The Latest In Financial #AdvisorTech (November 2023) (2024)

Practice Intel Launches To Help Advisors Go Deeper Into Measuring The Depth Of Client Relationships

Financial advisors have incredibly high client retention rates, with most industry benchmarking studies showing approximately 95% retention rates in the standard AUM model. Of course, individual advisory firms can and do vary in how successfully they retain clients, but amongst firms working with ongoing clients and managing their assets, the difference between 'good' and 'bad' firms in this regard amounts to only around 5 percentage points. More specifically, even 'bad' advisory firms may still have retention rates of around 92%, while firms that are the most successful at retention can be as high as 97% or more.

Notably, though, even though these differences in client retention may be small on a year-to-year basis, they're far from trivial when viewed from a long-term perspective. A 92% versus a 97% retention rate, despite being a difference of 'only' 5 percentage points of client retention each year, is really the difference between the average client sticking around for just over 12 years and staying for more than 30 years. Which means that a 5% swing in retention rate can result in more than doubling the lifetime value of each client.

One issue is that most financial advisors don't necessarily have a good way of measuring whether they're closer to the 92% or the 97% group, other than calculating it after the fact. But there are fewleadingindicators for financial advisors that can help them understand how satisfied and happy their clients are (or aren't) – which would come in handy, since knowing whether a client is dissatisfiedbeforethe client breaks the news that they're leaving gives the advisor the opportunity to address any issues and potentially convince the client to stick around.

In recent years, a number of providers have begun to emerge that aim to fill this gap by helping advisors distribute client surveys to understand how satisfied their clients really are. In many cases, these tools utilize metrics that are common in many industries such as CSAT (Customer Satisfaction Score, which measures overall client satisfaction) or NPS (Net Promotor Score, which provides an indicator of how likely the customer is to "promote", i.e. refer, the business to others) – an approach taken by tools likeNexa Insights, which can send out standardized NPS, CSAT, and similar surveys to an advisor's client base. Other tools, such as Julie Littlechild'sAbsolute Engagement, seek to go even deeper in understanding whether clients are not just "satisfied", but actually meaningfully engaged with the advisor and their services – driven by Littlechild's research showing that while satisfied clients tend to stay with an advisor, "engaged" clients are the ones who will actually refer other clients to the advisor.

Following in this vein of seeking to measure not just the overall satisfaction of clients but the actual engagement and depth of the advisor-client relationship itself, this month a new solution,Practice Intel, launched with a tool that's built specifically to help advisors measure the "relationship quality" that they have with each and every client.

Although Practice Intel's description of itself as a "growth platform" echoes the branding ofrisk tolerance/proposal generation/sales enablement tool Nitrogen, the "growth" that Practice Intel seeks to facilitate is in the quality of the client experience, by in its own words, helping advisors "align their practice with what clients value most". In practice, Practice Intel's tools include a survey that can be sent to every client to measure the client's satisfaction in several aspects of the client experience (including whether they receive comprehensive advice, whether the advisor understands their values and goals, and whether the advisor places the client's best interests first), ultimately rolling these individual scores into a single "Relationship Quality Index" (RQI) metric.

From an advisor perspective, Practice Intel makes it possible for advisors to understand which clients may be less than satisfied with the firm and thus may be at greater risk of attrition, or conversely, to understand which clients are most deeply committed to the firm (and perhaps more likely to refer it to others). Perhaps most importantly, however, the tool provides in RQI an actual metric for relationship quality that advisors can then try to improve over time – because as the saying goes, you can't manage something if you're not first measuring it. All of which could be appealing to many types of advisors seeking to improve their client relationships, but especially so for mid- to large-size multi-advisor firms where it's more difficult for firm owners and managers to know which advisors are doing the best job building relationships with clients, and who may need further training and support.

From an industry perspective, Practice Intel, along with its competitors like Nexa Insights and Absolute Engagement, have arisen to fill an area where many advisory firms are lacking by providing actionable data on the quality of client relationships – a gap that exists arguably because advisors have been so used to having year-over-year high retention rates that they've historically invested little or nothing towards improving those rates, or even understanding what their own retention rates even are (even though, as noted above, the difference between a 92% and 97% retention rate has an extraordinary economic impact over the long term).

At same time, when financial advisors aren't used to using tools to track client satisfaction, it isn't clear whether they'll be willing to pay for solutions to start doing so. Also uncertain is whether advisors will pay more for financial advisor-specific solutions like Practice Intel (which charges $195/month per advisor for its tools), when they can alternatively use a general-purpose NPS survey app likeDelightedfor less than 1/10th of that cost.

Nonetheless, the core point remains that relatively small incremental shifts in client retention rates can produce outsize long-term returns for advisory firms (albeit ones that they'll most likely only see years, or even decades, down the line). Meaning there really is the possibility of a strong ROI for advisory firms who invest in better understanding of their clients' satisfaction. The question is simply which form of measurement – traditional CSAT and NPS, Absolute Engagement's engagement metrics, Practice Intel's RQI, or something else entirely – is actually most predictive of client retention and referrals to generate that ROI in the first place.

FinanceHQ Launches As A Niche-Focused Pay-Per-Lead Generation Platform

The last decade has seen a proliferation of technology platforms dedicated to providing leads for financial advisors. Notably, lead generation itself is not a new service – financial advisors have long had their pick of methods for lead generation, ranging from direct mail to seminar marketing to even tear-out postcards in magazines. But the current wave of providers, rather than using 'analog' marketing strategies that were predominant in the past, are instead focused ondigitallead generation, seeking to gain visibility by using digital ads and SEO to capture people who are already searching for financial advisors to match them with the 'right' advisor for them.

In a world wherethe average client acquisition cost for financial advisors is thousands of dollars per client, there's a lot of money to be made for a platform that help advisory firms with their challenges in bringing in new clients. And those challenges have only seemed to become more acute in recent years, as firms' organic growth rates have slowed to the mid-single-digits (averaging 6.2% per year from 2018 to 2022,according to Schwab's annual Benchmarking Study), whilethe increasing number of comprehensive financial planning firms in the marketplace makes it harder for any one firm to differentiate on comprehensive planning alone. All of which has led to lead generation becoming one of the fastest-growing categories of providers on theAdvisorTech Map.

But the issue that many advisors run into with lead generation services (and which led to the category ranking dead last amongallcategories in average advisor satisfactionin the most recent Kitces Research on Advisor Technology) is that the leads provided by the services often simply aren't very good. The tools often allow prospective clients to filter and browse for advisors based on criteria like geographic area, asset minimums, and increasingly also by client reviews and ratings – yet those attributes by themselves arguably aren't enough to determine whether an advisor is actually a good fit for a given client. In practice, lead generation toolsoften send prospects to advisor who might nominally "match" with the advisor based on the tool's criteria, but whose problems don't align at all with the advisor's specific skills or expertise. And while many of these tools only charge their fees when a prospect actually becomes a client, there's still a material cost to the advisor in their time spent vetting the prospects sent by the lead generation service, and the expense can add up quickly if the leads prove to be low in quality.

It's notable, then, that in OctoberFinanceHQofficiallyannounced its launch as yet another competitor seeking to disrupt the advisor lead generation space. FinanceHQ takes a slightly different approach from other lead generation platforms by asking advisors first and foremost what their niche or specialization is. It's a feature that could both help advisors differentiate from other advisors on the platform itself (which could prove especially helpful if the platform becomes successful and each advisor needs to stand out among a sea of other advisors), and potentially more significantly, allows FinanceHQ to more directly target leads who are looking to solve very specific, actionable challenges – whom they can then match with an advisor on the platform who might actually be uniquely suited to solve those challenges.

FinanceHQ follows the path of lead generation platforms likeSmart Assetwhich charge a "pay-per-lead" arrangement for each prospective client the advisor receives, regardless of whether the prospect actually closes (as opposed to many other lead generation services which charge a percentage of revenue earned only after a prospect becomes a client). For advisors, paying a flat fee for leads could conceivably be a much less expensive source of organic growth than other channels that charge based on revenue. But then again, FinanceHQ's pay-per-lead model also puts the onus on advisors to proactively follow up on and close the leads they receive, or else risk losing both the prospect and the fee since FinanceHQ gets paid either way. This arrangement also risks the potential that advisors will need to compete with each other for the same leads: Because FinanceHQ charges per lead rather than for converted clients, they have an incentive to hand off leads to multiple advisors, who would need to compete amongst themselves for a common pool of prospects.

From a broader perspective, though, the real question isn't so much about the "pay-per-lead" versus "pay-for-success" fee models in advisor lead generation – it's simply about whether FinanceHQ can find and source an ongoing flow of interested leads who are actually ready to take action to hire an advisor, and how they can scale their platform for an ever-growing number of advisors who want an ever-growing number of leads. Because as noted earlier, client acquisition costs average out to several thousand dollars for a single client, meaning that lead generation can be extremely capital-intensive for platforms to build, and the risk is always that it will become even more expensive as other lead generation platforms proliferate and compete for the same pool of advisors and clients.

In other words, FinanceHQ's prospects for success are ultimately tied not only to the experience it creates to match prospects to advisors, or to the particular way it charges (as long their fee model adds up to enough revenue to keep its business operating); rather, the question is whether FinanceHQ's more niche- and orientation-focused sales funnel allows it to find a better flow of prospects – i.e.,those seeking specialized advisors who solve specialized problems– that FinanceHQ itself can acquire in a cost-effective manner to then re-sell to financial advisors.

Farther Closes $31M Series B Round For Its Digital RIA – But Can It Actually Achieve Tech Margins?

The typical scaled-up advisory firm spends around 35% of its revenue on overhead expenses, which include compensation for administrative and support staff, compliance and legal services, office (or virtual team) expenses, and technology. In recent years, however, there's been a growing focus on how firms leverage technology to become more efficient in the hopes of reducing overhead costs and improving profitability. Which seems simple enough in theory, but in practice, the ever-growing (and sometimes overwhelming) number of technology solutions – some of which integrate with each other in some way, but many of which don't, at least to the level that most advisors wish that they would – can just as easily serve to hinder efficiency as it does to help it. Suchthat in the latest Kitces AdvisorTech Research study, the average financial advisor rates their entire technology "stack"lowerthan the individual technology solutions they use; in other words, the whole is valuedlessthan the sum of the parts!

Over the past decade, two approaches have come along to try to solve the challenges for advisors around technology integration. One has been the emergence of "all-in-one" technology solutions, which effectively seek to fill every role (or at least the core functions) of an advisor's tech stack, ensuring that the most important pieces will work together seamlessly without any need to worry about their integration capabilities (because they're all built under one roof by the same provider in the first place).

More recently, however, another trend has started to emerge: Startup RIAs that eschew third-party technology altogether and instead build their own in-house advisor tech stack from the ground up. Typically these so-called "digital RIAs" raise outside capital to cover the significant upfront cost of building their native technology, with the hopes that their better (and better-integrated) technology will allow them to reduce their overhead costs and thus operate more profitably than a traditional (analog?) RIA, winning out their long-term valuation with superior profit margins.

In principle, the prospect of unlocking value by cutting down on tech-related inefficiencies makes sense; in practice, though, there's really only so much in overhead expenses that can reasonably be saved through better technology. Because in reality, the bulk of the expenses for advisory firms come not from administrative costs which could be lowered by improving efficiency, but fromthe direct costs of the financial advisors themselves who generate the firm's revenue. And while efficiency is still a fine goal to strive for, it won't necessarily be achieved by investing heavily in technology: In the 2022 InvestmentNews Benchmarking Study, the top-performing (i.e., most profitable) advisory firms spent less as a percentage of revenue on both administrative support staffandtechnology than the median firm on their path to achieving profit margins that were 9% better than the average firm. The top firms are simply more profitable because their advisors take on more clients than other firms, and meet with each of them slightly less often (because there's only so much time), allowing the firms to have better economics driven by a focus on charging full value for their fees and on not 'overservicing' clients – not necessarily by investing heavily in technology to reduce overhead staff costs.

All of which makes it notable that this monthFarther Finance, one of the new crop of digital RIA platforms aiming to build its in-house technology to optimize the efficiency of its advisors, announced a new $31 million Series B funding round on a whopping $131 million valuation as the firm approaches $1 billion in assets under management 4 years after launching in 2019.

On the one hand, Farther's somewhat eye-popping capital raise shows the sheer appetite that private equity firms have for investing into opportunities to find new paths for building RIAs that are even more profitable than the existing model. If Farther charges fees at an average of 1% of its approximately $1 billion in AUM, then the $131 million valuation of this fundraising round represents more than 13x its run rate revenue (and an even higher multiple on its trailing revenue as the firm started the year with a much lower asset base). Which indicates that Farther's investors are very excited about its recent 'booming' AUM growth (which at the aforementioned $1 billion is a 48% increasefrom the $675 million it reported in April 2023, which itself was reported to be a nearly 500% increase from the beginning of 2022).

On the other hand, however, it's worth recalling that the key feature of Digital RIAs like Farther isn't advisor recruitment or asset gathering, per se, but technology – namely, the streamlined efficiency that allows it, as it purports, to pay its advisors a higher share of the revenue they produce than other RIA aggregators do. And yet, as noted above, there's only so much potential improvement on the margins that technology can create through reduced overhead costs. Even a mega-firm with 30% profit margins that is valued at 15X its earnings (already a premium valuation for most RIAs) would amount to a 0.3 x 15 = 4.5X revenue multiple, and a super-profitable tech-savvy firm with 40% tech-enabled margins might get to stunning 0.4 x 15 = 6X revenue – which makes it hard to see how Farther's more-than-13X-revenue valuation is in any way connected to the operational efficiency that its in-house technology might provide… at least not unless it can sustain an absolutely extraordinary growth rate in its assets and advisor headcount, which suddenly makes Farther less a story about the tech-efficiencies of RIAs and more simply about its ability to acquire or recruit advisors quickly.

Which helps to explain why advisor recruiting seems to be an effort that Farther is now emphasizing,as the firm has gained headlines in recent weeks for recruiting a slew of former Goldman Sachs Personal Financial Management advisors in the wake of the former United Capital's sale to Creative Planning.

The challenge, though, is that recruiting and acquiring advisors in high volume becomes increasingly capital-intensive in the current advisor landscape amid heavy M&A activity and high competition for talent. And at some point, investors may ask what 'magic' Farther could really have that allows it to recruit advisors or acquire them at 'typical' industry multiples of 2X to 3X revenue, and then claim the same revenue is now worth more than 13X to its investors when even a 10% improvement in profit margins due to self-built digital-RIA technology can't possibly justify nearlythatmuch of a valuation lift.

Nonetheless, the opportunity does still remain for RIA platforms to improve their efficiency by building in-house technology to solve for the problems of managing and integrating technology. Still, though, technology alone can only do 'so much' to improve the margins, and valuation, of an advisory firm whose costs are still driven first and foremost by the compensation costs of its advisors themselves (not its overhead and administrative staff that technology tends to automate). Such that Digital RIAs like Farther have to buildsucha tech-savvy platform that advisors flock to itwithouthigh recruiting or acquisition costs – and become an engine that attracts advisors unto itself – to be able to grow into tech-style valuations. And even then, if Digital RIAs offer higher payouts to their advisors to attract them – and 'share' the cost savings of their self-built technology – then such firmsstillaren't more profitable than the traditional RIA, they simply try to grow faster by recruiting advisors faster by offering a better payout. Which means, ironically, that the success or failure of Digital RIAs like Farther may simply follow the path of any RIA aggregator: can the firm find a capital-efficient way to acquire and recruit advisors to ever-larger size and scale (regardless of its actual technology)?

Will Fidelity Blocking Screen-Scraping Data Aggregators Hinder Or Actually Catalyze Better Account Aggregation For Advisors?

The 2006 debut of Mint.com (RIP) provided one of the first successful examples of automated data aggregation:A continuous feed of data from multiple sources into a single aggregation platform, making it possible for someone to view their whole financial picture in one place and to track their progress over time. The early success of Mint (which within a few years was acquired by fintech giant Intuit) spawned a host of imitators that sought to provide the same holistic data aggregation capabilities to financial advisors, who recognized the opportunity in automating much of the time-consuming process of gathering and inputting client data. One of the earliest successful versions of a "Mint for advisors" waseMoney'sFinancial Dashboard, but today there are numerous AdvisorTech tools that rely on account aggregation, encompassing not only financial planning but also other functions like performance reporting and fee billing.

In the early days of account aggregation, many financial institutions didn't have ways to externally share customer data, and furthermore had few incentives to invest in the capability to do so (since many of the parties they would be sharing the data with could very well be competing with the institutions for their own clients' business). So instead of going through the financial institutions directly, the early data aggregation platforms used technology that came to be known as "screen scraping": A client would provide the platform with their login credentials to an institution's website, allowing the platform to log into the client's account, view the account balances as they appeared onscreen, and "scrape" the key information into the platform.

While screen scraping proved effective enough become the predominant method for account aggregators to collect client information for close to a decade, in practice the technology had some significant drawbacks. First and foremost was the fact that account connections tended to break frequently – either because the client changed their password or added multi-factor authentication (interrupting the data connection until the client could re-enter their authentication information into the aggregation platform), or because the financial institution changed its website or user interface in some way so that the screen scraper didn't know where to find the data it needed.

Another challenge that arose around the use of screen scraping was that, with Mint and other aggregation tools increasing popularity, the sheer number of logins that any institution's website needed to handle increased exponentially. Normally, a typical customer might log into an institution's website only once per week or month (if that) – but an account aggregation provider might log in eachdayto collect updated balance information. Which when multiplied across a significant chunk of the institution's customer base could strain their servers and create a material cost associated with supporting the infrastructure needed to handle that many logins.

So starting in the late 2010s, the industry began to shift to address the issues created by screen scraping technology. Institutions began to offer direct API feeds that account aggregators could connect to in order to get direct access to the data they needed, for which the client needed to provide only a one-time authentication which didn't require further updating. The result was a more stable data feed with fewer breakages.

From the financial institutions' perspective, direct data feeds also had the benefit of easing the crush of logins created by screen scrapers, which meant that many institutions that had previously been loath to share customer data externally, were now happily willing to create and provide the data feeds for free, since the cost of doing so would be more than made up for by the reduction in infrastructure costs associated with handling screen scraper logins.

In this context, it's notable thatas of October 1, Fidelity has gone so far as to cut off screen scrapers' access to client accounts entirely, forcing account aggregation tools to use a direct API data feed to collect client information. By Fidelity's explanation, the move was intended to ensure more stable data connections, reducing the burden on customer support, and providing more security for client data (being that screen scrapers, after all, had access to clients' login information and at least had the potential to gather – and potentially sell – client data that it found while nominally logged in for account aggregation purposes).

Fidelity's decision to end access for screen-scrapers represents another milestone in the large-scale shift towards direct data feeds. Which on the whole is a good thing, since it gets nearer to fulfilling the original promise of account aggregation by finally providing account connections that don't routinely break. There's also a benefit in the creation of consistent data standards that can be established and maintained at the API level rather than needing to be shoehorned in around the financial institution's user experience and requiring updates every time that UX changes – which in conjunction with cooperative efforts among financial institutions around setting data standards likeFDX, further increases the potential for more stable data in the long term.

One wrinkle that has emerged, however, is that because most financial institutions and account aggregation providers don't necessarily want to enter into individual data-sharing arrangements with every single one of the thousands of other institutions and aggregators that exist, most of the connections have been consolidated into a small number of middleware providers likeAkoya(which itself was spun off from Fidelity), which serve as a central hub by establishing core data connections to financial institutions and licensing those connections to account aggregators. Which adds a layer of fees to the data-sharing setup and raises the potential of increasing the cost of any tech stack that relies on account aggregation.

In Fidelity's case, Akoya – in which Fidelity still has a minority ownership stake – is the only licenser that will have access to Fidelity's data feeds. And while Fidelity has stated that they will cover Akoya's licensing fees and provide aggregators with "free" access to Fidelity data, the fact that data aggregators canonlyaccess data from Fidelity and its 86 million or so client accounts through Akoya means that they may be pressured to use Akoya forallof their data feeds, unless they want to sign agreements with multiple data licensors. In other words, if Fidelity's size and influence coerces data aggregators to set up all of their data feeds with Akoya, the added revenue through Fidelity's investment in Akoya might make up for the cost to them of providing their own data to aggregators for "free". (An arrangement which has aroused the interest of the Consumer Financial Protection Bureau for its anti-competitive potential.)

Ultimately, however, it does seem as though direct feed APIs really are the better solution for data aggregation, since they allow for more a consistent data flow, stable connections, and better security of client information – all of which can help account aggregation live up to its original promise as it pertains to financial advisors. But the caveat, as with many advancements around data in technology, is that direct data access hands more power to those who own and control the data – and so the question remains whether the cost of account aggregation will increase over the long term as providers like Akoya build and grow businesses around managing the flow of data – and whether the benefits of data aggregation are worth that cost in the first place.

Financial advisory firms have an obligation to ensure that their advisors actually doing the right thing for their clients, or at least have reasonable policies and procedures in place that would be likely to detect and capture any wrongdoing before it causes harm to other clients. Historically, compliance oversight occurred by having compliance personnel review advisors' activity, from advertisem*nts and marketing, to communication with clients and prospects, to their own personal securities transactions to ensure advisors weren't front-running trades in their clients' portfolios.

However, the challenge, especially for medium- to large-sized advisory firms, is that even if only a random sampling of advisor communication is reviewed, it's a very time-consuming manual process for the people doing the reviews. But the plus side, though, most compliance processes, by nature of their occurring regularly on an ongoing basis, also happen to be very systematized and repeatable – or in other words, conducive to at least some layers of automation through technology.

As a result, the industry has seen what was historically a cottage industry of compliance consulting services bloom into a growing array of compliancetechnologysolutions – and although the consultants still exist and can provide value in solving for more complex compliance problems, they are increasingly supplemented with technology to automate the more systematized parts of compliance. Meanwhile, compliance software offered directly to advisors – the pure technology itself – has become an increasingly hot business, with notable capital raises for companies likeSmartRIA, and M&A transactions forRIA in a Box(which was acquired byComplySci), BasisCode (which was acquired by Orion to becomeOrion Compliance), and Schwab Compliance Technologies (which was acquired byMyComplianceOffice). All of which underscores the idea that, with compliance procedures tending to be very repetitive andwith set compliance calendars needing tasks to be completed periodically at regular intervals over time, technology providers recognize the opportunity that can exist in providing ways to automate repeatable compliance tasks and lessen the compliance burden on midsized advisory firms.

Over the past year, however, a new angle has emerged on compliance technology: Where and how artificial intelligence (AI), including platforms like ChatGPT, can further improve, expedite, and automate advisory firms' compliance procedures. And while AI has found its way into AdvisorTech tools across many domains, it's especially interesting to see AI crop up in the context of compliance, since it has such important ramifications for an RIA's status in the eyes of state and Federal regulators.

Accordingly, it's notable that in the news this month has been the launch of not one but two new compliance vendors,RegVerseandHadrius, that in competing with the other technology solutions that aim to help streamline compliance, are both leaning heavily into the fact that they are leveraging AI to make advisors' (or at least their Chief Compliance Officers') compliance lives easier.

In the case of RegVerse (which calls its compliance tool Avery), the AI's function is to help summarize key regulations that may apply to the advisor, and generate action plans to complete compliance tasks, as well as offering a ChatGPT-style chatbot as an "intelligent chat assistant" to navigate various aspects of its compliance software. The impression, then, is that Avery's AI does little to automate the compliance work itself, but rather focuses on making information more digestible for the person doing the compliance work to help facilitate their decision-making.

Hadrius, however, goes one step further in the way it deploys its AI. According to the vendor, the AI tool can be used for monitoring client communications and identifying items to flag for review, as well as for reviewing marketing materials before they are sent out, leveraging their "ComplianceGPT" technology to try to spot problematic phrases and conversations (such as making unrealistic promises or guaranteeing certain outcomes in various communication or correspondence to clients)… in theory using GPT's language capabilities to detect more efficiently than the already-common approach of 'just' doing keyword searches for compliance-problematic words like "guarantee". Which still isn't quite doingthe work of compliance personnel per se, but can significantly aid the staff in pointing out specific areas to look at when performing reviews of advisor advertising and communication.

From an advisor perspective, arguably anything that expedites repetitive manual compliance procedures is helpful, especially as advisory firms grow to be multi-advisor and have increasing oversight obligations over marketing, client communications, trading practices, and other areas. And at least in theory, AI technology could be especially promising for compliance applications that often involve going through reams of information to try to find concerning patterns or data points, whether it's the sheer volume of email communication in a mid-sized advisory firm, or the reams of advisor trading data at a large advisor enterprise that needs to watch for front-running or potential trading by 'rogue' advisors against its own clients.

At the same time, as with almost any scenario where advisors rely on AI to solve for not just rote tasks but ones that require "intelligence" to accomplish successfully, the question remains whether AI can be trusted yet to handle tasks requiring technical knowledge in specific areas like advisor compliance. Can ChatGPT-like tools such as Avery can actually sift through complex regulations and provide actionable recommendations for advisors that are on-point regarding sometimes-very-technical-details in a particular regulation… especially whenthe actual ChatGPT has been found to have invented fake legal citations when a lawyer trusted it to write a legal brief? And can Hadrius actually comb through a pile of marketing and communications materials and reliably spot problematic promises from advisors to their clients any better than the keyword-watching that most large firms have already implemented (and if firms leverage the technology to reviewallof their communications rather than just a random sample, does that actually increase their compliance burden by implying to regulators that they now actually expect to catch 'every' instance of wrongdoing)?

In other words, the promise of AI automation is one thing, but reality of doing enough due diligence to figure out whether the technology isreallyeffective at doing what it purports to be doing, is a unique new issue that advisors face with the rise of AI in particular. Because at the end of the day, the firm and its chief compliance officer are responsible for ensuring that compliance tasks are actually done, and done correctly – meaning that if the technology isn't up to snuff, it's the firm and CCO who will have to deal with the consequences for choosing that particular AI tool in the first place.

From the industry perspective, while compliance, with its high volumes of materials to review on a regular basis, is especially favorable for automation-driven use cases, the question remains how close the technology is to the point where advisors can hand off those responsibilities to AI in particular, and count on the technology to do it accurately – and perhaps, how much of a margin of error they believe they will have with regulators around mistakes made by the AI. Because ultimately, if firms feel they have to actually review everything the AI actually does, it's not certain that they'll actually realize any time savings by using it – making it less likely that they'll be willing to pay for it (beyond what they're currently paying for their existing non-AI technology tools to automate more basic compliance steps).

Morgan Stanley Launches AI Chatbot To Help Its Advisors (Not Give Advice To Its Clients… Yet?)

It's been nearly a year since ChatGPT took the world by storm, virally growing to tens of millions of users in just a few weeks as the world experienced what it was like to chat with an AI tool that chatted back in a remarkably intelligent-sounding manner. Almost as soon as people began to 'talk' to ChatGPT and watched it reply back with answers that sounded more convincingly human than any other technology to date, the ideas began to flow about what applications ChatGPT could be used for. Which due to the technology's nature included not only the typical kinds of repetitive tasks that tech automates, but also increasingly complex 'intelligent' functions that had been done by paid professionals – from preparing legal briefs to potentially delivering financial advice.

In the months since, though, it's become clear that relying on the current generation of AI tools to actually do 'intelligent' work may be overstating the tools' capabilities. Because ultimately, the great breakthrough of ChatGPT wasn't that it had studied bodies of knowledge or learned how to apply them intelligently – it was in using Large Language Models (LLMs) to string together words in a coherent way and weave information from a large number of sources together into a single consolidated response. A response which, while not written by a human, sounds as though it was created by one.

But the combination of being trained to use language well but not to fully absorb information in specific subjects often gave ChatGPT and other LLM-based tools the air of, at best, a smooth-talking amateur opining confidently on subjects where it is far out of its depth, and at worst,a 'moody teenager' that tries to break up marriages. Since early in the ChatGPT era there has been awareness of its 'hallucinations', where what the AI says may sound intelligent but lacks any basis in reality – the most famous case being when an attorney asked ChatGPT to create a legal brief, only to later find (or rather to be told by the judge to whom the brief had been submitted) that it had invented fake legal citations. ChatGPT had added them simply because it had decided that that part of the document would be a good place to include a citation, and made one up to fit the need.

Still, while tools like ChatGPT may not have fully internalized deep knowledge domains like the law, taxes, or financial regulations well enough to be trusted to apply them, its language training can make it useful forsummarizingbodies of text with technical language so a human isn't forced to read through original source documents to gain enough understanding on a topic to, for example, answer a client's question. Which can make it a useful tool for someone giving advice, even if it isn't the source of the advice itself.

So it's notable that this month,wirehouse broker-dealer Morgan Stanley has launched an AI tool in its wealth management division, which notably doesnotinclude advice delivery to consumers on its list of features. Instead, the tool – built by ChatGPT developer OpenAI – focuses on 3 core use cases designed to lift the productivity of Morgan Stanley's advisors themselves:

First, the tool's current main feature is its ability to tap into Morgan Stanley's database of investment research (encompassing around 100,000 research reports and other documents), allowing advisors to type in a question about something they want to research and get not just a list of documents to read, but a response that weaves together information from a range of research materials, consolidated into one expedited response that advisors can then use in their own response to clients.

The 2nd case, which is still in pilot mode, is using the AI tool to listen in on client meetings with the client's permission, allowing it to capture meeting notes (which it can presumably then log into the CRM system) and generate a post-meeting summary that can then go out to the client – all of which could represent a significant streamlining of meeting-related support tasks that can be time-consuming for advisors with 10+ meetings in a week.

The 3rd use case that Morgan Stanley, which hasn't launched yet but has been hinted at, is taking information on a client situation and generating ideas around planning strategies that an advisor could then take to their clients. Notably, though closer to the advice realm than the other two use cases, this function doesn't sound as though it truly would give advice itself – rather, it would be more of an idea generator, or perhaps even more accurately an automated checklist, that proposes options for planning strategies (including some that may not be obvious or common, but could be recognized by the AI to take advantage of a client's specific circ*mstances) that advisors and clients can then narrow down to whatever is most relevant to recommend and implement.

As with all AI in new applications, the question going forward is how effective Morgan Stanley's new AI tool will actually be at doing what it purports to do. The new tool may be able to review thousands of pages of investment research, but does it know how to evaluate the merits of different research methods and not overweight research that might be out-of-date? Can it accurately summarize meeting notes, and pull out the most important takeaways, for 1- to 2-hour planning meetings day after day? Will it actually surface the right planning strategies for a client's circ*mstances, or will it be liable to miss strategies that could be important (or worse, send advisors chasing after strategies that aren't right for a client situation)?

These questions notwithstanding, it's interesting to see a multitrillion-dollar wirehouse invest in AI, and worth noting in particular that the core use cases are not to use it for generate advice but rather to support advisors on otherwise time-consuming tasks like research, meeting support, and brainstorming planning ideas. All of which are very human tasks, but if Morgan Stanley's AI can shave off just a little bit of the time it takes to complete each step, the aggregate time savings across its 16,000-odd advisors could very well be worth the significant investment in building it out.

In the meantime, we’ve rolled outa beta version of our new AdvisorTech Directory, along with making updates to the latest version of our Financial AdvisorTech Solutions Map (produced in collaboration with Craig Iskowitz ofEzra Group)!

So what do you think? Would quantifying the depth of your client relationships move your bottom line enough to justify investing in a solution for doing so? Is account aggregation an important enough feature that you would continue to pay for it if the shift to direct API feeds raises the cost? What aspects of your compliance processes would you trust AI to handle? Let us know your thoughts by sharing in the comments below!