Peter Grant: The GTM Reset No One Is Prepared For

Foreign. Hello everyone and welcome to today's episode of the Revenue Insights podcast. We're excited to welcome Peter Graham, Chief revenue officer@you.com Peter is one of those rare revenue leaders who's had a. Who has repeatedly joined companies before the revenue engines existed and then built it from Siebel to Salesforce and C3AI and now you.com he's helped scale multiple enterprise software companies through hypergrowth and IPO. Today at U.com, peter is leading global go to market for enterprise AI search infrastructure, helping organizations deploy AI agents powered by real time data, internal knowledge and trusted outputs. If you want to understand what it really takes to build a billion dollar revenue engine in emerging categories like AI, this is the conversation. So without further ado, welcome Peter. Thank you. Great to be here. Wonderful introduction. So let's get started. Peter, perhaps you could start by giving us a little bit of background as to how you got to the role you're in at the moment. Okay. So I've always looked for shifts in the industry. So I've been looking up Mazda Civil, that was a client server coming from what was mainframe. I wasn't old enough to be in the mainframe space that guy but. And then went from Seaboard to Salesforce such as shifted utility computing which obviously became SaaS that we know today. Then went to C3AI which sent by AI and then remember the 30th, 2022, my world changed. That was the launch of Chap GPT. Right. When I spoke consultant at times of basically advising young CEOs on go to market strategy for VCs. Yep. I thought I need to jump back in, I need to be part of this. Yeah. And I was looking at the right way to go. And to be honest, at my age it was quite difficult to get open AI, Plexig or anybody like that to sort of listen to you. But Benny offered reached out to Richard Social when he's looking to make change from consumer to enterprise and he said look, you need some of the enterprise deal that he's been done it before. Carl Schachter reached out to me and said hey, do you want to do this gig? I jumped at it straight away. I looked at Richard's o' Brien's academic background and he realized you want to be with some really, really solid technical people or the good salesperson that aside them. But now that you build the foundations, very exciting. So you saw it coming and made a conscious decision that that was the next step for you. Yeah, exactly. 100% because it's a learning. Right. It's a transformation it's going to do. So you do these things because you want, you want to build something, you want to grow something, you want to leave a legacy and everything else. I thought this is my opportunity. But there's a lot of AI companies out there now. Right. So. And if you know not all of them are going to exist in three or five years. So how do you assess the an organization when you look at it to work out whether it's actually got longevity, especially considering if you're getting involved at an early stage. I think just before we answer that question, the point is when we joined Seaboard, we were making a market for clm. When we joined Salesforce, we had to make the utility computing market. So to get traction took a lot of freaking hard work. Right. You know, you try selling in 2003, your data in London is going to be stored in American server here in San Francisco. Random times the phone got put down. I thought we were crazy and mad. This is different this time. To your point, the opportunity is huge. It's a TAM that we've never ever, ever seen in our lifetime before. However, it's the most competitive time that we've ever seen in our lifetime before. Yes. So if you look at the my where they might follow the money, where it's going right now, it's 250 billion invested last year going to VAD million the Valley into gen AI type companies but it's fewer companies with bigger chips on it instead of going into them. Interesting. VCs predict that 99% of all gen AI companies are going to fail in in the next one to two years, not the next three to five years. Okay. I would say it's going to be a terrible attrition rate. So you do have to choose really widely, wisely, incaptively. What you're also seeing is the LLMs are literally destroying the front office. Yeah, yeah. Any application, anybody's a wrapper around an LLM. You should look really careful with the value proposition and what they're doing. So going back to your question, I think the four fundamentals I risk for is one is the people. Yeah. But it's a team sport at the end of the day. You need to have the right people, the right skills, you know, in that particular team that have an absolute run focusability to execute and have a proven track record of doing it. Yeah. Now a lot of them don't have that in this space right now because it's younger generation, which is fantastic. Right but they do need some older gray hairs to guide them thank goodness keep us employed. Right so by things people there that academically really strong and not just putting a rapper that really understand the fundamentals of this. Yeah so with Richard and Brian I put them on the Mount Rushmore of generative AI. It was a very very lucky to meet with them Richards recognize Evon Deli he is one of the third most subjected NLP research scientists in the world and it's Richard and his Brian's work they get probabilistic models embedding that were the vectorized word vectors and now we know it's basically you know what Nell and Mount to content creation changing and everything else that was part of their research. Okay so OpenAI redbat meant oh my God, we'd stop doing gaming. We should do ChatGPT2.0 y credit Brian for his paper the Decathlon Paper 2018 so look for people that that tuck the news. You're looking for rock stars that that really know what they're talking about. Yeah, exactly because I, I you know my success was through following people a lot smarter than me and see G it being one of them Phil Robinson being another one young Jeff Squire and everything else as well. These people ideals of European tech industry they've built you know distribution for American software companies. They're much smaller than me PhD in particle physics everything else as well. So I followed them first of all and I learned from them because the people the first sort of one yeah I never changed Number two is a market opportunity. Yeah. Now pretty obvious right? Going back to my point before about some people here the amount of people left salesforce in the early days of this company's not going to be successful. No one's going to restore their computers or your information here then they're not going to be an enterprise. Grace set her company up until last year's a 300 billion dollar business. The market built but so is the market. What direction is the market going and how big is the market? Right. What's the dynamics of it? How could people buy and thereby is there a real pain in that market. Most software companies fail. They run out of money obviously. Yeah but they run out money. So you get a product market fit because there isn't demand for that particular product yet. So you've got to really qualify now. So you've got to go and say okay here's your ICP type of ideal customer. Like go talk to the sales vp, the marketing vp, the cto CIO of the particular companies you're targeting when with where you think there is and hear that feedback makes all the difference and make sure you're intellectually honest with yourself that you're not having a happy ear song because you really want the job because that's what you thought to be nice people you said can I really really be successful. The third one's a product. Yeah. Now in the olden days if you had the most amazing people on the great market opportunity the product didn't matter so much because you knew it'd always catch up. You could bridge it with the relationships with CS course and everything else. I think it's different. Okay I I read I mean Betty off used to talk about the deserialization of technology with Riv like you know his original page for Salesforce is that the nine year old girl can order with dad's credit card a video on Amazon.com why can't enterprise software be simple? Yeah. And it never was. Right. Salesforce became what it where they set out to destroy the Siebo too complex on the user interface with this new generative AI retail maha space yeah you time to first token getting value from it getting answers at a highly accurate level. It's instantaneous. Yeah right. And you've got this shadow it that's going around that Everybody's using a ChatGPT Claude com LexDIA or whatever it may be and they're all finding productivity and everything else the speed is done at so to put into context ChatGPT got to 100 million users in 60 days. Wow. You took Google for ideas to do that. Yeah. It's 60 times faster. Yeah. For all now we know lots of good bouncing in ROI here but that's because they're point solutions until we start to stitch them all together. But who's going to win the desktop we're not going to be in between I think Anthropic Gemini maybe a bit of Microsoft we're see and obviously OpenAI but the product is so important now running these PLG motions as well. Yeah. You know it's not about giving up freedom it's about the product selling itself internally and having that instant value. So I think that's really important and that's why you see a lot of consumer people working enterprise space to get that moment really quickly and of course forefront timing right space there's never been a better time it's moving so fast but timing here is a speed of change. If you need to move that as a company is can there go the market because it is absolutely crazy so those would be my 4 foundation. Yeah, it's really interesting because if you find an opportunity where there are meet three of the four in players, it's very tempting. Just oh, it'll be fine, we'll work it out. Yeah. Right. But of course all you're doing is you're just investing time in something that's eventually going to go to zero. So it's very tempting but you have to be honest with yourself about these things. Yeah. So a lot of the enterprise B2B software companies that are coming across are really trying to work out what their moat is. And again it could be a custom data set and others don't got access to. Which again is a great way of protecting yourself. The software itself is no longer a moat. No, we saw that with Monday. Com and the speed in which people can recreate software now is only going to get faster. So talk to us about the most that you're seeing that are working in this space. Yeah. So we've had to do a number of pivots in the last five years. The company for two years I joined when they pivoted for literally a consumer based brand and we were doing a few million dollars in consumer. We were per nk. But, but the interesting thing is if you go back and is to our Playbooks, the SAS Playbook, the gold standard is triple, triple, double, double, double. Everybody knows on music you start to go to 0 to 1 but go to 3, go to 9, 18, 36, 72 IPO. Yeah, you need to do that my six months to a year in these days. Yeah. From a growth point of view. And then you say well how the hell do you do that and how do you differentiate yourself when you're going out? Right. So brand and distribution. Yeah, I think are really important. And then it comes down to actually as you say, your data source as well. And where do you fit on the new stack? Right, right. Because now you've got the hyperscalers. So you've got the Nvidias, the squad, the ship providers. Right. Yeah. You've got the airline providers, Sybil Top the better. Then you've got who owns a middleware, the process and everything else. And then you've got this app layer. Yeah. Which can freely be changed as human interface. So how we talk and engage with computers is going to be so different in the next 12 to 24 months. Right. So we were prompted. Right. That's getting wind tapped. Yeah, that's absolutely getting wind tapped out to your Pong Monday.com, sage/markdown GERX now getting marked down like a crowdstrike getting marked down, you know, the legal companies getting marked down, everything else as well. So that's going to change dramatically. So we are now focusing on building an index. So building the world's first index to AI agents so it can accurately efficiently answer any domain question. Okay, what does that mean? We believe that humans will have anything from 20 to 30 agents as productivity to help them in their job. So we believe this hybrid model. Yeah. Will be a lot more productive and everything else. We believe a lot of junior positions will be eliminated like analysts, Wayland John data gatherers and doing reports and everything else as well. Because all that can now be sort of automated. But in the enterprise space, unlike Kusuma, they need ground truth while yummy domain expertise. They need context, they need freshness, they need to be kept up to date. Right. So we suddenly went, well, we can supply that to everybody that's building agents. So we're building an index first of all. And the original web index is like Google, which is a big, the Big Daddy 400 billion pages and you've got B comes down here. Then you've got brave and you've got these little players that come along here as well. We want to build an index just for agents, not for humans. Right. And for humans this keyword search is designed for advertising. There's core ways you do for product, different location for specific website by where you know the product you want, wherever there'll be more you search and everything. That's not how agents search. They're not lazy like humans where they can fan the query, lots more information and then go a lot deeper. But most people those trades that the deputy might unit could be a legal expert or a healthcare expert or financial services expert. So as an example, if you announce Google, which I will say how many drugs are approved by the FDA in September 25th. Yeah. You will say oh, many drugs were approved. These are the top eight or 16 that were done. Right. That's not good enough. Right. If you go to the FDA website and you search for up to 484. Right. They're approved. I want to know every single drug and I want to scrape all the PDFs and all the submissions course. And if that's not a normal human, Nikki Suma searches for. Yes. And that's when we think there's going to be a big moment for you can differentiate so effectively a search engine specifically optimized for the agents. Exactly. Okay, I love that. That's fantastic. So one of the other big changes that we're seeing or the big shifts is a, is a move away from kind of traditional SAX based pricing and people linking there trying to really focus on. We started, you and I are old enough to remember when we were selling Capex right. So customers would come in and they spend a lot of money but they'd spend it one off. And then we moved to this SATS model where people would effectively were renting their software. But the beautiful thing was it was recurring and that's how we got all our valuations. Now we're seeing a big shift to try and align value to proxy. So how are you seeing that evolve? Because there's nowhere to hide there. Right. You can get customers signing up but if they're not actually using service, if they're not getting the value then there's no revenue there 100% so there's two ways you can do that. So when I joined UACOM we were see based pricing just like ChatGPT and everybody else. Well you $20 a month and $15 if you paid for the year. When we switched to enterprise I went around and just met as many customers as possible. The original CBO and the number one question I said what's your adoption rate Mark? And they measure it and go worth over 25,000 people. We've rolled out to equal 4%. So how did you find adoption rates? How many queries are they doing per week? Basically yeah. On average 4% were doing one query per week. Right. That's the lowest I saw. Yeah. Right. The average is about 12 or 15%. Okay. And the best I saw was one of the world's most famous consulting firms got 46 or have 46,000 consultants. With this 53% answer to widget VOC based pricing we have a massive Patricia problem after a year. So only half of the users were actually using the software that they were licensed. It wasn't a discipline like CRM in a roll now CBO and crap or lick the sales VP equity forecast and it doesn't exist. And see sorry Salesforce, it doesn't exist at all whatever it may be. So I said I think we need, we need to align behind the customer success. I think we need to go usage based pricing now we had the advantage we had and this is where organizations change is you have a very good analytics department where yeah we need to know all you use during your platform is with amount of 10 million monthly active users. So we could see what our power users were doing 150 queries a month. And what our lower end users were doing and everything else. So we basically say we worked it out. We basically have a bell Kirk distribution ambulance. Up to a thousand users of. Fine. What bucket of percentage of queries do you need? Right. And we'd start below to make sure you're going to be successful. And if you want to add to that bucket later. Economics lesson, sell first. Yeah, right. A lot of customers love that. Okay. The next iteration of that is obviously Outcome. Obviously where you say, did you close our travel ticket? It will cost you $4 before $5 before the human. This cost you 20 cents or dollar. Yeah. Whatever it may be. I've heard some pushback on that price here. I don't think that's quite right at the moment, but I think that's where the industry will go. Yeah, it's really interesting. I like the bucket idea because it helps the CFO kind of get to grips with this idea that we're going to be paying based on the, on the usage. But, but don't worry, you know, you're not going to get different bills every month. You know, assuming that you stay in this bucket, you know what you're in for. But I, I had a really interesting. The like the last podcast we did was with Vicky Revots for Insta and Intercom were really early in the AMA space, so everyone did them a disservice. But they were the way I consider Intercom historically in the old days was it was a chatbot on the website. Right. And obviously they adopted, they launched their first AI product three months after ChatGPT came out and talking to them about kind of the journey that they'd gone on, what they can. So for example, if you're this AI based chatbot now and you're on a bank's website, well, if someone's asking for the opening times of the bank, and that's one thing, and you might pay 20 cents for that because you know the value is relatively low and they probably could have searched for that themselves. But if you can connect to enough of these disparate systems around the bank and put all the authority in place where if I've broken my or lost my debit card, I can now use that chat experience to actually get my new bank card audit. Right now, all of a sudden that's probably worth $5. Right. Because now, because now we can take out a human process. And as a customer, as a consumer, I've been able to get what I need much, much quicker. So I'm now happier as well. But obviously that raises all sorts of challenges because you have to make sure that the person requesting the bank card is the person who you're going to send the bank card to and it makes sense. So really interesting to look at charging based on different use cases and the impact that use case has on the cost base. So I think there's some really interesting conversations happening at the moment about the different ways of commercializing the beef models but the one thing they all have in common is trying to leak of the pricing through to value. Yeah it's actually really complicated at the back end so that's why he's alluding to are you amazing. We're also leaders. Right. You need really good billing engineers as well. But yeah we understand this because we take a lot of LLMs so they all price differently and then we build our APIs on top of that and then based on the use cases you're referring to. So for example you go and do a web search that's 5 to $8 per cpm per cost per thousand beer was 8 cents. Right. To bring to us a web search question and answer. You know who's Guy Ruby on like tons of out forecasts blah blah blah. Yao Korean. That's. That's one query. Yeah yeah Some people price it differently based on what he's doing and everything else and they make you think it's cheaper than it actually is and get a lot more complicated. So I don't want to go that if you go to the other end where your salesperson do research and you use a deep restart photo that got Ari product sales reference research insight about five minutes will go and talk to up to a thousand websites. It would do chain of thought thinking it would do reasoning and everything else right. Go read the site. It will bring back a bunch of things important to you. It will then structure it into PDF it will create an index it create charts for you and everything else and it will output it to you. You now have this amazing research report that would take a top analyst four days a week to do whatever it may be. Now that's $25,000 per thousand which is $25 per report makes sense but it will crush the houses to do it. Tens of bets. Yes. Look and what scares people for me to outcome based prices they get worried then that's they think well price how many queries might we get to see you be doing this but control of the employee fixed overhead and the other issue you're dealing with is obviously the internal politics sector is voting voted for prisoners. Yes Course, yeah. And what you are seeing is a bit of change with CEOs that basically say double my revenue, not my haircut. Exactly right. And they're rewarding people eternally who embrace AI and make it part of what's going about before 20 to 30 agents. Right. In this high grid model where they work to together very chemistically. So we were just today launching our P2P benchmark report. So every year we do a report where we've analyzed, I mean this year I think we've analyzed nearly $50 billion worth of revenue across over 700 businesses. I suppose the conclusion I got to when I looked at the analysis was that we're seeing almost a tale of two AI strategies. So there's still a big chunk of the market that are using AI as a kind of a brute force tool. So in those organizations we saw the top of funnel activity that the volume of deals entering top of funnel increase in a single year by over 60% for day. Huge amounts of volume because they're using AI to generate lots and lots of leads. But there's no real thought behind them. There's no that it doesn't match ICP. And the average deal values are going. And so when we look at the volume of deals closing also increases. But the, the, the revenue per seller is going down dramatically dropped by 25%. Right. These organizations are having to employ more and more and more sellers because they've got all of this volume. But the efficiency of the sales function is dropping through the floor. This, this is the off CRM problem. I remember Tom Siebel's book I read 96 on a drone. Siebel and he likened it to the car industry decade before, two decades before in the 70s when robots came in to make cars. And what happened is GM spent billions on basically turning their factories into automated robots of a building and everything. But the quality of the RIM is still crap. Exactly. The reason why it was crap was because the deadboard was going in the past. It hadn't gone all the way back into the supply supply chain to change. That's all they have is put more rubbish through it more efficiently. Yes, exactly. And if you haven't locked down your go to market motion, if it's not, if you haven't got the rigor in place to keep people disciplined around the gates and triggers between the different stages as you're progressing through a sales cycle or making sure you're qualifying well and after and making sure that you're not going through a sales process of anyone who businesses that just really aren't naturally near scp. Then all you're doing is using AI to accelerate the failure or the cost. And all these inefficiencies just get broader. The businesses where we're seeing the opposite where they're using AI to maintain that rigor to focus on OSCP to ensure that every customer gets qualified correctly to make sure that they're multi threading with the right stakeholders at every stage of every sales cycle. Effectively using AI to introduce that level of consistency. So getting under the scale of what top performers do every day and making sure that we hold everybody else to the same bar to the same other account. And when those organizations we actually saw the volume of deals they closed last year drop by 8% but the revenue per seller massively increased because the average deal values went up. Their win rates increased and the focus on ICP meant that they were building a real valuable business. And so in a way while everything is changing actually nothing is changing when it comes to the consistent way that the fact that relationships still drive revenue and rigor and discipline around how you run your sales cycle does is still is more important to go than ever before. Yeah. I tell a wood challenge is the speed things are changing and what I mean about that the sales process. And I could talk from genitive AI point of view. So I'm in unfortunate position. Yeah. But ras once the I loved we were taking any revenue before we were very horizontal. Yeah. I didn't go to market by vertical. Okay. And that was good. But it becomes a bit of a customer suck issue. Customer service issue to service. All these people demand it and everything else as well. So I stopped that and submit minimum more is a hundred thousand dollars. Yeah. Oscar ends the E3 type. This is not my API. This is a very different type of business. But what I found is it changed that the seller's behavior. Right. Every seller here has to be an AI first seller by the way. So sellers tend to spend 23% of the time actually from the customers. Yeah. We're supposed to do BS and you know research and that I live to double or triple that. Yeah. With broadband aim sense. But also the first call is super important now go back to I said the beginning. We're not educating them up. You don't need to Are you it and why are we different? How could I make you successful? Yeah. Then I need to qualify. Why us? Why now? Yeah. What problem did I solve? You got to make you successful. If I can't, you're already gone. Yep. Because there's Enough opportunity to go to the next person. Cool. If I can though we see a sell side for six weeks for three or four hundred thousand dollars. Wow. Right. So, so it's. So that's where we're applying the AI to get to a very quick discovery much faster than you have done historically. And then, and then you have to control the evaluation the pov what success looks like when then fit your arms around that customer especially when spend that random money with you course. Right. And then you, you, you know very, very good to whether you're going to be successful less sellers. But you actually need more technical type sellers and more creative type sellers to do that. Well that's the change I'm seeing. And in your organization do the sellers continue to take up an active role after the dealers have signed? Yeah, because we just go back to consumption. Yeah we, we, we just want more and more consumption on more and more your business onto every single penny you've got put onto our platform. Of course. Right. So yes. It's true. And the sellers continue to hold the account forever or they way too early. Right. You're on that journey. Right. Yeah. Way too early at the moment. Right. And we do a lot of chopping, changing or you know that's what I go back to. I said you need an amazing rev opsy which I'm fortunate just to have on team but we have their web constantly analyzing what's happening. Constantly analyzing combustion threats bouncing out analyzing our pricing and everything and then making sure. Right. Fitting our orbit up go to market. So is it geography is it there to bore size of accounts we're going to target and everything else. Yeah. There's always a little microscope and you've got rev ops doing all of that now is it? Yeah, 100%. I used to do it. Yeah. And I've had so that goes. You've got bigger. Yeah. And then much better. See. Oh no I hear that, I hear that. And and revolve still reporting to you though in our organization they estranged they reporting to finance but yeah I don't care. They saw a dotted line into me. Yeah. Yeah. So he's my right hand person. Yes of course. Meet I didn't report directly to be read. It actually doesn't matter. Okay. And we're seeing the finance role taking such an active role now in, in you know in the past if you were selling something under a certain value you, you kind of your manager could sign it up. Right. Yeah. Now everything seems to be going through fast margin CRM is right. So it's so in our own SaaS world you know we. We model on 87% gross margin. Yeah where this world's 3040 points is way lower. And if you think time to value for SaaS companies is 1.1 years of the salesforce. Right. Gen AI is more two years plus. Wow. So that's why finance are getting more involved because of margin pressures. Yes. Or in this race of distribution brand I was talking about earlier that we make sure we don't go under we have enough money to survive and is there a. Is there a time frame after the deal signed that you need to be delivering value before you if you don't then then you might have a churn with this. With APIs it's so fast where you do the ebal sets and but the house that's I mean that's a good and a bad thing. It's so easy to get in. It's also easy to be swapped out so so you almost. It's almost instant for it. So how do you build a moat in that environment? How do you get them to consume whatever the index where the. Yeah so so our whole value proposition is around accuracy. Okay right so it's all about giving highly accurate answers right so perfect answer to your search in the enterprise it's a really hard problem to solve for that's why they aren't going web indexes that's why you need epidemic horse card there's some of the best engineers on the planet new organization and need a lot of mates but a web index swap but as as a as a layman customer how do you how do you explain that to me how do I understand that you're more accurate than a another so so there are. There are four main metrics where so one is accuracy you care about secondly is freshness so accuracy where so this is about actually first right. So I have 100 questions that will ask you know how many k bad right Sort of like the human will blurry better released it gold data set back down where the ram truth is. Yeah so there were independent benchmarks that would do this not new.com saying okay against our competitors right Frames and meta and everything else so we use those to ground ourselves some truth where is it okay and then where the customer you work he he or she will give you their own then questions like 50 or 40 questions they will give you and then we're one of those different use cases right so as there's a lot so accuracy accuracy number two is freshness okay so how up to date Is the information as well. Zari Burn news index for duck girl which is very unique. So we now unify that with a web index and that pieces freshness number three, this latency. And, and so sorry, on, on freshness, how fresh is it? I mean if a news article comes out yesterday, it'll be yesterday within five minutes. Yes. Wow. Okay, fantastic. And again, how does that compare to your peers? They don't have an easel mix. They don't do that at all. Okay, interesting. And how, how, but how up to date are so. Well, it depends on the use case. Right. So it all depends. Does it actually matter to be up to five minutes? Yeah, of course. Recency. Right. So if you're in, if you're in social media, which a lot of our customers are, we do a billion calls a month onto our web index now then they do care about the latest information as goals. If you're doing a bit of research and go to new eating, cook and alcohol, you want a full briefing. Yeah. You probably don't care what happened the last five years in the last day, the press releases, data, coins and everything else. So all depends on the use case. Okay. So we've got accuracy, we've got recency. Yeah. The damage is freshness. Then we have basically latency. Yes. Where I say how fast did I get the result? Okay, go again. If you're in a consumer based business, right. You care about sub 400 or 300 milliseconds. Whereas the reason why Google was so successful over Yahoo and everybody else is they suddenly change the way they measured everybody internally. They're engineers, they put it publicly. Let's get to 250 milliseconds. Right. And that changed search forever. Okay. And the rest is history. What you're on the Google. That's what we need to do when we're saving those type of use cases. So speed. Speed. Right. If you're, if you began to cook or normally take you four or five hours to bring your briefing together, you're going to get it in two or three seconds. It doesn't matter so much. Sure. Well, okay. And the final one is cluster service called cts. Okay. How much did it cost me to get that answer? What value, how variable is it? So you go back to a webinar, five to six dollars a charge per thousand. Yeah. Go back to doing analyst report. $25,000 per thousand. Yes. So all depend again on the use case and what the engineers want. And actually your customer in our space is different with the cto. Right. Research engineers People go in the AI labs and. And people like that as well. Always the hedge funds need to know certain information they've got to plug in. Yeah. And to be honest with you guys, I actually love the AQI industry because you don't have to influence all the people, the desktop and all this. You're selling to the cto, the cio, whatever information that they can just embed into their own applications. That makes a lot of sense. And that's the way a lot of businesses are going now. Rather than paying someone else a big SaaS fee, they're just using these tools to build their own. So let's talk a little bit about your ghost market function and what's your tech stack look like? So actually the Redbox leader said to me on Monday, how many solutions we have in the tech stack? And they said, but it's like 40 or something. Which, like. Wow. Yeah, I said, wow. Yeah. A lot of it I don't see, especially for a company of our sapis. So. But the main one is Salesforce. Okay. Yeah. But over on the top of Salesforce is gone. Yeah, yeah. And then we have a lot of stuff where we get data in, make sure it's correct. And you know what else we use? Connect the dots. Yes. Massive fan of that. Yeah, I agree. Synchronous is good make to mind, but is it a break company he's building there. So. Yeah. This kind of relationship. Yeah, it's basically put your board, everybody there and see your LinkedIn connections and you use your investors and you use them to make introductions. Yeah. That you may not have known before, but you're doing it at scale. Yeah. I looked at another company called orw, a similar model, and I think again, in the world that we're now living in, those relationships are so important. And I think organizations have got access to so many relationships they don't know about. Especially the larger the business, the larger the pool of relationships they have. And the challenge with Lithium as a tool is that those relationships are owned by individuals and we need to get them available to the whole business. And I think things like connect the dots and all are great tools to. Yeah, no, I totally agree. But the building. One thing that I will double down and is gone. And then put a benchlaud over the top of every L. Because maybe you've read Andy Grove's book. Not only the paragraphs I'll think about the moment, but in there, Tom C. Will used to get us to write a report every single Sunday night about the business. What we did Last week. Yep. What are you going to do this week? There's a bang line. We do it benu.com he said this is a good discipline to have, but you fully understand your business, what's happening and everything else as well. I've now got Claude doing that with granola. So it looks at granola. I think it's recorded, observable, all that sort of stuff. Then from Gong, I get my Rev Ops team, they give me a full GONG report. Right. So every conversation has happened across the business. Yeah, yeah. And everything links to my V2 mom, to my okrs, everything's linked in for and it gives a beautiful output with a red, amber, green of what we need to do. Right. And then what my focus is, what Mrs. Worth for the week, what the successes were for the week. Well, my focuses are for next week and everything else as well. And then all the execs do that and we share it in the channel and we're looking to aggregate that together. So all aligned on where the focus is, the company and where we're going. But having that GONG over the top. Let's see. Derek, I think is working for a long week. Yeah, massive. And it's not just for sales Guy, it's for the Pride team as well. Yeah. Because they're hearing all the objections the sellers are getting, hearing more compressors are being mentioned. Yeah, it's really, really. Yeah. Not one of the things that we do when we work with a customer is we can analyze the last, say, thousand hours of call recordings or the last year's worth of call kills from a gone. And rather than allowing the sellers to tell us how well they qualified the customer or how well they dealt with objections, we'll go through all of that historical traffic to try and use the AI to score it. And you can see it when you do that for a year's worth of content, you can start to see that individual sellers are getting better at qualifying or not as good or dealing with objections coming up and so on. And it's amazing when you see that volume of data on a graph and you can see which sellers are getting more powerful and. And it's not coincidence that those that are getting better at qualifying, getting better at dealing with objections are the ones that their win rates are getting better and the level of coverage they need to hit that quota is reducing. Yeah, but. And again, it gives us as leaders that the tools and the insights we need to be able to support those sellers and help them to grow faster. 100%. So, Peter, I can't believe we've managed to get through so quickly. If if anyone wants to get in touch with you how would they go about it? Yeah just knock me up on LinkedIn. Very good and if they mention the podcast you'll accept the invite right 100 yeah yeah but that's actually on there right when people just reach out to you with no reason they can't be asked around why they want to connect I just want to upset with Graver Abbott I've already got four and a half thousand of my closest friends there on LinkedIn but really you know if you want if you are happy to connect the connection exchange or whatever right but 19 members are going to be trying to sell you son yeah of course and yeah if you take the time to write and say why please yeah no I agree with you absolutely well Peter thank you so much for joining us today and yeah looking forward to watching the progress of bu.com awesome guy I think thank you.