Matthew Cheung: Hi, I'm Matthew Cheung, CEO of ipushpull. And the following is a conversation with Brad Levy, CEO of Symphony, on the topic of what is your chat strategy? So Brad shares his thoughts on the growth of chat from DMs or direct messages to interoperability between messaging platforms, the importance of compliance and security for Symfony, and the impact of AI and large language models in the communication and chat Ecosystems enjoy.

Brad Levy: I mean, the word chat is either really generic or loaded. It sort of depends on the perspective where it could mean a DM. It could mean a bot on a consumer site, right? It could mean an actual automation, you know, workflow fully. We're sort of, and that's, I would argue ChatGPT is more of that, like automating a workflow based on a natural way of conversing.

Brad Levy: So like, I think that's, and then we obviously have a position in that from different perspectives, pure encrypted DMS on symphony, extending the consumer. And then obviously a lot of automations and bobbing up just if you think of AOL and Yahoo finance, Yahoo chat from back in the day, and even CompuServe, right, ways of, and then leap forward to WhatsApp.

Brad Levy: Right, in between those sort of consumer level, very widely accessible DMing, most systems I'm in now have some sort of a DM component, whether it's Snapchat or LinkedIn or Instagram, right? There's, and then they're even building bots and video and audio. So there's a. There's definitely a movement where I know IDM with people across 20 different applications now in life, right, from personal to professional and symphony is one of those as an encrypted version built for industry.

Brad Levy: So I do think, and that's a whole space in and of itself, just the idea of direct messaging in different ways on different platforms, you know, at different moments for different purposes. You know, that's a, that's obviously something that's of age with the generation and maybe even the pandemic kind of pushed us into a more social interaction on just about any platform.

Brad Levy: You know, even dating back to bots like Clippy back in the day, maybe the first GPT thing in my face, you know, can I help you do something, you know, can I offer you some assistance?

Matthew Cheung: And so how do you see the delineation then between the, the DM chat versus what your platform allows people

Brad Levy: to do? I guess I see the delineation as less of a delineation today and going forward, meaning.

Brad Levy: There's this idea of asynchronous, where you can have a person chatting on one end, physically typing, the other end can be a automated process that comes back as human typing. Right, so it feels like a human to human, but one side is a chat bot, the other thing is a person. You can then still remain as a person, but then have a full automation on the other side, where you're basically chatting into something that's a process and a real bot.

Brad Levy: On the back without even the human interface to it because the nature of the use case and then over time both sides can think more automated more chatbot and even more api and automated so i think one thing that we're seeing is less people having to match up at the same way at the same style like in the same mode as much as you can have a person chatting.

Brad Levy: And a bot chatting or fully automated process and then each can sort of take their own flavor or even go with their own pace to automate over time and you can have some moving faster or just in a different way than others on the same sort of network and that's kind of what symphony is becoming is that asynchronous automation platform that lets you be as personal touch or as automated and sometimes the volatility in the markets will drive whether people want that direct touch or maybe you're fine with just an automated process behind it.

Matthew Cheung: And so, since Symphony arrived on the scene, chatbots have, have kind of been unlocked for a lot of the big financial players, and the banks have had the resources to do it, and then as time's gone on, more, it's become easier to use, more configurable through either third party vendors doing it, or you providing, you know, developer toolkits yourself.

Matthew Cheung: How, how have you seen the trend to date? In chatbot users versus human users. And how do you see that? How'd you see that trend continuing over the next couple of years?

Brad Levy: Yeah, the trend on chatbot and that side is definitely up and probably more slope up into the right. There's still an incrementalism.

Brad Levy: I think that the industry goes at things, which is not bad, right? They take a big problem, break it down into parts, think about what we can solve today. And so it's more, I would say smaller problems solved in a more bilateral bespoke way. Then i say step two is taking some of those smaller processes and connecting them but still maybe on your side of a development initiative we're not depending on too many others you know than the cloud in the development kits that you have and then once you do something a bit more interesting those processes can connect to others in the marketplaces you know.

Brad Levy: Could be more other market infrastructure players in clearing and settlement or research platform so it's like do one thing small do three things small connect them and then externalize that behavior and that gives compliance people in tech people and business people comfort that there's some path versus.

Brad Levy: Everybody's it's sort of bots gone wild you know in a moment which is thirty years of bots in my opinion scripts on api's with databases automating a process we're not just able to do that much more. Easily with cloud and development kits, but there's still some real gates that we all have to deal with.

Brad Levy: And I think we allow people to do something quite small inside to much larger and outside over time and that trend. And I think the people interaction with that is increasing at an increasing rate rather than this idea that these bots are actually taking people out of their jobs or just moving on to more complicated orchestration tasks maybe than one thing in a row.

Brad Levy: So I definitely believe that trend, but you know, some people that do very specific jobs could be botted, but it's much. I think the human side is just accelerating, lagging, but at pace with the bot

Matthew Cheung: development. Yeah, it's all the co pilot assistance perspective, isn't it? So going back to two of your points, one around direct messaging and the other around internal versus external kind of outside.

Matthew Cheung: In its nature, direct messaging through all the other platforms you're talking about, like we have on our personal phone, in its nature is mostly external. Whereas what's happened with some of the growth initially around some of the automation was more internal just because it's easier to do and it's easier to, you know, plug into your own systems and so on.

Matthew Cheung: One thing that Symfony's had a lot of success with in the last couple of years is the, the, the kind of interrupt between WeChat and WhatsApp and Symfony because of the encryption tools and so on that you provide making it, you know, just super easy, you know, resistance. How do you see So two questions.

Matthew Cheung: How do you see the the interoperability between different platforms and symphony kind of evolving over time? And then the other is how do you see the internal and external piece playing out? Because probably actually one feeds the other as well. Because if you're connecting into WeChat and WhatsApp, for example, there's a whole bunch of external users.

Matthew Cheung: You can now talk to. But how do you see the growth areas coming in the interoperability between platforms and then the external? Chats or external conversation and communication that's, that's enabled.

Brad Levy: Yeah, and it's I would say it's, it's related for sure. So this idea of physical platforms being able to speak to each other.

Brad Levy: Again, I grew up at a time when text networks were separate, where you literally, you couldn't really text people outside because it was both not as fast and quite expensive, just physically, like there were different networks. Technologies often start as, you know, somewhat separate and either it's a winner take all or they become more interoperable more often than not.

Brad Levy: It's interoperability that wins and you could have cable television and streaming on the same Samsung TV run by an Apple TV or Roku. So that's coming. I think what will drive the need is some idea that internal and external or not easily bifurcated or personal and professional or not easily bifurcated and that's probably two planes that are worth so i know i have people on my whatsapp to symphony today because there are some conversations i'm having with them that are business and i want to keep them on my symphony now they're also very personal relationships for me at times which you I might peel off and go to my own personal cell whatsapp to talk to them because it's about my children or something but so i try to even do that today where i'm bifurcating i'm having a whatsapp with you directly versus a symphony to your whatsapp and that's a distinguishing that's distinguished between personal professional and life which is healthy the idea that you have internal systems and external systems is to me a temporary state.

Brad Levy: By fear or need saying you can only use this internally or that externally if you're using something a lot internally you're probably don't want to go outside and if you're using something a lot with your client you're not going to want to constantly move to a different place to do internal email is the one thing that hits all of that but it's become a very difficult tool to make.

Brad Levy: Useful in scale for people in teams. So I just think, you know, I know, bring your own device, you know, two devices is kind of a nightmare for people, but that's probably coming a bit now with some of the concerns, but that will converge again. It's already been back to one device. Now it's probably going to go to two with all the texting.

Brad Levy: Investigations but i do think they ultimately come back to one and things that are more interoperable and the technologies now are more available than ever whether it's cloud or sass or edge compute like all these things are making it much more possible to interact interoperate and have a. Your choice type of model versus here's the box in that box can't speak to that box that's three to five years out i have no doubt but it will take the internal external desire to use cases that make sense and then the platforms will then become more interoperable by need not by.

Brad Levy: Theory, if I'm

Matthew Cheung: if I'm a buyer of symphony, if I'm a bank or a broker or a customer, how, how how often do you see the split of people spending money with you coming from the compliance and security and encryption side versus what actually you can do on the platform and the enablement through automation and chatbots and yeah, this communication layer, because as the regulators have continued to tighten up, As we've seen it's, it's played well into your environment because you are secure and encrypted, but from a, from a, I suppose, a sales perspective, does, does the tail wag the dog or, you know, or is it a combination of both?

Matthew Cheung: In parallel when you're when you're selling symphony Yeah

Brad Levy: depends on the strength of the wind and the depth of the water I guess sometimes the dogs can do either if that makes sense. I made that up but Compliance was definitely a feature day one and I would even say encryption meaning The idea that you just can't get to this data without being able to get to the party whose centers receive the data But then compliance to me is a much more complicated platform of real tools that allow you to gate who is in and when and how that Changes as markets change or roles change so that investment in compliant minister compliance administration is heavy and huge and hard to do arms length as opposed to with the clients you're working with so that's always been there and i would say comes and goes in terms of new cycle like encryption is now very.

Brad Levy: In the news and even i would say information security whereas it used to be more about real data protection and even when gdpr came about right it was much more about personal protection which is still there but now it's about national security which has come around in the last few years so i think that's a bigger discussion and it could even layering quantum computing etc which really driving us now there are purposeful.

Brad Levy: Additions of value in workflows like settlements or client onboarding or getting to research or seeing a small bits of information in a big sea of data like those are the things we're targeting in a pretty granular level. Again probably uniquely as a company a symphony versus say some of the larger communication technology providers like you know whatever the slack and zoom and teams so we just see our purpose and our value and our problem solving is different.

Brad Levy: And then ultimately over time that drives some of the where the interop needs come from to come back to that other point. So I definitely see this compliance encryption. It's always part of what we do. It's become more involved. Great, but solving for really particular challenges in industry workflows is kind of what we do.

Brad Levy: It sounds boring and even too hard, but it is the nature of how we partner with I push pull

Matthew Cheung: moving on to a large language models. So the end of last year saw two things. Yeah, ChatGPT was launched and you acquired Amenity Analytics. So AI is, you know, By the

Brad Levy: way, we acquired Amenity Analytics and then ChatGPT was born in that order, November 29th to November 3rd.

Brad Levy: So we were a day smart, at least looked.

Matthew Cheung: AI has been around, you know, for, you know, 50, 60 years, a very long time, right? But, but things have changed a lot since, since, you know, the end of last year. And People looking at their strategies. How can they use chat because everyone's played around the chat GPT and it's like, okay, how can I use this in my business?

Matthew Cheung: Lots of questions come up about data and security like we're just talking about. But in regards to your business, what impact has large language models and AI had on your kind of product road map on solutions that you want to provide to customers on new business and how you're growing?

Brad Levy: Well, fortunately, we did acquire a company in the space the day before that, you know, that sort of November 30th day, which is helpful because we do believe we have a ton of value in the data space.

Brad Levy: We didn't have a lot of tools, picks and shovels or platforms to do it. Now we do. We still do not see this data. We are not a data company at that on that level. Like we're not selling, we're not analyzing and selling the data as much as the tools and the platform that you can do it against your data or maybe other public data sets that we've already built to.

Brad Levy: So, Number one, it's been well timed for us just to be incredibly relevant in the A. I. G. A. I. L. L. M. N. O. P. M. L. debate or or ramp up. yoU sort of alluded to it, but I never think that anyone technology is that big to drive a lot of change as much as a combination of a lot of technologies, whether it's cloud or, you know New materials and NVIDIA chips like all these things make these innovations that big and disruptive LLMs or have been there.

Brad Levy: They're just now much more accessible. And and pragma and practical. I also think combining these massive LLM capabilities, which many providers will have. We may be a bit of that. It's going to be the more. Detailed models that need to be delivered in much more tailored environments of like, say, the commodities market, which has 100 asset classes in it from palm oil to carbon to electricity.

Brad Levy: You can build one model to cover commodities. My guess is there's many smaller models that need to be built for palm oil trader and hog trader versus shipping container freight future, right? There's tons of these things that around commodities. So. I think it's combinations of people that have the user with the desire, are trusted enough to have access to information that they never see, and can then dovetail it with the largest LLMs that will be very dynamic and iterating, like right now they are, like literally.

Brad Levy: So I just think it's that convergence of lots of technologies, but then convergence of multiple models for different, very specific use cases. You know, again, predicting settlement fails is going to be very different than finding the best palm oil trade at that moment. You know, in a fully shipped containerized box at your port,

Matthew Cheung: just veering away from chat, you know, actual people chatting using their mouths.

Matthew Cheung: You'll see a quiet cloud 9 couple of years ago, so you're in the voice space as well. How do you see the intersection between what's happening in voice and what and where you want to head to in terms of innovation with that? And how does voice that interacts with bots that may be may be sitting in a chat?

Matthew Cheung: And I obviously connecting the two things together, right?

Brad Levy: And if you think of like energy policy is probably needing a lot of different types of energy renewable to come together to solve like communication is broad chat is a form of communication and DMing as our emails, as are the telephone as our videos, then your actions, even on a network might be part of what you communicate to someone with you clicked on something so.

Brad Levy: I take a very broad liberal view of what communications means. Fortunately for us, we're in that chatting space. Emails are highly accessible because they are, whether it's the corpus or the workflow phone is a lot harder to get to in our space, but we have one of the key leading products in that space called cloud nine.

Brad Levy: anD then you have video, which is probably that fourth horse of. Communication mode, which both comes with its own data. It's very heavy. In the Peta Terra, but it's also there's different tells, right? There's different information you get from a video where you could tell like retina scans and pupils dilating and people body language, right?

Brad Levy: I mean we're going into electric season in the U. S. There's a lot of body analysis experts that are going to pop onto the talking heads. You can AI that stuff and to me that's all communications all being communicated and all consumable by the LLMs or whatever the big AI machine. To put out a very full view and we think we play a role in almost all of that in theory So to

Matthew Cheung: finish off let's just look at the the chat ecosystem because it's a broad ecosystem.

Matthew Cheung: There's bloomberg that's been around for Decades there's there's there's you guys there's Ice chat. Yeah. Yeah There's a proliferation of stuff Which a lot of people don't like because there's too much stuff going on too many things to monitor too many things to to you know Having you know notification overload and you know real estate Screen space issues and so on.

Matthew Cheung: How do you see the back to one of the first questions around interoperability and about, you know, workflow and have an AI assistance and so on. But how do you see the proliferation of the different chat providers? Do you see ultimately everything being glued together and working cohesively? Or is it still very much some chat?

Matthew Cheung: Platforms are very much for one particular market, you know, power and gas people like this, you know, the palm oil trade is like this financials like this and they don't interoperate really at the moment, but when you've got your, your bridge between, you know, we chat, what's up and symphony, you can start to, you know, to eat into some of that and make it easier for the end user.

Matthew Cheung: Ultimately, how do you see the lap chat ecosystem in the markets

Brad Levy: playing out? Yeah. Again, I look to other places maybe to see what might the future. You look like so ten fifteen years ago i might have a netflix streaming app i had a crush on big hulking thing in my house maybe cable somehow a sonos for one speaker in one bedroom roco didn't exist apple tv kind of didn't exist right so there was but now whether you know all these providers you know.

Brad Levy: On my phone, like it or not, they kind of work together, right? I've got 20 places to DM. Now I have five that I tend to utilize. In business, you always have fewer, but you rarely have one, right? So I think in business, people will have what they need, which is probably something internal to do a bunch.

Brad Levy: That's just part of what they get when they start working. Then they probably, they may have a few different things that they bring together for chats or emails or document storage or video. Depending on the firm and what they've chosen, um, and then often you need something more specialized to be a bit more particular or go inside outside, right?

Brad Levy: I know I'm describing us, but I just see the technologies coming now with cloud in the lead and containerization and pads and APIs and sass and development kits being the core platform framework that we all have now. I know that there's people that are fairly, you know, a box. They will coexist in other boxes like monitors and framing of an application, whether you like it or not in your box.

Brad Levy: Those APIs will allow you to do a lot, interact as long as you have licenses. So I think you'll have your streaming apps. You'll have your big thing. You'll have your other ways of communicating, including email and all that will be pretty seamless in five years. I think the The Gen Z's are going to demand it.

Brad Levy: They're even going to want left and right swiping on monitors and everything's a touch screen like the five year olds today are touching every screen because they just don't know any other way. And they're the 20 year olds right now and they're going to be making decisions in five or 10 years. So I just think that world of interoperability.

Brad Levy: I've seen it come in my private life. It's absolutely coming in finance and we're all going to coexist with a lot more people around us competitively or interoperability wise because that's the nature of tech. It's what it does. And the demand from users is just going to, because people can see it here.

Brad Levy: They just can't quite apply it yet. To find capital markets and we're, you know, we're both a part of that fabric, you know, in driving a drop in our two companies. Thank

Matthew Cheung: you very much. This is the iPushball podcast series about what is your chat strategy. Over the next few months, we'll talk to leaders in the field and their views on the chat ecosystem, chatbots, and how AI is evolving the space.

Matthew Cheung: We hope you learn something new.