thousands of questions spread across multiple domains, uh, and multiple, uh, divisions within a single company.

Peter: So where the AI comes in is building a [00:04:00] semantic understanding of all of the content in your company. Um, and, and, and complex things like how that content needs to vary across products and geographies. Because the way that, you know, even for a single company, let's just say AWS right, uh, the way a WA salesperson to AWS.

Peter: Would answer an RFP for compute in the Americas? Be very different from how they're going to respond to an RFP for document storage in Europe. Right. It, it's, it's different products, different geographies, different regulations, different laws, different content, all within the same umbrella. Similar sounding content, which is an easy thing for kind of a naive approach to get confused about.

Peter: So yeah, so we spend a lot of time building a semantic understanding of our, uh, customer's content, and then when they receive a new RFP, in some sense, that's kind of the easy part, even though that's really sort of where the, the [00:05:00] magic happens from the customer's perspective. You've already built up the knowledge base, you've built, you know, kind of this, uh, this knowledge map of their content and their information and what they offer.

Peter: And then you just do your best job to answer questions. Um, and that just is like a tiny little slice of the whole process. Uh, and then from there, that's where the human effort really begins to kind of like jazz up your response, work with your collaborators, make sure that you've got everything the way you want it to be, to put your best foot forward and not just answer the questions in a rote manner.

Peter: Uh, but you know, we, we, we help with all of that, and our goal really is to let our customers. Uh, leverage AI in order to s kind of make best use of all the human's time, right? So that they're actually doing the value added activities in the process. Does that make sense 

Mehmet: now? Yeah, absolutely. So now, you know, again, RFPs 

Peter: Yeah.

Mehmet: Close to my heart. You can [00:06:00] imagine right now. Because you've also, you know, your previous startup was in, in the procurement efficiency, right? Mm-hmm. I want to talk about the elephant in the room here. So why we still have some time this inefficiency in the procurement process itself before, like I go back and ask you further questions about, you know, the AI and answering this from, from vendor perspective.

Mehmet: So you, you've done this, Peter, yourself. Um, why we still struggle in procurement because, you know, the other day I was thinking maybe one of the few areas which I still see. If I want to call them legacy is how we procure, especially enterprise software. Mm-hmm. We've been doing it the same way for a long time.

Mehmet: So what, what is stagnant there? Why, why it's not moving fast? 

Peter: Oh, uh, yeah, good question. It's a challenging problem and, and I, I certainly don't wanna [00:07:00] downplay it. I think procurement gets a very bad rap for, you know. They don't buy things the way that sellers want to sell them, but it doesn't mean that they're necessarily doing things in a bad way.

Peter: They've got a very challenging problem to solve, which is how do you spend money efficiently at scale? Right? Covering a whole range of spend from. You know, making sure that like slow dollar spend that adds up across the whole organization is, is controlled efficiently. So, you know, you've got 10,000 people buying laptops.

Peter: It is a challenge to make sure that they're getting the best price, that they're buying the right product, you know, from the right provider getting preferred pricing, that it's getting fulfilled in a way where there's not a million phone calls going back and forth to settle the invoice. Um, and that is, you know, its own challenge in procurement, procure to pay.

Peter: And then, you know, at the strategic level [00:08:00] it is also a challenge and a very different challenge to decide like, we're gonna buy an ERP system, say, or a CRM system. And if you're a large organization that touches a lot of people, it touches a lot of processes. You cannot take that decision lightly. 'cause you don't just have the cost of buying the software, you have the change management cost.

Peter: Right. And I think that's where a lot of sellers I think are, are, are kind of. Maybe not empathetic enough to what's happening on the buy side and why it is can be seen as so inefficient. Behind the scenes, there is this huge body of stakeholders that are going to be affected, like in the case of, you know, enterprise software by the purchase decision, and they all need to be consulted.

Peter: There's a lot of risk that needs to be managed. Before you sort of pull the trigger. And the challenge is that, you know, when you talk about enterprise software, you can do all the pilots you want. You can do all the demos and evaluations you want, get all the [00:09:00] references you can until it's actually implemented in your organization for your specific needs.

Peter: You don't know how it's actually gonna work out. You're just doing your best job to, to gather as much information as you can. So. It's a hard problem and I think they're just very, very focused on minimizing risk and um, you know, no one's come up with a better way to do it than going very, very slowly and asking a lot of questions.

Mehmet: Yeah, it's sensitive. I can understand. And because also no one want to be put into a, uh, situation where he would be asked, Hey, like, why did you choose this solution rather than that solution, or why? Um, it looks like you did a favor for these guys, um, more than the others, right? So no one want to be Yeah, I can understand the sensitivity.

Mehmet: Now let's go back to what you currently do with the, with the, with Fastbreak. So, and talking about, you know, [00:10:00] I've been there also myself, um, in, in B2B sales job, I did the SE job. I did the job also as well. Mm-hmm. So. We often hear that ai, you know, is coming to our jobs. Now, of course we know the answer, it's gonna make us more powerful.

Mehmet: But from the sales perspective, are you seeing AI replacing? Because it can take the RFP for example, it can act on it, it can give insights, so. Where are you seeing, you know, the role of salespeople, whether they are the AEs, the even the SDRs, even the ses. So how are you seeing these roles affected, um, you know, with, with the rise of ai?

Peter: Yeah, that's the big question on everyone's mind, right? And well. What I'm seeing, I don't think those people are going anywhere soon. I don't see account executives going anywhere soon. Um, SDRs I [00:11:00] think are maybe a little bit more at risk. Here's the model. I've got the capabilities of these systems, even the best systems, uh, and the best implementations of the latest models with, you know, the kind of, you know, most robust workflows and, and sort of agent frameworks around them.

Peter: They're really at the level of a highly skilled intern. Right. And I think that's the thing to, to bear in mind. So work that could be done by an intern or by an army of interns, that work is definitely at risk of being replaced by ai. By ai. And so, you know, SDRs to an extent like. You know, an entry level SDR without a lot of experience, maybe, you know, not that great at their job.

Peter: Um, just kind of an average run of the mill SDR. Yeah, a lot of that work, I think, ultimately can be replaced by ai. Great SDRs. I don't see them going away anytime soon. And I think everyone's [00:12:00] had this experience when you've worked with like a top decile, SDR, the difference between a top decile, SDR and, and like a median SDR, it's night and day.

Peter: Um, I. There's the reason why those top decile SDRs are so good is they really have mastered the human element of building rapport within 10 seconds, of connecting with a prospect. And I just don't see AI going there anytime soon. And then, you know, when you get sort of to higher and higher level activities, uh, it's, it's.

Peter: It becomes even less likely. What I think AI is going to do and what we are seeing in our business is that it is reducing the amount of rote work that you would push off on an intern if you had an intern or if you had 10 interns. Right? And. So the interesting thing, like let's set, you know, salespeople aside for a second, probably the people who you think might be most [00:13:00] directly threatened by software like ours would be the people who are actually writing proposals and whose job is responding to RFPs.

Peter: 'cause at large companies, there are entire teams dedicated just to responding to RFPs. You might think those people don't want to sort of bring in software like this because it's gonna replace their jobs. In fact, it is the exact opposite. Those are the most enthusiastic adopters of our software. Those are our best champions because.

Peter: What it does is it enables them to, you know, stop spending time on things like, you know, answering for the hundredth time. You know, do we encrypt our data at rest and in transit? I mean, the answer to that question for every single company in the world is yes, right? There's no serious company out there that does not encrypt their data in transit and at rest, and yet that question is never going away from an RFP response.

Peter: Um. And so like the proposal managers are tired of answering that stuff. They don't wanna have to go [00:14:00] ask somebody about, you know, some technical question like that, that they know has been answered a million times. What they wanna do is they wanna focus their energy on, uh, the creative work that they're actually very good at, right?

Peter: Like crafting the proposal, like taking a stock answer that an AI produces and adding in the value add that they have gleaned from context of understanding the customer to sort of shade like. Okay, this is, yes, this is what we do, but let me phrase this in a certain way to address the specific problem. I know that you have, that the AI can't know or can't fully understand at least.

Peter: And that is the value add of these AI systems. It's letting people spend less time on the low value work and a lot more time on the high value work, helping 'em win more business. 

Mehmet: Absolutely. Now, on the point of, you know, these RFPs and, you know, these questions can, I mean, is it something that also you can kind of, let's, I would call it modernize the [00:15:00] old, uh, RFPs that, you know, these companies they ke keep using again and again.

Mehmet: Is this something like. Also you focus on, I gotta give you an example. Yeah. Uh, I, I work, I work in the data protection space, so we, we, we've seen like some requirements which are obsolete. That comes again and again, and most probably because, you know, when these guys, they prepare the RFP all, what they have done is they changed the date.

They 

Mehmet: change a couple of things and then, you know, they copy past it from, from the old template. So is this something also you, you can maybe highlight on as well? 

Peter: Sure. So we're not, we're not doing that. Um, but I'll look, I'll tell you one thing I'm interesting that I'm seeing. We are seeing more and more telltale signs of procurement teams and, and buyers more generally, uh, using chat GPT to construct RFP questionnaires, which makes sense, right?

Peter: Um, you know, the, the old. Wisdom with RFPs, you know, from sellers. Used to be, if we didn't help write the RFP, we're not gonna win the RFP. [00:16:00] Well, the reason why sellers were helping write RFPs at all is because buyers are not domain experts in the things that they're buying. So, you know, if you are. An IT procurement professional, you're expected to be, to have a, a general knowledge of a whole wide range of areas.

Peter: Like you might have to buy managed DNS and database and cloud hosting, right? And you can't be an expert in all of those things. So when you need to buy a new managed DNS service, what do you do? You call up a vendor, say, Hey, can you help me with this? Like, I need to issue an RFP. Um, you know, I don't know what to ask.

Peter: Help me out. They didn't wanna do that, right? They didn't wanna have a vendor help them write the RFP, but the vendors were the experts they knew to go to. Nowadays, uh, we are seeing more and more evidence that they're not going to vendors. They're going to chat GBT or Claude or Gemini. Um, and they're leveraging kind of the expertise of the whole internet, uh, and large language models and, and the ability to synthesize that information [00:17:00] and building up, um, these templates using ai.

Peter: Some of the telltale signs are that all the sections. Kind of have about the same length. Uh, there's not a lot of overlap between the sections as, you know, kind of in your example, right? Like you, when humans have built up these RFPs over time, you get a lot of repetitive questions. Like the same question asked two different ways in two different sections of the RFP because you had multiple authors contributing.

Peter: Um, you'll see big changes in language like, you know, like the security section has like a very, very clear, uh, tone of voice because it was written by one person and then the. Requirement section is completely different because it's written by a different person. We're seeing a flattening out of that, and that is a hundred percent a sign that, uh, chat, GPT or some other LLM wrote the RFP.

Peter: So, um, definitely buyers are using solutions. We're not providing them, but they're definitely leveraging AI to help with that. 

Mehmet: Yeah. Interesting point on, uh, on that one, and this is why I asked you. You know, a couple of minutes ago about [00:18:00] the process itself. So I know for a fact, and we used to, if every seller right comes to us a a, a var mm-hmm.

Mehmet: Uh, with an RFP, we sometime we used to deny even answering it because you said if we didn't write it, we'll not win it probably. 

Mm-hmm. 

Mehmet: Um, which sometimes it looks like. Controversial. So you need some pipeline. You need also to be able to go and maybe last minute, flip the deal to your side, right? And you need to have this visibility.

Mehmet: But at the same time, you know for a fact that. You know, this customer got a champion and you know, sorry, the vendor got a champion and the champion, you know, asked them to write RFP in a specific way, which is again, okay. So we know who gonna be the winner. Yeah. And the funny thing is, if it, sometimes it happened, but of course like customers are smarter and you know, they know exactly they can differentiate.

Mehmet: So when they [00:19:00] float the RFP. You put, let's say, unique points for you as a vendor. Mm-hmm. And then you get everyone answering. Yes. Right? Mm-hmm. So yeah, we can do that. And uh, it's like, this is why sometime, you know, I think like this, something needs some, I don't know, some breakthrough I would say in the way how the whole thing works.

Mehmet: I understand the process, right? So I understand like, yeah, of course everyone has the. Right to spend time in the account trying to explain why they are better and then influence this decision somehow. But again, you know, the way it worked from procurement perspective, I know procurement people, I love you so much, but I mean, like, it, it sometime it doesn't make sense to me.

Mehmet: It's just my 2 cents on this. I think you want to say something, Peter? 

Peter: Yeah. Well can I, yeah. Can I give you, look, can I give you my kind of wild and crazy Sure. How I. I've now been dealing with RFPs in one form or another for, uh, for, for about nine [00:20:00] years, uh, now with two different companies tackling it from different sides.

Peter: My ultimate grand vision here and, and the reason why I'm so fascinated by this problem, I. Is, I would like for the RFP to go away. Um, that's my ultimate goal to like make the RFP obsolete. That is what I wanna achieve. And, and I think it is feasible leveraging AI in a way that it wouldn't have been 10, 15 years ago, or even, even five years ago.

Peter: Um, it's gonna require a lot of effort to get there, but I think that the ultimate look, the reason why companies run RFPs is they want to get. A uniform set of information as best they can. Uh, and they want to compare apples to apples as best they can and they wanna make an informed decision. The best way anyone's come up with so far to do that is to put out these lengthy questionnaires and have vendors respond to them.

Peter: But as we've discussed most of those questionnaires, most of the content in them is pretty rote. And so if you just had on both sides, AI sort of. [00:21:00] Contributing where the AI on the buyer side says, okay, look, um, you want to buy, go back to managed DNS, you wanna buy managed DNS. You know, here are, you know, 200 things that we, two, 200 boxes that we need to check in order to, uh, say, yes, we can buy this solution.

Peter: But here are the. 10 to 50 things that are actually going to make a difference for your decision. Right? The things that affect your business specifically, like why would you go with one business or another? Well, are you a, um, media company, like a streaming video company? I. Or you know, is your managed DNS really just about kind of internal servers and you just have like a public facing marketing website, right?

Peter: Those are very different use cases. Right. And they're probably gonna be very different solutions that serve your needs. No, no one managed DNS is necessarily better than the other. It's just what's best for us. So the AI can help focus the buyer on what matters. And then on the sell side, right, can just [00:22:00] tell the ai, the buyer's ai, like, yes, we check all these boxes, and now let's let the humans focus in on the things that are actually gonna make a difference.

Peter: And instead of spending all this time on the 80% of the RFP, that doesn't really move the needle. Like let's just get all the people talking about the. Handful of things that really matter and streamline the process. So like that's my ultimate vision is to kind of like turn this whole RFP process on its head and just like have people sort of asking each other, you know, a small number of questions rather than a million questions.

Mehmet: Peter, you would not believe this 'cause I just opened to see the date, uh, April 9th, 2018. So that's close to five years ago. Yeah. I wrote an article on LinkedIn. Uh. Selling the technology to robot. And I was talking that at some stage we're gonna see yes, the robot selling to the robot, right? So of course at that time I didn't know it's gonna be called GPT or something like this.

Mehmet: I was calling a robot, kind of an AI robot. Uh, and it seems like we, [00:23:00] we, we, we reached that like, and uh, we start even to see these, uh, videos where two AI models on voice mode, that they talk to each others. Uh, and, and I think this is good because mm-hmm. Steve, I, I believe one of the mistakes that even myself I used to do is I used to assume that customers doesn't always know what they want.

Mehmet: Right? So like. Yes, of course. But now more than anytime before they have the ai and even if they don't know, so they're using ai, kind of taking, you know, all these ideas, best practices, all the knowledge of the word is in our fingertips. We used to say it for search engines, but now with AI and the. Deep search modes and, you know, the thinking modes and all what I'm seeing.

Mehmet: So, you know, I, I ask, for example, I just get the access, uh, 'cause I'm not on the higher tier and, uh, open ai, but I get the access to the, to the search mode with, uh, with chat GPT and I ask it for like a detailed report on something and mm-hmm. It took like, I [00:24:00] think 15 or 20 minutes and Wow. Like I get everything I want and I'm, I'm sure now.

Mehmet: Maybe a CIO or a CTO, uh, or even the CO, like, they will write something there. Hey, based on the trends that are happening, blah, blah, blah, what should our B priorities be? Who are the main vendors that we should be looking at, right? 

Mm-hmm. 

Mehmet: Um, so I'm, I'm seeing this going to this one, but now I want to ask you something, and this is something also you, you, you try also to optimize.

Mehmet: Sales cycles in B2B are lengthy ones. Mm-hmm. I can't tell you how slow they are here. For example, in a region like the Middle East, it's very lengthy because also you need to build to build the personal touch the personal relationships. Yeah. Now the question is how AI can help us. I know it cannot shrink it from, let's say nine months or one year to one month.

Mehmet: Yeah. But how it can help us in at least minimizing it. 

Peter: Yeah. Look, that's a very good question. [00:25:00] And, and I'm, I'm definitely getting a little bit out of my lane here with my answer 'cause I'm, you know, I'm involved in a tiny little bit of the sales process, but mostly I think I'm, I'm thinking about how I use AI as a seller because in addition to selling, you know, software that is falls under sales enablement in some sense, I am also selling and, and, and leveraging AI to become more efficient at it.

Peter: The ways that we use it internally, and I think this probably points in the direction of, uh, of how it will impact sales in the near term. It doesn't, there's no sort of silver bullet that kind of, I. You know, radically speeds up the process. But there are small efficiencies at each step in the process where overall it's cutting cycle times and, and that adds up over time.

Peter: So like little things like, um, you know, we use, uh, fathom as [00:26:00] a call recording system, and one of the cool little features it has is that, you know, once it's built up, it's uh, you know, it's finished with its transcript. You can click a button to get a a, a draft of a follow up email and. It's not always perfect, but I would say 90% of the time it is 90% good.

Peter: Right? And so, like, you know, that means like essentially 80% of the time, like there's sort of like a, you know, maybe a 20%, you know, 90% times, 90%, uh, you know, you get sort of 80% good. Uh, and 20% of the time you need to actually kind of redraft things. Um. That's a significant time savings. It's like, kinda like 80% of the time we're basically just clicking a button.

Peter: There's the follow up. I don't have to have this long to-do list of like, oh man, I had, you know, three sales meetings today. Now I've gotta go back, look at the meeting notes, think about what to say, you know, what are the materials I need to follow up with? Like, I just click a button and, you know, I've got my, I've got my draft follow up and I can send it right [00:27:00] away.

Peter: Like, I can't control a lot of things, uh, that happen in a sales process. I can control how fast I respond. My people can control how fast they respond. Uh, and so, you know, little things like that, um, really compound over time. Particularly I think as buying teams are getting bigger and bigger, you're having to talk to more and more people.

Peter: There are more and more individual follow-ups to make after a sales meeting, after an email engagement. Um, and ai. The ability to draft messages to, uh, you know, suggest content sources to suggest actions, really helps reduce the cognitive load of having to think of each step. And, and that's the biggest deficiency we're getting, uh, from, from AI and sales at least today.

Mehmet: Yeah, absolutely. And, uh, maybe one thing I can think about, uh, so, uh. [00:28:00] I would put myself in the AE shoes for a while. So I did a call and then the guy asked me a question and then, you know, I need to go to the SE and do something and the guy, he have to go and find some documentation, whatever. So I believe that these things can be streamlined.

Mehmet: The other thing is, you know, uh, probably, I'm not sure if how, I'm sure like AI will, would be able to help. I can't figure out how exactly. So for example, if. Let's say it detects like it's taking too long till we get, for example, some responses or something like this. 

Mm-hmm. 

Mehmet: So probably they, I would have the insight.

Mehmet: I, I, I'm not sure. Maybe using some social tools or something like this. Hey, by the way, like you're talking to this guy, which looks like he's not the main decision maker. Like, why you don't do. Speak to this guy, right? For example, or, uh, have you tried, for example, to speak to the, uh, line of business manager?

Mehmet: I dunno. 

Peter: Yeah, [00:29:00] no. So like, there, it's funny, actually I met, I, it was just at an event last night where I met, uh, met somebody whose company does exactly that, right? It's it, they look at Salesforce data. And, uh, and try to find, you know, like, um, what does good look like, right? What, what, what is sort of your best sales processes look like and what can we learn from that, from, you know, what's the next action you should take?

Peter: What's the next person you should involve? Um, and then I, you just triggered my, my memory, like I'm talking about like how, you know, software can and AI can speed up the sales process. I'm forgetting actually, like our own product does exactly what you described, like the virtual sales engineer. Instead of having to wait for somebody to, you know, somebody gets, sends in a technical question instead of having to wait for somebody or go ask somebody, how do I answer this?

Peter: Just type it into the ai, like it's the same content that you use to respond to RFPs. Um, just ask, you know, we have a feature where you just ask a single question, like, how often do we test our disaster recovery procedures? Like, instead of having to go find somebody in information security or it just get an answer.

Mehmet: [00:30:00] Absolutely, yes. It would be much faster. And the only thing they need to figure out. Mm-hmm. Because okay, if we are doing the colon a Zoom or teams or whatever, fine. We can use any of these AI tools that can take notes. But, and I think, you know, it's good we started to go back to face-to-face meetings and visit, you know, so.

Mehmet: I can take notes on my notebook or even on my iPad or whatever, but I mean, it would be nice to have also someone to listen with us and then take this, you know, because I can tell you every single sales guy team is to find a way to automate. These entries in the CRM, right? Yes. So, so if we can find a way to automate whatever the insights real time in the CRM, I think like everyone will be chills.

Mehmet: Totally. Totally. Yeah. So, yeah. But I, I know like it's gonna be a, a, a journey and, uh, what you're doing, Peter, I think it's, it's great, uh, advancement in, in the whole process [00:31:00] because again, it saves time. I'm, I'm biased, coming from the same background. Loving technology, AI automation. So anything that would make, uh, my fellow AEs es life easy, it, it would be very, very good.

Mehmet: Now I want to a little bit like kind of shift gears and Sure. Talk about the broader ai, uh, thing and you know, you into the B2B business, uh, and SaaS offering. So if you, you know. Uh, couple of, I think it's been like now one month or so when Satya Ella said like, okay, SaaS is dead. Not, he didn't say it like this way, but what's your take?

Mehmet: You, you've, you've, this is your second, uh, company. Peter, what's your take on this? What's changing in the SaaS business? 

Peter: Yeah. Look, I will say it is a weird time to be, uh, writing software, um, [00:32:00] because every software company, by necessity now is an AI company. I. AI features have become so ubiquitous and so commoditized that if you are not including them in whatever product you're making, somebody else is.

Peter: So every, so let, let, let's, let's take that as a premise. Every SaaS company is an AI company. Whether or not it's a foundational AI company is, is a separate matter, but like everyone's gonna at least consume commercial models or open source models from open ai, from, from, uh, philanthropic. You know, or you know, from meta or whatever.

Peter: Um, even though, or because every company is at least partly an AI company, we're not only consuming these commercial foundational models or open source foundational models, we're also in some, uh, in some sense competing with them. Right? 'cause there is. No process that you can build around one of these [00:33:00] models that can't at some point in the future, be built into the model itself.

Peter: Um, the models are getting more and more capable, and you don't have to look that far in the past to see an example of how, uh, you know, sort of little niche industries have emerged. And then within a few months have been completely wiped out by a change in the foundational model itself. Like as an example.

Peter: When Chat GPT first came out, there was this little kind of, um, universe of companies that got built up to enable chat GPT to interact with PDF documents and, and, and with attachments and sort of ask questions of documents and understand documents. And then a few months later, well that capability was just built into chat GBT, because of course it was gonna be built into chat, GBT.

Peter: And so that whole industry just. Went away. Right. Um, and so I always feel like we're [00:34:00] kinda walking in front of this slow moving steamroller, um, where, you know, we're, we're able to stay ahead of it, but there's a chance it could speed up and just like roll us over and we could wake up one day and discover that, you know, actually chat, TBT, uh, just natively has the ability to, you know, complete an RFP.

Peter: Um. So, yeah, it's a, it's, it's an interesting and scary time to be in software. 

Mehmet: So finding what we call defensible mode, right? 

Peter: Yes. Much more concise way of saying it. Yes. Defensible modes are getting harder and harder to find, I think. 

Mehmet: Yeah, but see, sometimes, you know, I, I meet a lot of, of, uh, founders. Of course not.

Mehmet: I tell them, guys, I'm not challenging you, but like, yeah, let's, let's really think 

mm-hmm. 

Mehmet: Uh, if for example, OpenAI, Microsoft, Google would be interested in getting into this business. Um, because what I feel, if you are into very small, by the way, the people who are in some [00:35:00] niches, they are very safe. Yes.

Mehmet: Because. Like Google, Eric, like all, all of the other guys, they would not be interested in going, you know, getting these customers right. Oh sure. And I tell them like, don't forget, like also these guys, they get revenue from selling the API calls that you do. So it's a win-win situation for them unless you're doing something that like they would, they would consider it as an existence.

Mehmet: Uh. Threat to them, right? Yeah. Which I don't think much people, so, yeah. So here I tell them again, pivoting is key. Making sure like you, you have the vision, you study it well. Mm-hmm. Uh, and also I believe, you know, like speaking with customers as well. So some customers would prefer to deal with smaller companies, that they are doing extremely great job, not only on product perspective, but in support, uh, in, you know, after sales.

Mehmet: Because they say, you know what, like Microsoft would not care about us. We're just a small customer. So [00:36:00] I, I'm hearing this, but yeah, to your point, and one thing also, you know, uh, stopped me is you said like every company is becoming an AI company. It's, remind me when we used to say like, every company is becoming a SaaS company or a software company, right.

Mehmet: Somehow, which is uh, uh, which is very, very interesting. I would say now. Speaking again about, you know, being in startups and, and doing this. Um, so scalability today with the, in the AI age, how easy and or how more challenging it became in your opinion? 

Peter: Oh, uh, that's an interesting question. Um, I think it's, it's easier than it ever has been, right?

Peter: Like, um, and we actually see this, I think there are. The barriers to starting a company and to scaling it from an infrastructure perspective, right. The ability to serve large customers, just in terms of demand, [00:37:00] it's easier than it ever has been. I mean, the whole trend of the last, you know, 25 plus years in technology, you know, 30 years really going back to like the, the very beginnings of the internet, um, has been to make it easier for smaller and smaller companies.

Peter: To compete with larger and larger companies with less capital investment, right? Like, so first, you know, the internet enabled the delivery of software and the delivery of content, you know, through a free distribution channel. Very easy, scalable distribution channel. You still needed to run servers and you still needed to have, you know, physical infrastructure.

Peter: Then the cloud revolution meant that you didn't. No longer needed to build physical infrastructure in order to deliver your software and your services over the internet. Um, you could just, you know, rent a server for 30 to, you know, a hundred dollars a month from AWS or from Google. And so that further kind of reduced the barrier, uh, to starting a software company.

Peter: Building advanced [00:38:00] features and like, say, AI type features required a huge amount of data. So you needed a certain amount of scale in order to do that. Um, now with the democratization of AI and like these very, very cheap AI APIs, um, there's no barrier there either. You're able to leverage all of the advancements that you know, these.

Peter: Massive AI companies, uh, have, have produced, um, very, very cheaply, and it scales perfectly as long as you get your pricing right, right? Like, uh, as the demand for your product grows, hopefully your revenue is growing and you know you're paying more and it just kind of scales very naturally. Um, so yeah, it's, it's getting easier and easier to scale at the same time.

Peter: You know, those huge companies that are producing the foundational models, like they're the ones that control the data and they have this big advantage. Like, I'm never going to be able to match Salesforce's internal data set. Right? And so like, [00:39:00] you know, there are certain ways in which I will never be able to catch up to them.

Peter: Uh, and, and we're sort of relying on the continued existence of these APIs. 

Mehmet: Yeah, absolutely. Make, makes sense. Uh, so how this, um, you know, affects also these companies being able to find investment investors to, to support them. 

Peter: Hmm. Yeah. You know, that's, it's a really interesting time. I think. Um, I do a little bit of angel investing and, you know, I will say from my own perspective as an investor, like.

Peter: I'm not sure actually what I think about companies like mine, right? Like I'm being totally honest. Like, I mean, look, I'm very bullish on what we are going to accomplish in the near term. I don't know if, you know, we are going to, if there's a path for us to become like a Salesforce scale company, right?

Peter: Because you know that [00:40:00] that's, um. Uh, there's so many things that need to happen between here and there and, you know, unless we are going to tackle the problem of becoming a foundational AI company, um, which is a very, very different business from, you know, building business productivity software that uses third party APIs, right?

Peter: Like, I, I just, I don't know if it's a very uncertain and unstable moment in software, I think, and, and I don't know what the future holds, and I don't know who the winners are. It's, it's really, I mean, it makes it exciting. Um, but yeah, definitely scary. 

Mehmet: Yeah, absolutely. Like, uh, just today I was speaking with someone.

Mehmet: I said, look, uh, like maybe I. Until last year, although like we had all the hype, I could still kind of make some predictions. I can tell you like, okay, I'm seeing, you know, but now, you know, I, I sleep, I wake up, I see something completely new. Breaking the internet, people talking about it, starting to build.

Mehmet: Yeah. And [00:41:00] you know, I think our, I'm not sure if this is good or bad, honestly, but our attention span. From even, I mean, not as a, on the personal level, I'm not talking about the social media, the reels and all the thing uhhuh. 

I mean, 

Mehmet: I mean even from technology perspective and priorities. Yeah. I'm hearing this.

Mehmet: I'm hearing this from executives telling me, oh sure. It's going so fast. Like today, we sit with the board and we talk about that. After one week, we, we need to change because things are changing very fast. So really exciting times ahead of us. Um, Peter, final maybe words of wisdom, if you want to call it, some final thoughts and where people can get in touch.

Peter: Uh, ooh, boy, final words of wisdom. Um, I, I guess, um, you know, look, the, the.

Peter: I don't know. Look, I don't know what's gonna happen with ai. I don't know what's gonna happen with software more broadly. Uh, I can say this, that, you know, the lessons I have learned over the last, you know, nine years in [00:42:00] software, um, are, you know, they're, they're old. It goes back to the Lean startup. The same rules still apply.

Peter: Fail fast, fail often. Try things, keep failing, lean into the failure. Um, that's how you learn. And no matter what is changing around us, I, I believe that those, uh, those rules will continue to apply and, and that, that is the ultimate path to success in software. So, um, yeah, that's what I think where, 

Mehmet: where to get in touch, 

Peter: find me on LinkedIn.

Peter: Um, that's where I'm most active. Um, my company posts on LinkedIn all the time and, uh, yeah, I'm easy to find Peter Bonnie, it's not too many people with that name. Um. No worries. 

Mehmet: So I, I will make, I will make, uh, the audience life easy. So I gotta put the link in, in the show notes. If you're listening on, uh, your favorite podcasting app, if you're watching on YouTube, you'll see it in the description.

Mehmet: Peter, really, really, I enjoyed the discussion, like, uh, great insights and, you know, I love, you know. [00:43:00] The prediction at the end. Yeah, there is no prediction because it's exciting times for all of us. We're gonna see how, how things will move. Um, as I said, for the audience, they will find the links for the company and for your link profile on the, uh, on the show notes or in the description.

Mehmet: And this is how usually I end my episode. This is for the audience. If you just discovered our podcast by luck, thank you for passing by. I hope you enjoyed it. If you did, so, please subscribe and. Share it with your, your friends and colleagues, and if you are one of the people who keeps coming again and again, our loyal fans, thank you very much.

Mehmet: We are really making me, you know, very proud because, because of you, the podcast now for the second months in 2025, we are making it in the top 200 must listen, the Apple top charts for podcast in multiple countries simultaneously. This is, it wasn't happening last year and the year before because we were entering one country at a time and just.

Mehmet: Every now and then, for now, every week we are in a new country, which is make me [00:44:00] happy and of course can be cannot happen without your support and you know, all the, 

you 

Peter: know. 

Mehmet: Questions you come to me, you suggestions you come to me to do, and the enhancement you asked me to do. So thank you very much. And also thank you for making the CTO show as one of the top 40 most listened podcasts here in Dubai in the business category.

Mehmet: So thank you also for that, where I think now on the 15th, uh, uh, rank. So thank you again for all your support. And as I say, always stay tuned for a new episode very soon. Thank you. Bye-bye.