Transcript
Speaker: Data Mesh has been sort of around six, seven years, so pre-LLMs and the wider producer LLMs. You know, people that were building Data Mesh and weren't doing it necessarily 100% perfectly, then are potentially now struggling with how they adopt AI because their data is dislocated, or some of them, you know, their data is disconnected. the data sits within the businesses and then you you have AI within each department but you don't have that cross-reaching AI that is where it's at its most valuable where it can connect the dots across all the different teams. but Today's episode is brought to you by Omni.
Speaker: Most companies I speak to want AI analytics but are failing to put projects into production. That's where Omni is different. It's the semantic brain that grounds AI in the heart of your business logic, giving you governed answers, whilst also the depth to identify root causes. It's intelligence everywhere, from your spreadsheets to their chat feature. Even within your product, Omni moves you beyond the dashboard.
Speaker: Don't just take my word for it. Trust teams like Perplexity and Synthesia that are already using Omni to deliver intelligence that people trust. Check out them in the show notes or visit omni.co. That's O-M-N-I.co.
Speaker: Now, back to the show. Should your data team be centralized, decentralized, hub and smoke? It's a question that's divided the data community for for many years.
Speaker: and how does that actually change in the AI agentic era? Today's guest is going to be sharing their view on this approach and some of the AI projects and impact that they're driving.
Speaker: Welcome to the Stacked Data podcast, bringing you honest conversations with professionals driving the biggest innovation in data and AI. I'm joined by Paul Kelly, the VP of data at Booksy.
Speaker: And Paul thinks the industry obsession with data mesh maybe went a little too far and hasn't set folks up for success in today's era.
Speaker: Paul, it's great to have you on the show today. are you doing? Yeah, pleasure to be here, Harry. This one's been a long time coming. I know we've been struggling to get diaries together, so I'm delighted that we're finally finally able to meet after our future.
Speaker: Yeah, no, it's ah it's been a... I'm glad were we're we're here and excited for the conversation. to say I think you've got some really interesting insights on how data teams have been structured in the past and and how they should be um try get the most out of the AI era that we're firmly in at the moment. But um before we jump into that, um give us a little bit of context to you to yourself and your background.
Speaker: Yeah, sure. So, ah yeah I've been, I've been, i was in the army for for quite a while. And in the army, I got into, a yeah data, building databases, and things like that.
Speaker: um And um I left the army, sort of 2015, or 2014, and went and did an MBA, which was fantastic, really good opportunity to move between the military and the civilian environments without causing too much trouble. and And there I did a lot of stuff on on statistics and econometrics and and and you know future innovation and strategy and AI. It was a really good bedrock to set me up for civilian life.
Speaker: And then I got my first job at Barclays, the bank, and that was a as an interesting experience and sort of, you know, submersion into the corporate world. I did a couple of jobs and found myself ah working in their compliance team doing reporting. And, you know, boss said to me, how can we make these reports better? enough I went to you know build a database back. This was back when we did Hadoop and open all that sort of stuff. So we built we should build all of the Hadoop clusters to connect all of the data across the bank for compliance reporting purposes. And that got me a job doing
Speaker: cyber, was a bit of a cyber background in in the military, and that, you know, bringing those two things together, social cyber analytics and um data, data provision, data preparation to enable analytics.
Speaker: And then got promoted and moved to, you know, front office, which was, it was a big move to go from sort of back office to front office and work with the, you know, Barclay car team um and did that for a couple of years, which was brilliantly successful using a lot of the data preparation, data management stuff, really just to, just to coherence and organize the team to generate and it did generate, generate you know, improvements cash.
Speaker: um And then i was lucky enough ah yeah to move to Intuit and that was my first experience in technology. um Well, a technology company, it was a big big difference to Barclays, various jobs, but I've been there for four or five years.
Speaker: And Intuit was a fantastic experience, great company, really opened my eyes to the technology world and the technology industry and technology generally and I loved it and worked for QuickBooks for a couple of years and then over to Trustpilot which was my first real scale up, you know, Trustpilot was a company that was on the up and it was scaling, but there was, you still felt it was a startup in places and not in others. And, and you know, that was the sort of my experience in reforming and really driving change, not just incremental change or change over a long period of time, but quick and decisive change to to move needles.
Speaker: and then And then, yeah, recently about a year and a bit ago now, I moved to Booksy. So again, slightly less mature than Trustpilothood IPO. So this pre-IPO now with Booksy. And you know getting them um getting them into you know data fit. It's a good data environment already, but getting them data fit for the AI environment.
Speaker: Excellent. Who who are Booksy? Booksy is a great company. They're Polish. They started Polish but now they're they're global.
Speaker: And we've got offices in way Poland, Spain, France, UK and the US. And Booksy is a ah scheduling software, so you can schedule, you know, barbers use it for scheduling haircuts and things like that. You know, masseuses use it scheduling they're at their clients appointments. And we have a marketplace, so we have the consumer side and we have a provider side. And so we try and you meet those in the middle.
Speaker: um And yeah, and Anybody who needs to and you know schedule an appointment on there you know theyre whether they're like a mobile hairdresser, whether they're physiotherapist, whatever it is, they can use Booksy to manage their clients.
Speaker: Excellent. So Paul, getting into the meat of the the episode, um it's all around what's the operating model of ah of a data team?
Speaker: Why do you think the industry went too far on data mesh? Yeah, well, it's evolved. It's evolved. And, you know, the faster you can, you know everybody loves dashboards, right? So, you know, they they all go back to the dashboard. The faster you can deliver that, the better.
Speaker: And so there isn't there is an, attract if you look at the evolution of it, there is an attraction to data mesh, which is to put the, to put the analysts in the teams and then there's the the they're right with the run by the business they're owned by the business and uh the business who own that data already um can then use uh the dashboards and the data immediately there is there's no process there's no jira tickets there's no you know getting prioritized and waiting so there's no there's a there's an attraction and that actually you know speeds particularly scale ups and and startup speeds you up you don't need a heavy centralized team to uh to do that i think the problem if you're trading off is governance and structure potentially now there'll be people that are data mesh uh aficionados who and i agree with them there is it's a really good method but you must have
Speaker: almost perfect governance. You must be a very, very disciplined organization. And what's happened, I think, is um the attraction of data mesh for speed and pace has meant that in certain cases, you know, corners have been cut.
Speaker: And, you know, people have done their own thing within those teams, where the governance, the central governance wasn't potentially there, which means you get metrics that don't quite add up and both are called revenue, but they're calculated in different ways. And you, you then try to understand your you spend most of your time trying to understand exactly what you mean by revenue, and and particularly have different business lines, and you have different products, whatever revenue, if it's if you don't have really good classification in your in your warehouse, and in your in your, you BigQuery or whatever, then you're going to really struggle to understand what these things mean from a central perspective.
Speaker: And why is governance so important, I suppose, to data, but particularly in this AI era? yeah Yeah, it's always been important. It's always, always been important, but it's just getting more and more important.
Speaker: I don't want to bang on forever about governance because I know, you know, everybody knows that it's important and that it ah nobody will ever say I don't bother with governance but then it's it's but it's actually getting the effort to put yeah put people into governance when they could be on revenue generating activity and so there's always an attraction to just do the revenue with generating activity and i think one of the things I found sort of you know books he's very keen to um
Speaker: grasp AI with hands and really, really get every sort of benefit you can out of AI, both for our consumers who are getting haircuts or whatever the to the the providers who are providing them to us, for our employees, to make their their employee experience better and easier and faster.
Speaker: So how do you do that? Well, I think, um yeah, I'm quite lucky i and I inherited when I joined a a fairly good data environment and and we've been able to build on that, um which is great. But I think there's other places where that have fully embraced yeah something like Data Mesh where they're now trying to build AI tools ai tools and they're trying to get the AI to go and read their data to be able to do things like, you know, LLMs or whatever and retrieve that information to do the RAC and retrieve that information and give it to ah to a customer or or an employee or whatever it is.
Speaker: And if you've got seven different versions of revenue and 15 different versions of what churn is, the agents are really going to struggle to to make sense of that. So then what you have to do is isolate all of that down and are really narrowing you your ah you know your data pool that the the agents can learn off of because you're having to strip it back to to provide those answers.
Speaker: So you know the the issue therefore, when you you can link that back to data mesh where data mesh was not put in perfectly. People were yeah were, you know, where people were kind of corners and they but had eight analysts in the teams and all of the different metrics and the different ways of calculating stuff was was done, not centralized, but decentralized, the the problem arose within trying to deliver agents on top of that because that data is you know trying to join those. The power of AI, the power of this stuff is to connect the dots across different organizations and teams. You want to know, for example, you want AI to tell you
Speaker: if if If you've got a dispute with a customer, don't send them a marketing email. right that That sort of level of connectivity between your simple, simple connectivity between different teams. Or if you're the customer support team, you want to put a an AI agent in to support the customer support team. It needs to be able to make changes in the product or it needs to be able you know retrieve information on the customer and fix it. So if if you can't do that, because your customer support data and your product data doesn't quite align, or it's calculated in different ways, the agents are going to struggle to to solve those problems.
Speaker: Amazing. So how have you actually set up the operating model in Booksy to enable these AI projects and the governance and the the impact that you're hoping for?
Speaker: Yeah, we're we're very much, um the way the way we've done, so we're a central team. but we're embedded in the business. So we we want the best of both worlds, right? This is probably the best model.
Speaker: You we've really, and I have iterated on this over the years of different companies in different places. And I think this sort of is is the best method of doing it for now anyway, which is very much this, you know,
Speaker: centrally managed but but embedded in the business and and the way I sort of phrase it is is is the who and the how is defined by me if they're in the data sphere so who they are, what their qualifications are, how good they are at their job, who selects them to do that you know all of that sort of stuff that human recruiting and standards that that that sits with me on my team and we make those decisions So that's the who, the how, you know, that is the methodology. Do they, um you know, how how they create a algorithm, how they do, how they, you know, build a lead school, how they govern GCP, whatever it is, right they those things.
Speaker: That all sits with me. um Then the why and the what, that sits with the business. So what are they doing on dayto day to day? What is that task? yeah Whether it's yeah marketing, personalization, segmentation, whatever whatever all these things are. um the the the sort of task list and the prioritization and the what they work on at any given time, that that sits with the business stakeholder and the owner. And as long as as long as the task is ah being done to the standards that I want by the right person, ah then yeah know they they have um they have the remit to do that.
Speaker: And i think that works quite well. Now, sometimes it can be, um you know, i get involved if people, you know, there's, and and people always have opinions, this isn't the right thing to do, this is not the right, you know, I, that's kind of where I step in and sort of look at the problem and then discuss it with, know, senior stakeholders, but generally these things aren't that fairly quickly, you know, ultimately,
Speaker: we all want the business to succeed and, you know, doing well. we all We all want the same thing. It's just how do we how do we get there? And so that's really the the the structural issue with sort of the Hutt & Spoke model.
Speaker: What are some of the biggest challenges you've seen? you You mentioned you iterated on this a few times. What are the the biggest challenges that you see with this model and how have you overcome them? Yeah.
Speaker: We iterate it little bit. I think we you know fully centralised is very slow. If you centralise everything and you just say, no, it all's coming through. uss here's Here's a Jira ticket, no ticket, you know no no pity, no chitty. So then um but're Sorry, no chitty, no pity sort of thing. You know, you don't have to take it. We're not interested. When you, when you sort of go down that route, i understand why people do and it's actually drives discipline and it drives prioritization or so it does slow the business down. If, if we want to, um,
Speaker: If we want to move fast on an issue, that sort of sitting and having people sit in a backlog somewhere is is it's difficult. Now, we still have backlogs because the team also have been reporting to me and we we work through the prioritisations. But the prioritisation, the important difference is prioritisation isn't really done by me. I contribute, but I'm not the driver or the approver. That stuff is done by the business. So they're the ones picking the priorities.
Speaker: what do they work on first? They have a finite resource that they can use. I just make sure that they're doing in the right way and that it's, ah you know, it's done by the right people and they're not being, you know, they're being used in the right way.
Speaker: um I think if you fully decentralise, it's what we've just spoken about. And it it has, you know, that will have a big impact on your ability if you haven't done it correctly. And bearing in mind when we were talking data, data mesh has been sort of around six, seven years. So pre LLMs and what the words produce at that end. So we didn't know people that were building data mesh and weren't doing it necessarily 100% perfectly.
Speaker: then are potentially now struggling with how they adopt AI because their data is dislocated, or some of them, you know, their data is disconnected, their data sits within the businesses and then you you have ai within each each department, but you don't have that cross-reaching AI that is where it's looking but at its most valuable, where it can connect the dots across all all of the different teams. You know, a customer only wants one touch point, really.
Speaker: You know, if you think about, you know, chat and having a conversation with an agent in Booksy, it kind of only wants to do that with one touch point. doesn't, you know, it doesn't want an agent for this and an agent for that and an agent for that.
Speaker: So yeah, so that problem wise, that's the sort of problem with centralization and versus decentralization is why we've tried to you know, ride both horses a little bit and give people, maybe give ourselves... The flexibility. Yeah, clearance, governance, but the flexibility for the business to to execute.
Speaker: Amazing. So tell us about some of the projects that you guys have have worked on. I know we've touched on ah on a few. what What has been the most impactful project that you and your team of Booksy have worked on? Well, there's loads. So there's some product ones which which are coming up, which I'm really excited about. So there's assistance that we're going to use for customers. So the customers will be able to, the product teams are delivering on, that they're going to be able to, people will be able to book via free text and all rest of it, which is exciting coming up.
Speaker: um But ah for me, the one that the most impactful, I think we have built an agent that is a customer support agent. And um we, i you know, I think with these things,
Speaker: I think with these with with some with these things, the the thing is, is to, we call it, you know deflection and we say, oh, it deflects this. and In other words, it doesn't it doesn't necessarily go to a human. But the point is, is trying to take pressure off, um,
Speaker: the existing customer support team. I think with the agent we've built, it genuinely is useful. it's like it's It's not about trying to remove backlog from the CS team, although that's a benefit. I think this thing is genuinely useful to customers. their Customers use it and get the answers they need simply and quickly.
Speaker: rather than waiting, yeah trying to phone in or send an email in or or wait on chats for ages until a live agent turns up. They can get the answers they need quickly and simply and then carry on with their day.
Speaker: We've seen like we've doubled the the success rate. So but putting this in with double the success rate, we've seen like thousands with you know thousands and thousands of checks that are being being handled and and successfully closed by the agent that it that that has taken that is the solving you know customer problems. And we've seen you chat that that volume reduce because they're having their problems, which is great.
Speaker: Exactly what you want to do. So I'm very excited about it. We've got more in the pipeline. looking at you know supporting prospects as well, making sure that they're making the right decisions. Again, it's all about being customer centric. For me, it's not about how can we take the pressure off you know the sales team or take the pressure off the customer support teams. how What are the customer problems? what is What is the customer problem? How can we solve it? How can we make it better using agentic tools?
Speaker: How can we remove those frictions so that they get more value from the product, more value from Booksy Marketplace or whatever it is? How can they understand all of the different options that might suit them? Those the sorts of things that we're we're looking at and that's how we're using agents at Booksy to support customers.
Speaker: Fascinating. I think, yeah, the the impact on the customer journey and and clearly customer support sounds sounds sounds like it's it's not just increased your operational efficiency, but also the impact on how your customers actually perceive Booksy. How have you done that technically? whilst the Can you share some of the, obviously not all the secret sauce, but ah what are some of the challenges that you and your team face when implementing and building out these these agents?
Speaker: Yeah, i think um I think, I think, you know, it goes as said, guardrails are important, right? Making sure you've got the the right guardrails. I think there is, you know, with agile, there's a lot of, um you know, let's get the skateboard, yeah get get it up and running and stuff. Sometimes that's really useful if it's on small testing environment or a small pool of ah of customers. I think it'd be quite difficult to do like large scale.
Speaker: um particularly if it is a skateboard because you don't want ah you don't want to pollute people's minds for the future like as you as you refine it going on. um so we spent a lot of time we spent a lot of time on um ah studying the old chats from our old system and seeing where the friction points were and understanding you know where the value added.
Speaker: um and making sure they were solved. We spent a lot of time, knowledge base is incredibly important, right? Building the knowledge base so that so that the agent has a really good solid foundation of of data. And it might be that it has, you know, it pulls from, this is and this is my point around you know different data points and having your data, you know, be managed from, you know, centrally so that it's, so that it's retrievable across all of the different teams.
Speaker: But if you've then got three or four knowledge bases, one in marketing, one in sales, whatever, and you've've you've got Argentic in the center of that, and it's and it's pulling data from all of those different places, and then bundling that up as a utility for an employee and yeah or ah consider a customer or whatever, then um that that is just' incredibly powerful.
Speaker: that's incorrect lus really It's really, really useful. the other thing we've done um which is interesting is is it pre pre-populate like things like emails and chats and all the rest of it so it reads the chat or it reads uh you know what's happened and if it is escalated to an age uh to a to a rep then all of that stuff is ready for them you know the whole thing is pretty popular they can read it all they can get up to Whereas before they'd have to sit and read through the whole chat. There's now a summary, it's all summarized and you right, okay, I know what we are, right, know what we need to do. Solve the customer problem. and It just all brings the the time the customer has to sit around waiting down so that it feels a better experience. But yeah, to do that is knowledge basis. It's guard good guard really good guardrails, really clear, tested guardrails.
Speaker: And then make sure that you are using agents that learn. So you you want to have that unsupervised learning. They need to be watching what reps do and learn from the, you know, learn from how the rep handles good reps, your best reps handle certain situations, and then play back that into chats in the future. So that it's a... It's CICD. Yeah, exactly.
Speaker: It's exactly what it is, yeah. Amazing. and um yeah You touched on a few things, but like what's the the vision for for you guys moving ahead? What are you most excited about on your roadmap and what capabilities are you going to be looking to to build out?
Speaker: Yeah, I think um um what I'm really excited about some of the, you know, some of the product features of Booksy have got coming out is really they're really innovative. And, you know, the product team and and and the CPO have come up with some so really innovative, I'm not 100% sure how much I'm allowed to sort of, you know, say pre pre launch and the rest of it. But they're exciting, right? they're They're really, and the most important thing is that they're really useful to customers, like their customers go, this is really helpful.
Speaker: um and And, you know, that's the starting point for for all of it, but, you know, AI receptionists and all of that sort of stuff. So, you know, that you know that that I'm really excited about seeing that go live in the world. I think, um I think I'm personally excited about ah having you know agents that can triage prospects. So I think that's a really interesting space. So there's you know a myriad of different reasons why someone might consider
Speaker: um a using Booksy and they all, you know, some are sole traders, some are mobile, some of them, they don't have an office, some have, you know, some have five, 10, 15 locations.
Speaker: and locations and they've got 50, 60 members of staff. you know These very, very, very different businesses. They're all in different categories. yeah As I said, barbers, hair, massage, nails. You've got all of these different... It's it's it's very kaleidoscopic.
Speaker: And so having... ah building an ability to sort of triage all of that and really, really extract the need from the customer. What is the customer need? it Really extract that need so that then either the product or the sales team or whoever can can really configure Booksy, the product, the the offering.
Speaker: to suit them, to really personalize it, I think is a really interesting challenge for us. um And the same for marketing, right? The marketing challenge of, you know, moving away from, you know, mass segmentation, but to hyper-personalization. So Harry, what does Harry want Booksy for?
Speaker: You know, get really down to the super granular. Why is it not, not why does it like companies like Harry, but why does Harry need Booksy and have that really bespoke conversation that immediately when the customer needs it or the prospect needs it because, you know, two thirds of, you know, the US is a big become country, it has a lot of different time zones. yeah same You can't have people up all night, you know, humans have an hour shelf life. And so having an agent running 24 seven that can answer these questions with real clarity and be consistent about it is very advantageous, not just for Booksy, but for the people that are asking the question about, is Booksy going to work for me?
Speaker: and um And so, That I think is really powerful. I think that's really powerful. Again, it's yeah I can go back to the customer's problem, you know, solving the customer problem. The customer problem is I want to talk about Booksy when I want to talk about it, not when you're available. So creating the capability to have that conversation with with an agent is, I think, really fascinating.
Speaker: And how you do that, um you know, the yeah how you do that with... with AI and how you make that human like, you know, the Turing test and all the rest of it, you know, all of that stuff. I find fascinating. I think i'm really excited about doing that over the next months and years.
Speaker: Amazing. of Looking forward to seeing that unfold.
Speaker: Paul, before I let you go, um what do you look for in in data professionals, people that would be joining your team? um I imagine it's changed over the years.
Speaker: What are the skills that you look for in a data professional? Yeah, I mean, it really has changed over the years, isn't it? I mean, the data scientists from from years ago, i mean, it's it's it's the last couple of years has completely changed.
Speaker: um What we're looking for now, like, used to be people, engineers, data scientists, defined themselves by the code that they could write, the code that, you know, their knowledge of it, and ah how they solve problems using using code, whether that was Python or SQL or Scala or anything else.
Speaker: I think now it you know, generic code, i.e code from, you know, Claude or Gemini,
Speaker: um is is so abundantly available that and you yeah have people vibe coding stuff. Now, I'm still suspicious, but maybe in a couple of years, but I think i think there's a lot of tech debt being built up by vibe coding that will one day become apparent.
Speaker: um But hopefully not booksy, but I think there's definitely um that sort of, you know, that that need for being able to use code and understand code and that that threshold has has gone down because because of LLMs.
Speaker: And so i think it goes back to what what what is the purpose of an engineer? And an engineer historically has always been someone who solves problems.
Speaker: the The nature of an engineer is to solve a problem. that The character of the engineer over the last 15, 20 years has been to write code.
Speaker: Well, certainly data and and those sorts paste software. But the nature of an engineer is to solve problems. And so the LLMs and the code and all of that, that that is a support to the engineer who is in the ring solving the problem.
Speaker: And they're the people I look for. I look for people with really high level problem solving acumen. And we test that, right? We use code to test that, but we we when we write the you know case studies and we write the you know the code tests, we do that with the knowledge that people will use um a a andm So we do, we do it in really novel ways to to test that problem solving capability.
Speaker: And I think the other one is, is to, I think there's, you know, the, the, the engineers that sort of, say no until they say yes in terms of, your I need to know everything and I'm not going to do anything until I really understand every single bit of this process or whatever it is, whatever it this problem that you're asking me and I'm going to spend ages asking question after question after question.
Speaker: versus the engineer or the data scientist who is solution first, which is, okay, I kind of get the problem. How do we solve it? How do I fix this?
Speaker: how what are the What are the options? How do i run these things and work the room, work the network to really meet all of the different teams that are involved and and and get to a position where they're yeah, I've solved this problem. Let's stick it on an architecture board and let's see if it let's see if it works, let's get it live.
Speaker: know And then let's get it into testing mindset, I think is really critical. yeah It might not work, but let's get it into testing and find out. let let's Let's get it, you know, let's not you know, spend the ages in the in the sort of design and and and the, you know, questions. let' let's let's Let's move forward. And so, that you know, that that I think is a really important skill.
Speaker: um And again, none of this is is is, I'm not talking about code. I'm not talking about Python or whatever. I'm talking about mindset. And the mindset is the most important thing.
Speaker: It's can do people. People are actually going to put this feed up and go for it and really problem solve
Speaker: his name now, there but the engineer who built you know the Clifton Bridge and in Bristol, the guy, Isabel, I'm not going to say it, but a he dug a tunnel under, say again?
Speaker: Was it Thomas Telford? Bernard. Bernard. And he built he built the tunnel. That's an engineer. He's an engineer. He solved the problem. The problem was how do you get how do you build a tunnel under and under the Thames?
Speaker: yeah That's an engineering problem. and and and these and sorry It didn't have JCB's away. It's about the nature of an engineer and really understanding that the nature of an engineer is to solve problems.
Speaker: I think that's a very pointy answer and and what we've been seeing across the the industry, that ability to solve a problem, gather a context, understand a problem and then come up with a variety of solutions is, yeah, the solutions. Yeah, I think it's a really good point, Aaron.
Speaker: You know, the lateral thinking bit, you know, you know, the lateral and logical thinking, it's a logical thinking, you take a problem, you come up with a solution, whereas the lateral is you just keep going, you just go, right, that's one solution, I'm going to do another one, now I'm going to do another one, now I'm going to do another one. How many times, different ways can solve this? And then you sort of look you know, six or seven solutions to it.
Speaker: And then you narrow down to the best one. You know, they're the they're people like that, they're really are they really inspire me. I love working with people like that. It's a joy when when you're with someone that can do that.
Speaker: Yeah, as say, I think problem solving is is the biggest skill that a lot of our our clients are are looking for because the execution has become the the easier part of that that equation.
Speaker: Fascinating insight, Paul. um Final question. You've had a fascinating career career. Looking back across your whole career, what's one thing now that you you know that that you wish you'd known sooner?
Speaker: Oh, crikey.
Speaker: Crikey.
Speaker: It's really difficult. it's It's really difficult to do this, but it's a skill that um I'm getting better at. But I think the one piece of advice would be to stay ahead of the curve.
Speaker: the the the The thing, do not get caught in a trap of defining yourself. Going back to defining yourself by code. i i do Python. ah dead it the the The industry is so fast moving and getting faster that you have to jump with two feet into the next innovation.
Speaker: yeah And and yeah if we think yeah five years time, we've got quantum coming. Now quantum is not readily available today, Google's working on it. You've got Rigetti working on it. There's loads of quantum, IonQ. There's loads of companies out there that are building quantum computers.
Speaker: and they will crack they will crack the that you know um the um that problem space and when you take something like quantum and you take cyber security and you take ai you know that those sorts of things that's coming together to be a very interesting space in the future and You know, my my my one piece of advice that I would give to my pre-self is to always make sure you're thinking forward and you're staying ahead of the curve.
Speaker: Don't define yourself by the current state. Don't define yourself and say, I do this or I do that or this is... this is You know, this this is my niche because I guarantee you that niche will one day disappear.
Speaker: And so you need to keep moving forward. You need to keep swimming because if you don't keep swimming, you'll sink. Moore's Law. Moore's Law. Moore's Law. and The greatest innovators and companies I think have always come from the ones that have looked ahead of the curve. um Right, yeah, your Microsofts of the world. they they They knew what was coming. You know, NVIDIA's obviously huge tech companies, but I think they, you know, them inspiring, know,
Speaker: global leaders were thinking ahead of the curve. So I think that's that's really poignant advice and great to help people sort of step back and try and philosophically think, where where where where is this going? how can i How can I make sure I'm in the right position?
Speaker: And and let's let's, you know, with NVIDIA, it's a really good example because NVIDIA, if we remember about, know, four years ago, five years ago, Game console.
Speaker: They weren't, you know, but they'd been around for years. Like Nvidia was in my laptop, you know, Nvidia GPU was in my laptop from from like 2000 or whatever. They'd been around for years. They were there were very, very, and suddenly, suddenly they saw the they saw the future and they they took off and they're, you know, the most valuable company in the world. So so it is it really is to, what it's about it's about seeing the future and then going and grasping it.
Speaker: Amazing. Paul, it's been an absolute pleasure to have you on the show. Thank you for all of your insight. I really enjoyed it. And you, Harry. Pleasure to finally get on your podcast. Thank you for inviting me.
Speaker: Amazing. And look if you want to connect with Paul, and we put a link to his LinkedIn in the in the show notes. I'm sure they'll be hiring as well so please do reach out to Paul and the team but for this week that's everything thank you everyone see you in a couple of weeks hi everyone just a quick one from me if you've enjoyed today's episode I'd be so grateful if you could hit that follow button or leave us a rating even better ask the show on to a friend who might also get some insight from it it really helps us grow the community and continue to share amazing conversations
Speaker: I also wanted to take a minute to talk to you about Cognify. Those of you that don't know, Cognify is the leading recruitment partner for modern data teams. We help some of the world's best organizations scale data and drive real value from the players that they make.
Speaker: you're thinking about building a team or making a hire and you're struggling with talent or just want some insights on the market, then I'd love to jump on a call with you and tell you a bit more. Equally, if you're looking for a job and want to find your next dream role, then reach out to myself or any other Coventify team.
Speaker: We'd be happy to see if there's anything on our books that we can help you with and give you general advice on the industry. Finally, big thank you to Omni, this season's sponsor. If you'd like to learn more about the AI analytics that Omni can deliver you, then check out the link in the show notes or come speak to me. i can happily point you in the right direction.
Speaker: Again, thanks for listening and look forward to seeing you a few weeks time.
