Worry About Spreadsheets, Not Robots

By Chris Ainsworth, Chairman and CEO, Pave Finance

I've been in this industry long enough to have seen the wreckage of the dot-com bust, watched the financial crisis unfold in real time, and sat through more "AI is going to change everything" conversations than I can count. Through it all, I've seen technology change in ways that were vastly different than the consensus predicted.

So let's talk about AI, but not the AI you have probably read about.

Not the one that's allegedly going to render analysts obsolete and "disrupt the entire asset management industry," that leaves advisors scrambling for relevance. That AI gets plenty of column inches and breathless LinkedIn posts but it doesn't move the needle, at least not for our standards.

I want to talk about another type of AI, a quieter one, that is frankly more interesting and more effective.

Let's start with a scenario.

Part I. The Friday Afternoon Problem

It's 4pm on a Friday afternoon and a CIO at a $3 billion RIA is at his desk. Let's call him Jim. Jim's got an Ivy-League MBA, 20 years of experience, and a market view that's genuinely sharp. He can talk about duration risk, factor tilts, and the geopolitical implications of semiconductor supply chains for 45 minutes without notes.

Jim is also, at that moment, manually rebalancing accounts on a spreadsheet. 50 strategies. 300 advisors. A compliance queue. He will be at his desk until 9pm tonight, just as he was last Friday, and will be next Friday.

It's not that Jim lacks talent or experience, his credentials and results speak for themselves. The problem is infrastructure, the kind that implements machine learning at a level where it matters, at the portfolio level. The social media posts and dinner-table soundbites getting all the attention revolve around chatbots writing your client emails and LLMs summarizing meeting notes, not to mention the companies selling "GPT wrappers" at a premium. Those tools are fine, but they miss the point. The real opportunity, the one sitting right under the industry's nose, is the kind of AI that handles the Friday afternoon spreadsheets so that Jim can actually be a CIO.

Part II. The Squeeze

Here's the situation most RIAs are facing right now.

Fee compression has been grinding advisory margins for years while the cost of running an investment operation, from hiring analysts to third-party managers, has certainly not gotten cheaper.

Meanwhile, client expectations have only increased as the days of calling 60/40 models "personalized" are over. Today's clients do not want a model portfolio with a couple of tweaks anymore, they want ESG screens, religious preferences, generational tax planning, directed holds, all of it.

People should be able to invest according to their values and preferences, regardless of their tax bracket. That's what this industry exists to do, to deliver on client expectations. But meeting these expectations at scale, with margins already under pressure, required tools that didn't exist.

So the industry defaulted to one of two answers: hire more analysts — find more grads, put them on a desk, throw money at the problem; or reduce personalization — run model portfolios, water down, shrink the menu.

Personalization at scale was in and of itself an oxymoron.

That said, there is a third path that doesn't require more headcount or compromise, but an understanding that not all AI is the same and that the industry hasn't caught up to that distinction yet.

Part III. The Right Robot for the Job

Something I find myself saying a lot these days in boardrooms, conferences, and to family members is that there are two fundamentally different kinds of AI, and the wealth management industry is obsessing over the wrong one.

First, there's the generative AI: Large language models (the ChatGPTs, Geminis, and Claudes of the world) that dominate headlines. These are remarkable tools for certain things like writing, summarizing, drafting, and searching for dinner recipes. They are useful for note-taking, client communication, marketing copy, and I'm not here to dismiss them.

But here's the problem: they're black boxes. We give them an input, they give us an output. For the most part, the reasoning behind how or why they got to an answer is invisible. Sometimes LLMs hallucinate, make errors, produce confident-sounding answers that are wrong, which is manageable when it comes to drafting an email or LinkedIn post. But when it comes to managing a client's money? No way. One simple slip-up becomes an existential crisis when you are managing a client's retirement portfolio and a regulator asks you to explain every decision you've made.

You can't explain a hallucination to the SEC.

That's where the second kind of AI comes into play. This AI is what the quant world has been quietly building for decades: mathematical, predictive, traceable models that analyze data, produce auditable recommendations, and leave a documented paper trail.

No black box, no hallucinations, every output is defensible and every decision is verifiable.

This is the AI that belongs in portfolio management. Not because it sounds impressive, it is probably less glamorous than the other kind, but because it works. Every decision it makes can be shown, traced, and defended to your client, compliance team, and regulators.

The future of AI in investment management is not the LLM, it's the evolution of quant models into active, personalized portfolio management. Technology that is explainable, trackable, regulatorily sound, and accessible to firms at any size.

Part IV: Be Your Own CIO

Institutional-grade investment infrastructure used to be reserved for the biggest desks, wirehouses, and hedge funds. If you were running a $2 billion RIA and wanted to internalize your CIO function you were looking at OCIO fees of 25 to 50 bps on top of your advisor fee—Bloomberg terminal subscriptions, a minimum of one or two dedicated analysts, and a compliance operation built and maintained entirely by hand. The overhead for most firms below $5 billion was simply unaffordable. So the solution was to outsource, simplify, or go without.

Under the traditional third-party manager model, RIA gross revenue as a percentage of total client fees sits at approximately 66.67%, meaning roughly a third of what the client pays flows to outside managers before it ever reaches the firm. With a flat-fee structure that replaces that model, that number moves to approximately 98%. Not because the client pays more but because the firm stops paying it out.

That's not a rounding error you gift your spoiled nephew, that's a business move.

The infrastructure that once belonged exclusively to the largest institutions has finally caught up to the ambition of the CIOs who always wanted to operate at that level. What changed is the infrastructure available to build it: a proprietary optimization engine, AI-powered research across thousands of securities, compliance documentation that runs itself. The gap between what a $100 million firm can do and what a $20 billion firm can do is closing.

Part V. What Optimization Actually Looks Like

I know how it sounds. Optimization engine and institutional-grade infrastructure are exactly the kind of buzzwords that sound impressive in a conference presentation but rarely translate to something you can actually use on Monday morning. So let me show you what this looks like day-to-day.

Custom indexing and screening: Direct indexing across 17 indexes, with ESG constraints, religious screens, or any client-specific exclusion applied at the account level automatically. What used to require a manual review process—cross-referencing a screener, flagging exceptions, following up with an analyst—is now handled by the engine.

Every account gets a portfolio that is benchmarked, optimized, and screened to its own specifications. This is not about tweaking a model anymore. It's what allows advisors to construct genuinely individualized portfolios for every client they serve.

Tax management at scale: Tax-loss harvesting and concentration management running across every account simultaneously. The advisor sets the parameters, the platform executes.

A client with a large inherited position in a single stock gets an optimization that manages around that concentration without triggering an unnecessary tax event. A client managing around a specific tax liability gets a portfolio built to accommodate it.

And you never have to open a separate account. Any type of workaround becomes obsolete when the solution is built directly into the platform.

Research that actually scales: The platform analyzes each security across 156 data points on a weekly basis, running predictive analytics to inform portfolio construction. What used to take an analyst, or a few, several hours of work running models and updating spreadsheets now happens automatically, continuously, and at a scale that no human team could match. Those recovered hours are not spent on emails or admin. They are spent on strategy, judgement, and the heart of this job: time with clients.

Compliance that works in the background: Every recommendation the platform makes is tracked and retained on a rolling six-year-plus-one-day basis. Every IPS change logged, every trade documented. The CIO still makes every meaningful decision (what to buy, what to avoid, how to weigh the portfolio) while the platform removes the operational overhead of documenting it. If a regulator ever asks why a particular security was included in a particular account at a particular time, the answer is already there, organized and defensible. The CIO did not have to build that infrastructure, it came with the platform.

This is the transparent, traceable AI I was describing earlier. Not a black box. A fully auditable, documented system that can defend every decision it makes, which is exactly what regulators require and what LLMs cannot provide.

Part VI. The Firms Getting This Right

The firms seeing results here are not the ones with the biggest technology budgets or the most advanced IT departments. They are the teams who are sophisticated enough to have a real investment philosophy, nimble enough to actually change how they operate, and hungry enough to care about the margin difference between 66% and 98%. For these firms, owning the CIO function is not a technology decision but a strategic one.

Think about this example: A firm has recently acquired an RIA and needs to transition hundreds of advisor templates simultaneously (think complex, multi-strategy infrastructure, across thousands of portfolios) without disrupting a single client relationship or workflow. Pave migrates the strategies, automates the rebalancing logic, and rebuilds the model infrastructure on the platform. No re-custodiating assets, no manual account-by-account rebuilds. The firm's investment philosophy stays intact, and the Friday afternoon spreadsheet is nowhere to be found.

That is the difference between a technology partner and a technology vendor: one adapts to how you work while the other asks you to adapt to it.

For the firms still growing, the platform grows with you. Start with small-account automation, layer in direct indexing, add tax management and custom screens as the firm scales. The tool matches your ambition at every stage without requiring a painful re-platform every time you cross a new AUM threshold.

Part VII. The Quiet Revolution

Back to Friday afternoon.

Jim still at his desk, working till 9pm. Same as last week and same as next week, unless something changes. The AI revolution that is actually underway in this industry is not the one making headlines. It is not the one that makes for a good conference panel or an entertaining LinkedIn post. It is the one that analyzes hundreds of data points per security per week, documents every portfolio decision for years, and returns an average of 18 hours a week to the advisors who could be spending those hours on strategy and relationships rather than spreadsheets.

It is quieter than the AI everyone has been bracing for, but it is more useful. And if you are still running an investment operation the way it was run five years ago, it is considerably more urgent.

The firms that figure this out in the next 24 months will not be the ones who read the most alarming headlines. They will be the ones who looked at the Friday afternoon spreadsheet and decided there was a better way.

Disclaimer — For Financial Professional Use Only

The information in this article is for general informational purposes only and should not be considered investment, tax, or legal advice, or a solicitation to buy or sell any security. Investing involves risk, including the possible loss of principal.

Pave Investment Advisors, LLC is an SEC Registered Investment Advisor. Securities offered through Pave Securities, LLC, New York, NY. Member FINRA/SIPC.

About the Author

Chris Ainsworth is the Chairman and CEO of Pave Finance, a platform built to give independent advisors access to the kind of institutional-grade infrastructure that was once reserved for the largest firms on Wall Street. Before Pave, Chris spent two decades in senior roles at Deutsche Bank Wealth Management, where he served as the Head of Americas West Coast, and at U.S. Trust, where he led a team of 40 professionals across six offices with more than $16 billion in assets. His experience on both the institutional and operational side of the industry is what led Chris, alongside his co-founders, to building Pave. Chris founded the California Hedge Fund Association and has served on the boards of the Ronald Reagan Presidential Library and the Children's Hospital of Orange County Foundation. He holds a BS in Economics from Southern Methodist University.