About
Most family offices, wealth firms, and family-owned companies have stayed out of AI for…
Activity
26K followers
-
Ignacio Lopez shared thisThe first thing the machine found was nine formula errors. A 4th-generation real estate developer hired me to speed up their investor reporting. Investors wanted it quarterly. It was going out about twice a year, because interpreting a quarter and getting it onto one page took as long as it took. Before building anything I ran the engine backwards, against a quarter they had already closed and signed off on. It came back with nine formula errors. Not in my code. In theirs. Nine places where a spreadsheet built by a real person, and reviewed by another real person, was adding up wrong and had been for a while. Nobody in that room got defensive about it. That surprised me a little. Then their controller did something better than not getting defensive. She looked at how the thing was built and asked me, "The row number doesn't have to be the same across the reports, right?" It doesn't. Rows shift by property. So what happens if nobody says that out loud? Every figure gets pulled off the wrong line, on every property that does not match the template, quarter after quarter. And none of it looks wrong on the page, because a number is a number, and nobody goes back to check a report that has never been wrong before. She caught it on a review call, before a single quarter had shipped. The machine found nine of hers. She found one of mine. That is more or less the whole job. Not automation. Two checks pointing at each other. I feel like most people buy the first half and skip the second one, because the engine is the part you can put on an invoice, and reading what it prints every quarter is just somebody's Tuesday afternoon. If you sign off on reports somebody else builds, when was the last time anyone checked the math underneath them?
-
Ignacio Lopez shared thisBuilt a company with my best friend from kindergarten. Then we stopped speaking for two years. Nobody tells you that part before you start. We met when we were five. Started Plored together in our thirties, raised money, hired people, and for a while it was the best thing either of us had done. Then it closed. The closing was not the hard part. Companies close. What I was not ready for was that the friendship had been quietly carrying the weight of the company the entire time, and when the company went away the weight had to go somewhere, so it went between us. We did not fight, which would honestly have been easier to fix. We just got quiet. Two years of quiet with someone you have known since you were five. We are fully back now. It took time and it took both of us wanting it back, and I do not think there was ever a shortcut available. Here is what I actually took from it. Every company puts weight on the relationships inside it. Cofounders, first hires, the person who joined because they believed you. You are not only building on top of those relationships, you are spending them, and you usually find out the price after. So I run things differently now, small on purpose, with a tech partner I trust when a build gets big and no permanent payroll to keep afloat. People read that as playing it safe. It is not safe. I just know what the other thing costs. If you built something with someone close to you, what did it cost? Curious whether it was the same currency.
-
Ignacio Lopez shared thisBuilt a company with my best friend from kindergarten. Then we stopped speaking for two years. Nobody tells you that part before you start. We met when we were five. Started Plored together in our thirties, raised money, hired people, and for a while it was the best thing either of us had done. Then it closed. The closing was not the hard part. Companies close. What I was not ready for was that the friendship had been quietly carrying the weight of the company the entire time, and when the company went away the weight had to go somewhere, so it went between us. We did not fight, which would honestly have been easier to fix. We just got quiet. Two years of quiet with someone you have known since you were five. We are fully back now. It took time and it took both of us wanting it back, and I do not think there was ever a shortcut available. Here is what I actually took from it. Every company puts weight on the relationships inside it. Cofounders, first hires, the person who joined because they believed you. You are not only building on top of those relationships, you are spending them, and you usually find out the price after. So I run things differently now, small on purpose, with a tech partner I trust when a build gets big and no permanent payroll to keep afloat. People read that as playing it safe. It is not safe. I just know what the other thing costs. If you built something with someone close to you, what did it cost? Curious whether it was the same currency.
-
Ignacio Lopez shared thisThrew away three things I built in three days. None of them were broken. All three ran fine. The last one was a control dashboard for my own business: dark background, monospaced figures, the kind of screen that looks like a flight console and makes you feel productive for about nine minutes. Opened it and said, "I don't like the design." Then I went and opened the two dashboards I actually use every week. Cream background. Serif headings. Boring. Built months ago, still open on my second screen every morning. Never looked at either one before building the replacement. Not once. Why does that matter to anybody but me? Because it is the same mistake I get paid to go find inside other people's companies. Somebody signs for a tool that replaces a tool that was quietly working, nobody opens the old one first to ask what it was getting right, and six months later half the team is back in Excel and the postmortem blames adoption. Third time in three days for me. Same rule every time. Look at what already exists before you build the thing that replaces it. Two pieces survived the wreck and I kept both. A loop only counts as closed when somebody consumes the output, not when the task runs green. And a panel sitting on stale data should show you that it is stale instead of showing you a number. Two useful things out of three thrown-away builds. I don't know if that is a good ratio. Probably not. Before the next tool goes into your operation, has anyone actually opened the one it is replacing?
-
Ignacio Lopez shared thisRead 26 family office job ads this week. One of them asks the same person to evaluate private equity deals and call the plumber. Not a joke. That posting wants ten years of experience and the judgment to kill a bad venture deal quickly. It names Bloomberg and Cap IQ. Same ad, further down: pay the bills, coordinate repairs, book travel, and a stated willingness to do household tasks without dedicated support staff. It is also, in its own words, "the first hire for our family office." That is the job. Not the exception, the shape of the whole category. Some things I found across the 26, all posted between March 2025 and June 2026, every one with a public link: The most frequently named piece of software was Excel. Seven postings. Addepar came second with three. Every purpose-built platform you have heard of appeared once or twice, if at all. Not one posting used the phrase "single source of truth." Not one. But they describe the need constantly, in older language: account reconciliations, ongoing data reconciliation, consolidated reporting, review and validate monthly and quarterly performance reporting. The problem shows up everywhere. Nobody has a name for it yet. Roles with investment in the title are mostly operations jobs. Read the duties rather than the headline and you get onboarding, KYC, capital calls, distributions, data validation, reconciliation. Very little of it is deciding what to own. And the language around scope is remarkably consistent. First hire. Newly created role. Operational backbone. Build and lead the finance function. Every member contributes beyond their title. One posting describes the job as serving as "the operational backbone of the office." That is a person being hired as infrastructure. The honest limits: 26 is a small sample, it skews to New York and to recruiter-posted controller roles, and Florida is thin enough that I would not draw state-level conclusions from it. Sources are public and I am happy to share the list. What it adds up to: these offices are not short of investment talent. They are short of a system, so they keep hiring a person to be one. If you have written or answered one of these ads, I am curious whether this matches your experience or whether I am reading the sample wrong.
-
Ignacio Lopez posted thisTold a prospect what I build. He said he gets pitched something similar most weeks now. He was not being rude. He was being accurate. Anyone can talk to a chatbot for an afternoon and walk out holding something that looks like software. So the offers arrive constantly, they all demo beautifully, and the buyer has no way to tell which ones survive contact with an actual business. As far as he is concerned I am on that list too. Fair enough. So here is what I would ask if I were the one being pitched. Ask what happens when the input is wrong. Not the tidy file from the demo. The real export, with a merged cell, a blank row, and a date in the wrong format. Most of the code in any working system is there for that file. If they have never met it, they have not built anything yet. Ask where the numbers come from. If a language model is computing a figure that ends up in front of investors or a regulator, walk away. Models read and summarize, they do not do arithmetic you can bet on. Code calculates. Anyone blurring those two has never sat through an audit. Ask what happens when they disappear. Not out of spite. They might get busy, or bored, or hit by a bus. If the answer is that the thing stops working, you rented a demo. Ask who has run it on real data for six months. Not a pilot. Six months of month-ends, weird exceptions, and someone in accounting complaining about it. None of those questions are about AI. They are what you would ask anyone building anything for you, which is sort of the point. The demo was never the hard part. It is just the part that got free.
-
Ignacio Lopez posted thisThe compliance officer was not worried about the data leaking. He was worried about sitting in a deposition, being asked what he knew and when, and not having an answer he could defend. He also assumed I would be sitting next to him. That reframed the entire engagement for me. He is the compliance officer at a multi-family office serving Latin American families, and he is the person who has to sign off before any AI touches anything. Once we got past the polite version of the conversation, the fear turned out to be personal rather than institutional. Not the firm could be exposed. More like, I am the one who gets asked, under oath, whether I knew this was happening. Vendors do not sell to that person. They sell accuracy and cost savings to an operations lead. Then the compliance officer kills it, and everyone calls him risk averse. His question was never whether the AI is accurate. It was: what exactly touched client data, when, on whose instruction, and can I produce that on request. Here is the part I did not expect. He solved it himself before I could propose anything. He drew the line at client identity. Anything carrying a client name or an account number gets processed locally, with deterministic code that never calls out to a model. Everything else, the research, the drafting, the reading of public filings, can use cloud AI freely, because nothing identifying ever leaves the building. That is not a compromise. It is a cleaner architecture than most vendors ship, and a compliance officer drew it on a whiteboard. What I took from it: in wealth, the person who says no is usually not being obstructive. They are asking a question the industry has not bothered to answer properly. If you sell into family offices and your pitch has never been read by a compliance officer, it will be eventually. What would yours survive?
-
Ignacio Lopez posted thisMost people prompt Claude. Almost nobody teaches it their actual job. That gap was the whole reason for a workshop I ran. Two weeks before it, four friends got one question from me: name the task you repeat every week, and send me the real file behind it. Not a sample. The actual one. Then I taught Claude how each of them works, and turned it into a Skill. A Claude Skill is a reusable set of instructions Claude follows the same way every time. Yours, built on your data and your steps, not a template somebody else wrote. Prompting gets you an answer. A Skill gets you the same answer at the same quality every Monday. Three of them were built and running before anyone joined the call. Juan Ignacio Guevara Lynch, supply chain: cross-checking four tables every Monday to feed a dashboard. Thirty minutes, every week, forever. Franco Tiberti, analytics: he wanted his results written up in his voice instead of chatbot voice. Manuel Diaz Reynolds, sales: drowning in follow-up across a few thousand contacts. The fourth could not send me anything. Where he works, the data does not leave the building. That turned into the most useful conversation of the night, and it is the same conversation I have with every family office I work with. Two hours, live. Each of them installed a context file first, which is how you teach Claude what their business is, who the clients are, how the numbers work. Then they opened their own Skill on their own data. Juan ran his, and then spent longer than the run itself checking the logic line by line against how he does it by hand. That was the moment the room turned. Not because it worked. Because he could prove to himself that it worked. Franco's crashed on a formatting error, which is the honest half of running this live. Fixed afterward. Nobody left convinced about AI. They left convinced their own Monday got shorter. Same method I use with clients. Bring me the boring thing somebody repeats every week and the real file behind it, and the answer is almost always a Skill rather than a better prompt. Impressive demos convince nobody who has to live with the result. Save this if you have an AI training on your calendar that has slides in it. Thanks to Juan, Franco, Manuel and Ignacio Agustín Aceguinolaza for being the first four. Doing it again in September. Free, eight seats, same format: send me one recurring task and the real file, I build your Claude Skill before we meet, you run it on your own data and decide for yourself whether you trust it. This round is for people running operations or finance inside a family office, a wealth firm, or a family business. Want a seat? Comment with the task you would bring.
-
Ignacio Lopez posted thisFour posts crossed my feed this week making the same argument about family offices. Stop buying headcount. Buy the discipline instead. A staffed office runs about three million a year and most of that is salaries, so build the system and skip the payroll. They are right. Not one of them said what the discipline actually is. I build these systems, so here is the concrete version. One current picture. Not a dashboard. A single place where a number is definitive, with a date on it, so nobody has to ask which version is real. In one engagement the person assembling the quarterly investor report was losing about four days per cycle to copy-paste before the picture even existed. Provenance on every figure. If a reviewer cannot click from the number back to the file it came from, they will not trust it, and they will rebuild it by hand to be sure. Trust is not a feeling here, it is a trail. Exception review. A person looks at what the system flagged, not at all of it. This is the part that gives the hours back. Continuity. The process has to survive its owner being on a plane. Written down, encoded, runnable by someone else on a Tuesday. Now the part the discipline argument skips. A system does not remove the person. It changes what the person does. Someone still has to own it, feed it, and decide what happens when it flags something strange. Get that wrong and you have not escaped key-person risk. You have moved it from a person who might leave to a system nobody maintains, which is worse, because at least the person told you when they were unhappy. The families doing this well are not choosing between staff and systems. They are hiring one person to own the coordination and giving that person a system instead of a team. If you run an office in that fifty to five hundred million range, where is your single current picture actually kept right now? Curious, because the honest answer is usually a spreadsheet with someone's initials on it.
Ignacio Lopez commented on a post
16h
Every time I automate one of these workflows, the person ends up checking the machine's output instead of producing it.
-
Ignacio Lopez liked thisIgnacio Lopez liked thisThe Economist finally caught up to Miami. The city where power came to play is now where power lives. And the numbers in this piece are wild. Miami’s metro economy grew 23% from 2020 to 2024, almost twice the national rate. Central Miami’s economy is now roughly half the size of Manhattan’s. Fifteen years ago, it was a quarter. More than 150 financial firms have moved here. Shout out to Francis X. Suarez for his leadership on this. I’ve lived here almost my entire life, so this headline means something to me. The article is also honest about the next part. Rents are higher than New York. Prices have jumped nearly 40% since 2020. Florida spends less on research than every other top-ten state economy. And a low-lying coastal city can’t shrug off climate risk. None of that makes me less bullish. It makes the assignment clearer. We have to make it possible for working families and young builders to stay, put real muscle behind universities, research and homegrown companies, and build infrastructure and resilience with the same urgency we bring to towers. New capital found a city generations of immigrants and entrepreneurs had already made global. Now we have to make this boom last AND make sure the people who built Miami can still see themselves in its future. I would never bet against Miami or Florida 🇺🇸
-
Ignacio Lopez liked thisIgnacio Lopez liked thisBittersweet as I close the chapter at Kaseya after an incredible 4.5 years. I’m incredibly grateful for the leaders, Andrew and Andrew, who advocated for me, challenged me, and pushed me to grow, and for the technical teams and resources; Loren, Cesar, John, Sara, Armando, Nicole, Daniel, Enmanuel, and Nicole, who spent countless hours alongside me on calls, demos, implementations, and everything in between. Thank you for sharing your expertise and teaching me so much along the way. Most importantly, thank you to my partners. Many of you became much more than customers. You trusted me with your businesses, challenges, teams, and goals. I had the opportunity to learn firsthand from some incredibly talented people in the IT community, and the relationships I built are what I’ll value most from my time here. These 4.5 years brought wins, challenges, lessons, friendships, and more growth than I could have imagined when I started. I’m leaving this chapter incredibly grateful for the people who were part of it and excited for what’s ahead. One chapter closes, and another is about to begin. Stay tuned.
-
Ignacio Lopez liked thisIgnacio Lopez liked this“I don’t trust him” comes up frequently in family business work. The obvious next question is, “Why don’t you trust him?” I’m not sure that gets us very far. Asking “why” often invites people to explain and reinforce the conclusion they have already reached. The explanation usually locates the problem in the other person: what he did, what he failed to do, or what kind of person he is. In my experience, blaming blocks problem solving. It makes it harder to see the situation systemically. What has been happening between these people? What expectations were unclear? What decisions or commitments broke down? What patterns have developed over time? So, when someone tells me, “I don’t trust him,” I usually want to get more specific: · What are you no longer willing to rely on him to do? · What happened that changed that expectation? · What would you need to see happen differently? · What would you need to observe repeatedly before you began to revise your judgment? Perhaps the issue is that he doesn’t follow through on agreements. Perhaps consequential decisions are made without you. Perhaps you hear one thing while employees hear another. Now there is something concrete to work on. Our judgments about whether to trust someone are shaped by what we observe them do over time. Advisors can help people identify the behavior they would need to see without deciding for them what they should eventually feel. A family may still disagree about the past or remain less close than it once was while becoming more reliable about sharing information, honoring decisions, following through on commitments, and raising concerns directly. If those behaviors change consistently, people can decide for themselves whether their judgment about trust should change too.
-
Ignacio Lopez liked thisIgnacio Lopez liked thisAccording to Carta, 93% all pre-seed rounds in Q2 were structured as Simple Agreements for Future Equity (SAFEs). So what exactly is a SAFE? A SAFE is a simple agreement for future equity. With a SAFE, an investor gives you capital today with the promise of getting equity in the future when a conversion event occurs, like raising a priced round. Carolyn Levy, a former partner and lawyer at Y Combinator, invented the SAFE in 2013 as a cleaner alternative to convertible notes (which are debt instruments that accrue interest and require hefty legal docs just to raise often small amounts of early capital). There are some huge pros: → Close capital quickly (in days, not weeks) → Standard document = minimal legal work and fees → You can raise rolling tranches from different investors But also things to be aware of: → Multiple tranches at different valuations = messy cap table → Uncapped SAFEs or discounts can dilute you way more than you expect → If the market shifts, conversion terms can get ugly SAFEs are great for raising small amounts from believers early on, and ultimately the structure you pick matters less than understanding the implications! Happy Fundraising :) 💜
-
Ignacio Lopez liked thisIgnacio Lopez liked thisLong-term incentive plans that die on transition Your long-term incentive plan is a bet the candidate has not prices in yet. Its not retention. I’ve watched senior investment hires accept packages built on carry, phantom equity, or deferred comp without asking the only three questions that determine whether the money is real. What happens to this if the principal dies? What happens to this if the family sells the office or winds it down? What happens if I’m let go in year four, one year before the cliff? A package that looks strong at signing can be worth nothing by year three if the family transitions. I’ve seen exactly that outcome. The executive did good work, the office restructured, and the long-term piece evaporated because nobody had defined what a change of control meant in a structure with no shareholders. For principals, here’s the uncomfortable part. If you cannot answer those three questions in writing today, you don’t have a long-term incentive plan. You have an expression of goodwill, and your best people will discover that at the worst possible time. The offices getting this right are defining the mechanics before the search starts, not negotiating them under pressure with a candidate who has a competing offer. The ones getting it wrong are the ones telling me they can’t understand why the CFO left for a smaller office with a smaller number attached. Write it down. Vesting, clawback, and what happens when the family’s circumstances change. That document does more retention work than another fifty thousand of base.
Experience & Education
Licenses & Certifications
Languages
-
Español
Native or bilingual proficiency
-
Ingles
Full professional proficiency
Recommendations received
4 people have recommended Ignacio
Join now to viewOther similar profiles
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content