Technology consultant Steven Marsh started Smarsh in 2002 as an email archiving service for financial firms. Over the years, Smarsh quickly outgrew his Brooklyn apartment and added many other communications channels.
It has expanded into a global firm with more than 6,500 clients, including many RIAs and independent broker/dealers and some of the world’s largest enterprise financial services firms.
Its technology now tracks over 100 types of communication channels, including social media, texting, voice and many others. Today, its products go far beyond just archiving to include monitoring and analysis, employing sophisticated technology, including artificial intelligence in many forms.
WealthManagement.com recently had a chance to catch up with Smarsh CEO Kim Crawford Goodman during a recent visit to New York. She has run the company for 2 1/2 years, having previously served in senior executive roles at Fiserv, Worldpay, Dell and American Express and as a partner with Bain & Company.
The following has been edited for length and clarity.
WM: I’ve followed the company’s AI technology dating back to when Smarsh acquired the machine learning and natural language processing pioneer Digital Reasoning, and just last week the company announced its AI-powered Intelligent Agent. How do you view AI?
Kim Crawford Goodman: Smarsh has an extremely rich history in AI, the most longstanding history of any company in our industry of communications compliance. We acquired Digital Reasoning in 2020, which has been an excellent acquisition. We acquired a team with very deep capability and AI experience. In fact, that team just published a book.
We also inherited kind of a first-generation product line, which has done good things for clients. But as Smarsh, we needed to both uplift and integrate the capabilities from Digital Reasoning as we do with all of our acquisitions. So that product line that was originally Digital Reasoning, we’ve now developed and delivered to the marketplace; the second generation is native to the cloud and is repeatable and is more scalable and serviceable inside a company as well as across the industry.
With Intelligent Agent, we are building on our AI prowess by integrating capabilities with gen AI and developing large language models. But it also reflects a partnership with clients. We are working hand in hand with the most senior-level technologists and compliance people inside a couple critical clients.
It’s actually their data, but it’s in our platform. So, we can develop a large language model that’s specific to them. That is useful for compliance and surveillance. We can then use that large language model to develop products and services like the Intelligent Agent.
The other thing I’d say about our trajectory on AI was our recently announced partnership with Open AI. We are anticipating what the needs of our clients will be and what their needs will be working with these new technologies. So, that partnership is a perfect example of how we expect over time that our clients are going to start allowing these AI bots to speak on their behalf, answer questions in live chats, do some element of customer service, etc. And when that happens, we anticipate that there’s going to be just as big an obligation to capture all of that information coming out of the gen AI bots and agents. So we’ve gotten ahead of the pack and formed this partnership and are fully capable of capturing that content today.
WM: Does Smarsh have relationships with any of the other gen AI providers, like Google, Microsoft, Anthropic or others?
KCG: We have a longstanding relationship with Microsoft because as it develops and delivers new communication tools, they’re going to be in use in the financial services industry. Those tools have to be compliant. So, we work quite diligently with Microsoft, and it’s a two-way partnership. When they’re developing new communication tools and the APIs for those tools, Smarsh is typically at the front of the line, testing and introducing the capture for the Microsoft tools to ensure that those tools are actually compliant.
WM: Open AI also has a big partnership with Microsoft, too, so I assume there are some synergies there as well?
KCG: I would describe it as having independent partnerships with the two of them because the two are creating their own products. They manage their own partnerships in the background. But what we care about is that as their products and services come to market, we are able to facilitate, on behalf of our clients and our technology, the use of these things in a way that is compliant and reasonable.
The last thing I’ll say about the Open AI partnership is that we expect our clients to start speaking through these bots. I’m not a regulator, but regulators have said they expect every communication to be captured and evaluated. So, we are anticipating a huge growth there, too.
And we are also anticipating a responsibility for this, right? What we don’t want to see with our clients—there have been a lot of fines for channels that started to be used that people weren’t monitoring.
We’re trying to get ahead of that.
And you asked me about our journey with AI, which is the next greatest technological innovation. It’s no different from what we did with the public cloud. We are still, in our industry, the only scalable archive application that is native to the public cloud. The public cloud was a major innovation, and we took advantage of that very quickly.
By acquiring Digital Reasoning, upgrading it, and maintaining it, we are taking advantage of that for the clients. By working with clients and partnering with companies like Open AI, we’re now helping them [our clients] take advantage of gen AI.
WM: Do you see an explosion in communications generated by AI products? For example, even smaller advisory shops are employing marketing firms or fintechs with AI platforms for generating client communications that imitate or mimic their own writing voice.
KCG: The volume of communications has been growing nearly exponentially now for several years. And it was accelerated by the pandemic. We can just look at it in our own everyday lives, the number of conversations or communications that happen electronically. That may be on email, text messaging or social media, which we capture as well. So, the trend has been toward an incredible increase in digital communications, period.
It is also true that we expect this trend to continue with all the new forms of communication; there are always engineers trying to create the next TikTok.
Smarsh has the broadest portfolio of electronic communication capture tools of anyone in the industry—over 100. This is part of our commitment to ensuring that our customers capture what they need and are protected.
WM: How many developers do you have at Smarsh?
KCG: As best we can estimate, Smarsh is the largest and most strategic company in the space. We are at least two times the size of our nearest competitor, and we are fairly confident that our R&D budget is somewhere between two and three times the size of our nearest competitor.
I won’t give a specific number of engineers, but we have several hundred.
WM: What about LLMs?
KCG: Large language models are at the heart of how Smarsh is leading on gen AI. We talk to our clients about the most effective model you can build inside the enterprise, which is actually with Smarsh on the data that you have archived with us. We have every single communication, every single way that your traders are talking to the market, and every single way that your bankers are dealing with their clients and one another.
The language on which you want to build the model is sitting in this Smarsh data. Of course, we have the scientists who have a view on how you advance and build those large language models because, as it relates to financial services, we actually believe we should be the premier partner working with our clients on the development of these models.
WM: With your big enterprise clients, they have their own data repositories. Is that on your own cloud environment, or is it a cloud environment hosted with Amazon Web Services, Microsoft, Google or another third-party provider? And what is your own cloud architecture?
KCG: We have an enterprise-class platform that is native to the public cloud. We do work across all three major public clouds to support our clients. We do have a significant and primary relationship with AWS, so it is our cloud environment.
WM: Is AI going to end up replacing compliance officers?
KCG: I’d say that risk management and compliance within financial institutions and in wealth management companies of all sizes is an excellent use case for AI and gen AI. Because the data is rich, the need to analyze the data with accuracy is important. And the need to find the biggest problems is very important. So, it’s an excellent use case.
I would say that the key to making the AI and the gen AI effective in this use case is how the AI combines with the human effort; in fact, that’s our recent Intelligent Agent announcement. We do see instances where the AI can do things at the most basic level that humans are doing, and we’re developing the technology to do that. However, we believe what will happen with that is the AI will allow the institution to surface more of the true problem areas and the risk and, therefore, not have to do case studies and deep dives on things they used to have to do to find them.
So, no, I do not see a future where there is no chief compliance officer because it’s a bot.
If anything, I see a future where the chief compliance officer and his or her team are more impactful or more effective because they’re utilizing more tools; they’re reducing risk at an even greater rate because the tools are letting them find more risk. They are more technically sophisticated because they’re helping to tune the technology to achieve what their institutions want.
They’ll ultimately have more work to do because groups from across the enterprise will want to utilize this data for a whole number of things.
WM: How do you see data management evolving in wealth?
KCG: The data that chief compliance officers and compliance teams have assembled and managed over multiple years is extremely valuable. And I see that clearly at Smarsh because we do that on behalf of these folks in the most significant way in the industry—we manage on their behalf over 100 petabytes of data.
What I see is that the usefulness of this data, given the right governance, will likely be used across the enterprise in a more significant way.
WM: What’s top of mind for you on the roadmap that may not already be in Smarsh’s repertoire, or areas where you may partner or acquire?
KCG: We hear quite a bit about iMessage on the Apple platform. In order for these things to be compliant and captured, you have to have APIs made available to you, right? But iMessage is certainly something that financial institutions would be interested in.
I would highlight that voice has been around a long time—it still matters a lot—and now the translation of voice, transcribing it no matter what language it is in and putting that into digital comms so that you can measure it.
The third one—we talked about it earlier—is ChatGPT and what the other gen AI vehicles are doing. Those are the three areas of digital capture that we hear a lot about from our clients.
We’ve effectively combined internal development with acquisitions of excellent technology companies, not just the acquisitions of those companies but the integration of them. We’ve continued to drive that integration deeply and now have a platform, whether it’s for mom-and-pop and smaller clients or whether it’s for enterprises that integrate these things together so that people can get the advantage of the entire platform. That puts us in the position of being open to all kinds of partnership opportunities. And we see ourselves continuing to do acquisitions over time.