How to 1000x Your Deal Flow: AI-Driven Self-Storage Underwriting with Noah Starr

THE SELF STORAGE REPORT — FULL TRANSCRIPT Guest: Noah Starr (Co-Founder & CEO, TractIQ — self-storage data platform; institutional background; new dad) Host: Chris Berg (Abernathey Development — "The AI Storage Guy") Topic: How TractIQ's new AI connector (MCP) brings its data into any AI workflow (Claude, ChatGPT, Excel); live demos of market analysis, model population, and an "achieved-rent" predictive model; Chris's own four-agent CRM/deal-screening system built on TractIQ data; the thesis that AI lets operators 1000x screening and refocus on human superpowers; the 1960s–70s stagflation rate comparison. Notable references: TractIQ AI connector launched ~1 week prior; upcoming TractIQ webinar June 18, 1pm Central (customer builds showcased; deal pitched by Austin McCloud). Chris built his CRM + agents with Replit + Claude + Whisper (voice). Recorded: June 2026 Source: original recording transcript (speaker-labeled) ======================================================== Chris Berg (00:47): Today we're talking about how technology at TractIQ is transforming — how's that for an alliteration? — the way you underwrite deals, to save you time. Welcome to the Self Storage Report. I'm Chris Berg with Abernathey Development. If you've got land in California or Arizona you think might work for storage, definitely hit me up. And subscribe here to the YouTube channel to be the first to know when we put out these pods. Joining us today, the CEO of TractIQ, Mr. Noah Starr, doing incredible things with tech. If you're not familiar with MCPs, we're going to dive into that, and maybe show you some things I'm building as well. Noah, happy Thursday. Great to see you, my friend. Noah Starr (01:29): Likewise. You're not sick of me already? Chris Berg (01:31): We love having you on. We're going to get into tech, and I want to share some things I'm building because of what you've put out. But before we do — let's talk about things that actually really matter. I want to give you a chance to touch on this great post you put out about your daughter and how it changed your perspective on what you're up to in life. I'd love some more color and context on why you wanted to put that on LinkedIn. Noah Starr (02:05): I appreciate you bringing it up. It's never been easier to put out content — never been easier to put out AI-generated content — and what I've found is when I overthink a post, it tends to be over-analyzed and over-AI'd. In that moment, when she was just sleeping on me, being so cute, I put her to bed and wrote this personally, then figured it might resonate with people, especially on the weekends. The whole reason we're doing all of this in storage is to provide for our families and make sure they're okay. Being a parent has changed me a lot in the last 13 weeks — this is just one tiny post about it. Chris Berg (02:57): That's what I want to get at. How has being a new papa — congratulations, by the way — impacted you and your perspective? Noah Starr (03:06): Everything's fundamentally changed for me. I used to think about my life in terms of my career and wife, and everything was built around TractIQ — the company, the employees, the mission. It's fundamentally reset, where now everything's built around Charlotte. I was saying to my co-founders and friends: I could not start a company today. I would not want to, because I just want her to be okay, and starting a company is really challenging mentally, physically, and spiritually. So I'm glad I did that before becoming a parent, especially with a newborn. Lots has changed in the last 13 weeks. Chris Berg (04:02): Congratulations. Being a new dad — hopefully you're sleeping, but I remember those days, walking around like a zombie: okay, where am I, who am I? Super happy for you. Before the tech piece, I want to touch on your take as someone from an institutional background. I had a great conversation earlier this week — I don't know if you've been watching the ten-year — and I put together this graphic overlaying roughly 1967 to 1983 as the top line, and then now, 2015 to April 2026. It's roughly a 2% spread but following along. If this were to bear out to the worst-case scenario — inflation is tracking a similar way right now — you're looking at, what, a 13% ten-year? We don't know where it goes. But as an institutional investor, and I've talked to a lot of brokers lately who say activity's picking up — I wonder if some of that is people looking at this going, "my cap rates might start to rise, diminishing my asset's value, so I'll just get rid of this now." What's your analysis? Noah Starr (05:27): I haven't looked at this specific relationship between 1967 and 2015-present. There are lots of things you can compare that might look similar but have no correlation. I haven't thought enough about this one in particular. If you look from 2023 to present, rates have been really steady — pretty consistent in the low fours on the ten-year. Lots of geopolitical and economic factors at play, but rates have been consistent for three years. Obviously, if inflation and rates continue to increase, that puts lots of pressure on the ability to buy properties at today's cap rates — buying at negative leverage, we all saw the pains of that over the last five years. Back in 2022-2023, even the notion that rates would increase put a lot of challenges on investment companies raising capital. I'm hopeful that's not the case — we don't want anywhere close to 15% interest rates. Warren Buffett always says the best investment is in America, and that's been true for a hundred-plus years, so hopefully we don't see the same trends as that inflationary period. Chris Berg (07:22): It's interesting to watch. There's been this 40-year run of rates going down where a lot of real estate investors might think "I'm a genius," when part of it might just be you bought at a good time and rates falling made it easier to make money. Everyone I talk to lately says those days are probably over — if you're going to win now, you have to be a really good operator. Do you concur? Noah Starr (07:45): It definitely helps to be a good operator. A maybe-contrarian take: the quality of operators today versus five or ten years ago is night and day. The average-to-good operator today is materially better, and technology continues to pave the way. So the really good operators will still have an edge, but that gap is narrowing as the whole industry has become more institutionalized. Chris Berg (08:18): Let's jump into tech — I'm assuming you'll say maybe this tech can help. You look at what Public Storage is doing — we talked last time about Welltower — to increase NOI. I'm presuming you've got a presentation on some of the tech you released. I want to say thank you and acknowledge you for what you're doing at TractIQ, because what you gave me — an early version to play around with — what I've been able to build is blowing my mind. I had no idea it was even possible, and I'll jump into some of that. But I want to give you the floor first. Noah Starr (09:04): We just released an AI connector last week, which connects all of TractIQ's data into any AI workflow of your choosing — Claude, ChatGPT, Excel, PowerPoint, or on your phone. You can now talk to TractIQ's data. Let me share my screen. Historically, in TractIQ you export all the data to Excel and analyze it there. Now you can bring all of this data into your workflow. Let's anchor around this Extra Space here — say we want to look at this 1000 Main Street deal, run a market analysis. I'm going to open Claude — I don't work for Anthropic, but I guess effectively now do, because they're amazing. I just typed, "Can we run a market analysis around this address?" You can see Claude is calling our market summary tool, our pricing tool, geocoding the address within Claude, and it'll build the analysis. While that's going, I also have my Excel model — a fully dynamic, institutional-grade model I built — and I asked it to pull the rates for that same address into the model. Chris Berg (12:11): Can I ask questions while you do this? In Claude, is this an agent — is that why it's spitting out this dynamic graphic? Or is it a simple chat? Do you have specific instructions? Noah Starr (12:27): Over time Claude kind of knows how I want it, but I literally just asked it to run a market analysis using the AI connector, and it pulls certain tools that provide structured data we've deliberately built for Claude. You can see it pulling the downtown number of operating facilities, rentable square feet, supply per capita, number of climate-controlled, demographics, percent of the market that's REIT versus non-REIT, some stuff under development — and here, rate trends within a three-mile radius of this address. All pulled while we've been on this call. And it's still doing demographics. So in 60 seconds, Claude has basically pulled all the data to analyze this market — think how much faster I can move. Meanwhile, as we talked about 60 seconds ago, it populated all the rates for North Main Street in LA into my Excel model — my entire rate-analysis tab, all within the last three minutes. Chris Berg (14:06): It's stunning. What are you solving for in your underwriting, and can you show what it came up with? Noah Starr (14:15): For this tab in my model, I was trying to understand the 10x10 climate and non-climate rates in this area, and the discount percent between them, so that in my institutional memo I have a rate analysis. Our AI connector in Claude pulled all the data from this address after geocoding it and populated my model just like that. Three or six months ago I'd have had to export the data to Excel and manually paste it. Think of the time savings and the speed to quality insights — it's genuinely unmatched. Chris Berg (15:12): You're doing a lot with the MCP, Claude, and Excel. Can you just talk to it and say, "give me a stress test if rates go to X or cap rates go to Y," and it'll do that automatically? Noah Starr (15:27): Let's do it. I'll say: "Can you help me understand how interest rates over the last five years compare to the 1960–1975 period using the TractIQ AI connector, and estimate what rental rates could look like the next five years — please back-test this and look at cap rates." While that's going, Chris hit me with that chart on the fly — I have no idea what we'll see. It's going to pull in the Fed funds rate, some surveys, clarify some stuff. Chris Berg (16:50): Just so people know — the chart I showed you was built by Claude. I just said, "Claude, give me this comparison." It's incredible what it's doing. Noah Starr (17:00): While it works on that, let me demonstrate something else. Here, in Claude, we've got a demographic analysis of this entire market. In the background — we're in LA — I also asked, "Could we build a predictive achieved-rent model for this market based on all our CMBS occupancy data and CMBS achieved-rate data, and simulate some scenarios?" So we have the initial Claude chat, my Excel model, and now this achieved-rent model — all while it also runs the 1970s stagflation analysis. Claude and I built a model to predict achieved rates, pulling in data from a U-Haul, a StorQuest, a US Storage facility in this market. Depending on the 10x10 standard rate, it predicts what these stores would be based on our data. You can back-test it and build all sorts of things to understand what relationships could exist. Let me pause — I'm seeing really interesting stuff. Let's watch the stagflation one build. I picked LA, a 25-mile radius around downtown, looking at our pricing data with a macro overlay. It's building the last 36 months and a projected 60-month period, with minimal direction from me. Before it shows the answer — Chris, what do you think it'll come up with? Chris Berg (19:36): Regarding rates, utilizing the Fed funds rate? Noah Starr (19:42): If we go back to the prompt — understand how interest rates compare to that period you showed me, and what rental rates could look like the next five years. So it'll do a rental-rate estimate for the next 60 months. Chris Berg (20:01): My guess, if it's pulling accurate data — because I believe inflation is going up — it'll show rental rates trending more than 3% upward. Unless it factors that LA is watching the "stagflation of the 70s" and lowers rates. No — they're projected up. That's what I anticipated. Noah Starr (20:22): Let's see. We've got the 10x10 street rates for climate, non-climate, and the 5x5, and it's forecasting up and to the right. It shows what the Fed funds rate was from the cycle start and how cap rates have moved over the last number of years. Here's the initial conclusion from Claude: it's a milder version of the seventies, street rates are tracking the textbook stagflation pattern. We can ask a thousand more questions and get really deep — that's the point. We're empowering everyone with this data to derive insights that have never before been possible. Chris Berg (21:27): It's been unbelievable. If you don't mind, I want to share what I've been building, and give you a chance to poke holes in it — this is the first time I've done this, so interrupt me anytime. I'm building four agents right now: one for initial analysis, one for zoning, one for scoring, and one for underwriting. One thing that's been really powerful — I built my own CRM. And I want people to realize I know nothing about tech. I'm using a thing called Replit and Claude. I built our CRM, and I built an MCP for our CRM. Why does that matter? Because now my CRM can talk to Claude and communicate back and forth with other tools I'm using. This is our dashboard — it's how we track and monitor potential deals. I've got a "dead" column way over on the far right. Why this matters is it lets me compound. Right now I've got roughly 1,000 deals — 450 dead, 284 in source, some in screening — but once this is at 10,000 deals, the beautiful thing is, because of the MCP, I can literally talk to my computer and re-underwrite deals every three, six, twelve months, and see if something that didn't work 12 months ago works today. There's a compounding effect, and I'm also helping Claude get better every time, so it gets more refined at helping me underwrite the best deals in the country — even though we're focused on California. So that's the CRM and the compounding piece. Have you ever used Whisper, Noah? Noah Starr (24:13): The voice app — yeah, I've used it. Chris Berg (24:15): That's what I use. So — keeping it super simple today — I'll say: "Go to the AH CRM. Grab two sites from the source column that you've not previously run through Agent 1, run them through Agent 1, and set it up so I can copy and paste the instructions into Agent 2." It goes to the CRM with the MCP, grabs two sites it hasn't looked at from the ~250-300 in source, then runs — because of the instructions — searching TractIQ and using TractIQ information. What's powerful is I can run a list of 10, 25, and as it gets better, 200 sites, throw it into Agent 1, and it does its initial run. Noah Starr (25:41): It's really cool — wait, scroll down. So now it automatically knows to geocode using TractIQ, it's pulling demographics from TractIQ, it's pulling pricing — and you built your own deal engine on top of our data so you can have your own agents underwriting sites. Chris Berg (26:00): Exactly. Agent 1 gives me a brief demographic analysis, then a copy-and-paste I put into Agent 2 — my zoning agent, which does a thorough job on city documents and zoning. Eventually this will all be one button. Here's site one — a 24-month CC blended rate of $219 per TractIQ. Now site two. I don't know off the top of my head how big these sites are — this is all first iterations — but it gives people an idea of how powerful TractIQ with the MCP is. Whether I've got two, 10, or 200 in here, while it's doing this, I'm talking to brokers, sending emails, shaking hands and kissing babies, while my tech does the work. Noah Starr (27:24): To drive the point home — you can do this from your phone. Here's the demographics analysis of LA we did together, the pricing trends. You can leave your computer, go to a lunch meeting, and have it continue to run on your phone. Why is this so important? You can 1000x not just the deals you look at, but the quality of information brought initially. Five years ago, initially screening a deal, you might look at four or five things. Now you can bring all that deep underwriting data immediately, so if you'd have missed a site five years ago because you didn't have time to understand the dynamics, now you can understand them much faster — and focus on your superpower. Chris, your superpower might not be building memos and models. It could be developing relationships, meeting with owners and brokers — the human aspect that requires your presence, your brain, your integrity. Not how great you are with Excel keyboard shortcuts. Chris Berg (28:58): Thank you for pointing that out, because Excel is not my jam. Claude makes it much easier, but I'm not going to sit and do a bunch of formulas — that's not my jam. Talking to people, I love doing that. I'm kind of embarrassed I didn't know this about myself, but I've loved building stuff. Here's what's awesome: I copy this, throw it into Agent 2 for the zoning aspect. And watch — every time I do this: "Can you make sure you put these notes inside the CRM with each site card, and confirm that's been done?" Now all these notes are in the CRM forever. So if I go back six or twelve months from now and a site didn't work today but is still listed or off-market, boom — it's all there. I just talk to my computer: "Re-underwrite 2325 Statham Boulevard," and it'll re-underwrite it, redo the zoning, whatever I want. Then I run it through Agent 3 — a scoring rubric we built, thanks to Noah's tech. I've got it set up so if a site scores 50 or more, it automatically moves from source to screening for a deeper dive. Just for fun — this is a site-scoring rubric. Anyone watching with feedback, I'd love it. You put an address in, hit look up — thanks to TractIQ — and it fills in the fields and scores. I did this one yesterday and it was an 80. There you go — 24-month blended rate, projects in the pipeline. The zoning stuff isn't accurate, so that's not really fair, but the point is building scores to sort sites more efficiently. The rubric gives rental rate 40% of the score, supply scarcity and the demand score 30%, market demand 10%, entitlement risk 15%, plus some contextual bonus like high-traffic corner. It's really built for California because rates are so high — if the 10x10 CC 24-month blended rate is $260 or more, that gets 21 points; down at $185, our floor, it gets 9. Any feedback, Noah? Noah Starr (32:58): It's amazing. If you're watching, you might say "this is a little overwhelming" — but simplicity is the ultimate sophistication, and that's what Chris is doing: simplifying his own process for underwriting and screening so he spends time on the right opportunities. What's incredibly transformative about this era is you can build whatever you want. TractIQ no longer tells you precisely how to use the product — our platform is the home base, really simple to navigate, an amazing way to get data on any address in the country. But then we empower you with all that data to build whatever you find interesting. That's the new differentiator. Chris can screen all the deals in the country significantly faster — something that historically would have required a million dollars a year in personnel or technology. And here's the edge question: once Chris underwrites 1,000 deals, what about the company that's underwritten 5,000 or 10,000 over three years — what edge can this technology give them? As groups lean into all of that, we'll see really interesting insights and activity across not just self-storage but the rest of real estate. Chris Berg (34:48): Thank you for what you've put out, Noah. Again, I know nothing about tech — I just talk to my computer and say, "Hey Claude, if I want to do X, how do I do it?" If you're not familiar with Replit, check it out if you want to build stuff. I use Whisper, talk to Replit, it builds it, I don't know how. Yesterday, an example: I may have unlocked something for finding off-market deals. Claude said, "I'm going to run Python code on your terminal" — I don't know what that is — and next thing I know I had a full list of off-market deals in the snap of a finger. It took your information, geocoded it into clusters, and analyzed what could be viable — "these are the ones I'd go hunt down." Blew me away. I don't know how viable it is yet — first trial — but if it's as viable as I think, holy cow. Any closing word? Noah Starr (36:11): In TractIQ, build your custom prospect list, save it as "Chris's off-market deal list," and then in Claude say, "Can you analyze Chris's off-market deal list I just built in TractIQ and score every property most likely for me to acquire based on my criteria?" I'd encourage you to check that out. The point is we want to empower the industry with incredible information wherever they work — and that's now the case. Once you use it and realize what you can build, I've already seen unbelievable things our customers have built — CRMs, deal-screening tools, email automations, all sorts of things to have this incredible data everywhere. That leads to so much better decisions. Even the 1960-vs-75 analysis would have taken me a month to understand; in 10 minutes we put out something pretty dangerous. I'd encourage folks — we fear what we don't understand. Spend an hour talking with Claude or ChatGPT and try to do something with your job: ask it to write an investment memo, write a model, do something legitimate you do every day. Your eyes will open up in a meaningful way. Chris Berg (37:53): You've got a great webinar coming up — tease us a little. Other than what we saw today, what are two of the coolest things you've seen built? Noah Starr (38:05): That's what we'll highlight on our next webinar — June 18th at 1pm Central. We'll highlight incredible customer builds, plus a deal pitched to us by Austin McCloud, and some other surprises. It'll be awesome — and we should work together separately on you presenting some of what we walked through on this call. Chris Berg (38:40): Reminder — if someone out there is building, they can submit to you, correct? Noah Starr (38:47): That's right — email support@tractiq.com, or me or Hayden or anyone on our team, and send a link or video of what you built and we'll take a look. Chris Berg (38:59): Submit by the 5th, and the webinar's the 18th — you can register. With all that said, because of your tech brain: what would you do to improve my process based on what you saw today? Noah Starr (39:16): You have a pretty good process — I'd have to spend more time off the cuff, but you're in the top 10th percentile of customers I've seen, already having built an incredible quasi-CRM with agents — I imagine those are Claude skills you've built with a specific purpose, so you have an effective strategy and process. The next layer is probably: okay, all the lights are green — yes, yes, yes — what happens next? You want to build a memo, a model, a full package to send to a lender or investment company: "I'm raising capital for this project, here was the process, I back-tested this deal across 500 others I underwrote, and this is the best one for all these reasons." That's probably your next step. Chris Berg (40:23): One last button on this for people — the compounding. There's a lot of conversation about how AI will let you anticipate what's coming. It'll get to the point where I press a button in Cowork, Noah, and run all four agents at once — Cowork pulls a list itself of 10, 200 sites, runs them through all four agents and the CRM, and it's sitting there for me to go get lending. And then the idea is it self-develops to where it gets predictive: "Chris, based on the last 1,000, 10,000 deals" — kind of what Welltower is doing with Public Storage — "here's where you should be looking over the next 6, 12, 18 months." It's fascinating to watch. Thanks to Noah Starr and TractIQ. If people want to sign up for TractIQ, how do they do it? Noah Starr (41:26): Go to tractiq.com, come chat with us — we'll find a plan that works for you. Every plan now includes AI connector access. Looking forward to seeing the excitement across the industry and empowering everyone with great data. Appreciate you having me on, as always. Chris Berg (41:45): Great to have you. This is the Self Storage Report, I'm Chris Berg, that's Noah Starr. Please subscribe to the channel to be the first to know about the content we're putting out. We'll see you back here again soon on the Self Storage Report. Noah Starr (41:59): Thanks, Chris. [END OF TRANSCRIPT]