Andrew Capewell: [00:00:00] All right. Hello, and welcome to another edition of ipushpull's podcast. Today, I am joined by Richard Quigley, CEO and founder of Ventriks, and has a long background in the energy commodities data space. So without further ado, I'll let Richard introduce himself. Yeah,

Richard Quigley: thanks, Andrew. And, and appreciate being part of the podcast.

Yes, it is. As you said, I've been in the industry for over 25 years now. I guess a little bit of background. I, I sort of started in banking and became a petroleum economist. And a friend of mine started a data management company in London. And he said, I think you'd be quite interested in it because I had an interest anyway.

in in computer science and economics. And I started my foray into data management, and that was an upstream. And they sold the company to Harley Burton, and then sort of moved on to downstream data management. And I'm [00:01:00] sure we'll get into that, and what I've been doing over, over the last 25 years.

But that's my, my, my, my, my, my quick summary of, of entry into data management.

Andrew Capewell: Awesome. Thanks for that. Firstly, I suppose what they, the, what's the petroleum economist, economist do? Is that

Richard Quigley: Yeah, they do various things. My particular strand was working on fiscal systems. So we used to design fiscal systems for countries like Mexico, Australia, Russia so you're trying to you're trying to get inward investment from the shells, the bps, et cetera, and for most of these countries.

Because they didn't have that much money, there was a, a particular structure that we used to encourage folk to get some money up front and then you would share in the proceeds as, as, as they dug oil at the ground and can be quite complex and can be quite simple depending on, I mean, there's, there's obviously politics involved as well as economics and my, [00:02:00] my role was to, was to engage with countries in that and we actually worked on behalf of the IMF and the World Bank to provide some funding for, for, for some of these countries.

Andrew Capewell: Cool. I suppose we should probably talk a little bit about what you're doing now in terms of ventyx.

Richard Quigley: Yeah. So, so, so I, I, I guess to, I guess just to segue in there. So, so we, we, we, we managed to successfully sell genic mm-hmm . To, to, to inverse. And I, I sort of left and then started up Ventyx.

So, so Ventyx is. is very much I guess it's, I guess it's my 25 years of experience. Manifesting itself into a data platform, which is cloud native, so it's built directly from the cloud. Of course, you have no technical debt, so you don't have to worry about all that stuff. Every company builds technical debt, so we don't have that.

It's a blank sheet of paper. And we Obviously, we've learned a lot from, [00:03:00] from the other two companies, the Fame Energy, I was part of, or SunGuard and Datagenic the good, the bad and the ugly, and you try to you try to map out you know, what, what, what the system's going to look like. So we started it in April 2020, just, just, you know, lockdown had happened in March.

Nothing better to do. Yeah, nothing better to do. It was actually, it turns out to be. Actually, it was an amazing time because my co founder Kaushik we had time over a few months to release a blueprint, how we wanted to, to, to, to, to bring the company together. We, we just him and I really no one would want to jump ship during COVID, you know?

And only when COVID sort of died down, did we actually bring, bring people on board. And I think we have 18 people from my last company, Datagenic, that's now part of, excuse me, part of Ventric. So Ventures itself is, is I guess a, a, a very it's technology led data management [00:04:00] platform. We. We, we, we embrace as much as possible, as much as we can, we we, without sort of bringing in risk new technology cause new technology always brings in risk as you know, Andrew, as a product manager, so you have to, you have to understand the risk.

So, but we embrace as much as possible. So we use our primary focus is serverless technology. So, so the client doesn't really have to think about stuff and scaling and stuff and so that everything's done for them. We, we brought into we, we have a, a number of different things within the platform.

So we have our classic sort of we've devolved into three main areas. So data integration, which is your classic sort of ETL, ELT, bringing data in, ingesting data. And, and what that means the data management side, how you manage that data, you know, the master meta reference data management and all of that encompasses that.

And then a data analysis side. So three sort of, three sort of main areas. We have a fourth one, a data marketplace that we [00:05:00] introduced as well. So Gartner have got this amazing statistic that I can't remember what it was, 70 percent or 80 percent of. Data is going to be quite transactional as opposed to having a sales guy having this bespoke conversation when you just want to buy, you know, I just want to buy the price of Brent from you, you know, it's a very transactional.

So, so there's been quite a foray into, into marketplaces for various domains and I'm in the sort of energy and commodities domain. And we've, we've built up a number of different data providers. That are on the marketplace, both, both commercially premium private providers like Argus and, you know, and Opus and ICIS, et cetera, and exchanges like ex and ice, et cetera.

And, and then, and, and free data, like sort of, and so e and so G, you know, Noah, mm-hmm . On data, et cetera. So but extremely important data. So within the platform itself. You can do a [00:06:00] number of things. So you can do forward curve construction. But we took a step back and my last company, Datagenic, we had a rules framework.

It was a third party rules framework. But it had it's own, it had it's own language, it was called Drools, and it was, it was quite an unusual construct and actually, even a normal programmer was kind of taken aback by the construct, and it took them a while to get into it, and it was quite difficult, and we also operated with, you had to operate with it was a kind of, it was kind of dual construct, you had Java and you had this construct of drools and it made it, it made it, in some respects, it was a barrier to a sale because a lot of folk didn't have Java.

People didn't really want to learn another language, drools. So I, so one of the lessons learned was actually. to make it super simple for clients to use their own language of choice and to never have that sales barrier. So we now operate in 11 languages, 11 programming languages. So [00:07:00] whatever, whatever AWS supports, we support and currently support in 11 languages.

And so you can write a rule in any language and that rule is then reusable. In the construct of fault curves, for example. So if I have one month in, in, in a quarter, what, and what do I do to derive month two, month three, for example, and quite simple calculations. But that, that's a reusable rule.

And it makes the platform very, very powerful because of that. You've got this. Got this, this rules framework and we use it in a number of different areas. So we use it in quality as well. Yeah. You can do analysis. So we, we have a. We have a like a formula editor where you can do obviously formulaic expressions.

We bring in, we've got like 200 functions, mathematical functions and stuff, but you can also bring in rules as well, and it makes it just unbelievably powerful, but underpinning all that, of course, is the auditability, traceability, everything's versioned [00:08:00] and that really is, is a big thing for clients to know that.

And of course, we're ISO 27001 SOP2 compliant, which really does help, especially when you're handling people's data, especially when you're handling people's data from a security point of view, you know and so, so yeah, so, so taking a step forward, you know, what we kind of what we do now, so we're, we're, we're, we're kind of embracing or trying to embrace.

I guess AI.

Andrew Capewell: Yeah.

Richard Quigley: And AI is, I mean, it's, it's a huge area, you know, it's been, I mean, I learned at university back in the day, you know, but it's, it's now to the stage where it's, I, I mean, a, a normal person now understands AI because they have the concept of chat GPT. Yeah. Before, or, you know, and, and, and, and, and they can use it and they're going, wow, this is amazing.

This conversational survey, AI is amazing. And and we are starting to, over the next quarter, we'll be sort of introducing just [00:09:00] our own sort of LLM, so Large Language Models. So all the data will be sucked into an LLM and you can start to query the data. You can start to ask it things like, you know, potentially from forecasts.

One of the other things we're, we're going to be doing from a an ELT, so ELT just for the audience is, you know, extract and load and transform or extract, transform and load, whatever way you want to see it. But the most complex part of that is the T part is the transform. So when you get a, when you get data can be coming all shapes and forms.

But let's. Let's say a classic CSV which is a semi structured format, and everyone has their own version of a CSV. So when people say CSV, they think it's a homogenous item. It's not. It's heterogeneous, really. And so we have to do something with it. And you have to do something with it. There is it makes it that it's usable for the consumer as in the client or a system that's consumable, not just for that particular provider.

Let's [00:10:00] call it PLATS. But Argus has got CSV and you want to make it so they can, it's apples for apples comparison. So you want to do what's called a normalization of that data. And so we, we normalize at the moment semi automated fashion and we're trying to get it to to use AI to embrace that, to, to start to normalize that structure.

And that will help us. scale it even faster. Yeah, it's kind of there. I mean, like anything in life, it's, it's not 100%, you know, and you do need that bit of intervention. But we're, we're sort of doing a bit of R& D on that sort of right now. And another use of AI that we're using as I said, we have this rules based framework for quality.

is anomaly detection in data. We already have a bunch of rules that we use for anomaly detection, but AI actually could embrace it even better and help us scale. So we're starting to, starting to use it, but what excites me [00:11:00] more than anything is artificial general intelligence, and probably most folk haven't heard of it, I'm sure you have had to, but it takes AI To, to, to the next revolution.

And it's, it, it's, I I, I mean, I mean, it, it, Elon Musk, it is obviously a, a huge advocate of ai. He's, he's, he's got he's own AI company. I think it's EIX, it's called, isn't it? And of course there's open AI and all the rest of it. But it has, it potentially has, its, its problems. 'cause what is, what does a GI do.

So a GI. AGI's first building block, according to OpenAI they published a white paper in April. was to do a conversational AI and they've already, they've accomplished that. It's GPT 4 and another version of that out there. Google has their own version of Meta, etc. But it's, it's amazing, and it can do a drawing for you, it can do a video, a PowerPoint, you know, and it can answer questions.

But it's, it's, it's, it's, it's, it's quite [00:12:00] black boxing. And, but the next stage of, of AI is what we call reasoning, and, and reasoning is the hardest thing to do, and basically you're wanting to take the, the concept of a human brain and think like a human brain at a PhD level. And, and reason in, in a problem set without any access to any data.

So it has to do it, it has to do itself. And that reasoning is where we are just now in, in, in the journey for AGI. And that's a step two in reasoning. The step three is what we call actors. And actors already exist, but the, so actors will be they already exist to do certain things. Well, maybe we have robotic process automation, which has their own sort of actors, but the actors and AI are autonomous actors.

So now I can go away on holiday, Andrew, and it does all my sales for me, you know, in a, in a sort of a [00:13:00] cheeky example there, so the actor can, can work autonomously in whatever field that you're doing. And the next stage is the innovation. So you're, you're looking at AI to, to actually self innovate. So come up with its own ideas.

How to, how to, it looks at problem sets and how can I solve the problem of, I don't know, a new drug for cancer or whatever, to solve cancer or whatever. And, and, and, and that is a, I mean, that's, that's super exciting. But the one, the one that, that keeps me awake at night is, is the, is the fifth one. It does keep me awake because it, it, it, it has, it has its own challenges is organizational AI.

Well that does Andrew is it, is it takes all of those four steps and it now can run autonomously an organization with no human contact whatsoever. It can set up a company, it can start, you can, it can, it can have conversations, it can transact, it can do anything you want. without human [00:14:00] intervention.

And of course, It needs guardrails. Mm-hmm. And, and you know, the EU and the US and all that's working on, you know, potential frameworks for guardrails and stuff, what keeps me up at night is not, and I, I get no doubt there'll be regulatory guardrails. What keeps me up at night more is the, the bad actors mm-hmm.

That are out there. Yeah. That, that will use that and, and, and cause chaos in the world. Yeah. And, and we're not far away from it. Andrew, I, I'm not joking, you we're not far away from it. Mm-hmm . From, from that, what, what, what would seem like a. science fiction to now, we're probably a handful of years away for that as a reality.

And then if you throw into the mix quantum computing, and we're really, we're really in a, I don't, I don't know what sort of era we're in now. So it's exciting, really exciting, but it needs its guardrails. Yeah, fair enough. I think

Andrew Capewell: that's like the main takeaway, right, is around there's, there's a lot it can do, but I think the limitations or the potential for, for going wrong is also very high as [00:15:00] well.

And I think. Yeah, it's interesting. I think I was talking to a group of I push both customers last week and I think there was a lot of sentiment there around the concept of, you know, it has all of these potential say like these things it could do, but I think at the moment a lot of people see it as very much as it's a tool in the same way that we've used.

Computers to generate documents rather than handwriting or calculators to do math, to do mental arithmetic rather than trying to do it in their heads or on a piece of paper, etc. And I think, at the moment, a lot of the business applications actually stem from using it as that next extension of a tool, the next extension of growing that productivity.

And I kind of, I think, I see it at the moment as that, that kind of the next step for it is to, to grow the UK's had a lagging productivity problem since the financial crash in 2008 and you see the, these, these tools as a natural way of increasing productivity of each worker quite substantially. With, with the ability to use these tools, rather than, you know, the, the, the, I think the doomsday [00:16:00] scenario of it going off and replacing humans altogether, I think there's a lot to be said for it to, the incremental productivity increases that we can see over the next decade, I think, are, are, are huge, and I think it's, it's very much a time of embrace it, or Or find a way around it, but like, there's, there's, there's very few ways around it compared to the embracing inside of it.

Richard Quigley: Yeah, I, I, I think you're right. And I, I, I think a lot of companies, most of the, most FTSE 100 or whatever, S& P 500, they've, they've kind of embraced, they've got their own LLMs, they've got their own internal data. Yeah. And they're using it to interrogate and to, to, you know, to question and stuff. And I, and I think that's a really good, a really good way.

But it's the, it's the, the next evolution that I just described. Yeah. That, that, that's going to be. It really, really interesting for, for companies and how, how they go about embracing that. And it, it's, yeah, I, it's, I, I, I, I, I'm excited by it. I'm, I'm very excited by it. I've been in technology for most of my life but I, hopefully there are, as I say, some guardrails and mm-hmm

You know what, what [00:17:00] the thing about quantum computing that that, that, and I've been talking about for, for, for years is. You know, a concept to, to, to crack all, all our encryption which would take a millennia to do due using normal compute power, but quantum, they can crack it in seconds or minutes.

And you know, it, it, it has its own you know, dystopian sort of views on, on, on where we are with financial system. If that happens, we no longer have a. Have a, have a banking system, you know, a real estate system, et cetera. So, but, but, but I think there's an amazing amount of people out there.

They're involved with this right now to try and to keep the world going. We're still, we're still spinning every day, you know, and, and to, as you say, to hopefully embrace it and to, as humans, that we, we can, we can work with it side by side. And I'm, I'm, I'm trying, trying to evolve society to try and make it a better place to live in, you know, and, you know, we've got what, 8 billion human beings [00:18:00] now, so, you know, we'll work out how we can, you know, I don't know look at, look at, you know create more food stuff for us at the right place at the right time, there's ways to, you know, things like this using UI to do stuff like that.

I

Andrew Capewell: think, I think it's interesting you talk about the banking system encryption and all that problem and it almost is an advert for the blockchain at that point. It is an advert for the blockchain. It is.

Richard Quigley: Yeah. Yeah. I mean, you know, A

Andrew Capewell: very good one, in fact. Yeah, it is. It is.

Richard Quigley: And blockchain's amazing.

And, and again, it was, it's, it's kind of to some degree a bit like smartly trained. There was every single conference I went to mentioned blockchain, it's a bit like what's happening now. I was at a conference. You know, two weeks ago, and the whole conference was kind of underpinned by AI. Yeah. Most, most, most of the conversation was AI related.

Blockchain was that. And then it sort of died down a bit. There's a, there's a, there's a reticence for change.

Andrew Capewell: Mm.

Richard Quigley: As you can imagine. And, and, and blockchain, you, you need to change to embrace it. And [00:19:00] there's a, there's a ton of value add to be had. You know, it's a, it's a huge oil tanker. Yeah. That's, you know, that, that, that, that, that, that, that, that you, you, you have to turn around, you know, and it's going to, it's going to take a while, but you're right, you're right about blockchain.

Andrew Capewell: I think on that note, I think we've covered everything. It's been a nice long conversation. So thank you very much for joining me. It's been it's really interesting to hear some of your thoughts and some of your history as well, so. Yeah, thanks Andrew. Thanks, thanks for, thanks for having me. I really appreciate it.

Cheers. Thank you.