Flow two · Market profile
This is the profile of the people Addi is built for, and what they keep struggling with. It comes from the research we ran in early July 2026 on fourteen months of Addi webinars and coaching-call recordings. We read it again this week and added two things: what it says we should talk about, and what it says we should think about building next. Nothing in the feature section is being built. As Mufaro put it on the check-in, "I don't want to up and just jump and build stuff without having your final thoughts on that."
Agenda · Market profile
The short version
Addi's market is the people already around CCL Properties: course graduates, Owners Club members, and the regulars on the Thursday co-living call David runs with Kyle Karges. We listened to 39 recordings, roughly 140,000 words of what those people said. One problem sits above all the others. Before anyone can make an offer on a co-living house, they have to find out four things. Is the house legal for co-living? Does the town actually enforce its rules? What will the rooms rent for? Will a bank lend on it? Done by hand, that takes two to four hours per address on a good day, and two days when the zoning is unclear. Get it wrong and it is a twenty-to-fifty-thousand-dollar mistake at the small end, and a several-hundred-thousand-dollar mistake at the large end.
Addi already answers the zoning question and the time problem, in the towns it covers. The value question, the bank question and the "too many tools" problem are only partly answered. Those are where the next posts and the next features should come from.
1. How we did it
Thirty-nine recordings, from May 2025 through June 2026, in four groups. First, the five Addi sessions CCL Properties hosted: the live demo with David on May 28, the public webinar on June 11 and the team's own debrief later that day, the one-on-one with Marianne and Jeff in Naples on June 18, and the underwriting and compliance webinar on June 25. Second, five guest sessions from the Thursday co-living call: Fernando Corona on DSCR lending on June 12 last year (a follow-up on June 19 was billed to him, but he did not attend), Clara Arroyave on multifamily co-living on July 3 last year, and two sessions on July 10 and July 17 last year that were billed as Miller McSwain, where Miller did not show and David ran the call alone. The people reading the calls caught both mismatches and credited the content to the right person. Third, the Thursday co-living call itself, the one David runs with Kyle Karges, read in four stretches: May to September 2025, October to December 2025 (twelve calls), February to April 2026, and May to June 2026 (seven calls). Fourth, one long one-on-one with Logan Reinard on June 19 about how he works with AI.
A different reader took each recording, or each stretch of calls, and wrote down every moment someone described a problem: the exact words, who said it, what kind of problem it was, how bad it sounded, how often it came up, and whether Addi could plausibly help. Each reader also wrote a plain account of what the room was struggling with and who kept showing up. That produced 679 of these moments, in fifteen batches of notes (one batch per recording or stretch of calls). The readers were told to be honest about limits, and they were. The stretches of coaching calls squeeze months into one pass, so quotes from those are treated loosely. We do not use them as pull quotes here unless someone went back and checked the wording.
Two checks, both by people who did not do the reading. First, a quality check. It took 45 of the 679 moments, leaning toward the five claims the whole study rests on, and went back to the recordings line by line. 43 passed. 2 needed a fix: one coaching-call quote had stitched three separate sentences together, and one note called a human virtual assistant an AI tool. Zero were made up. Then a second analyst was told to be paranoid and attack the first analyst's twelve biggest conclusions. Six held, five were softened, and one was thrown out (the first analyst had said the quality-check files did not exist; they did). Where a conclusion was softened, this document carries the softened version. Section 5 lists what changed.
One thing to remember with every conclusion
Everyone in these recordings is already inside our world: a course graduate, an Owners Club member, a regular on Kyle's call, or CCL leadership. The problems they raised are the ones this group feels safe admitting. Some almost certainly apply to co-living investors everywhere; the zoning problem is built into the business. Others, like how many of them would pay a monthly fee, tell us about our own community and nothing more.2. The people
The recordings show at least thirteen different kinds of person. Here are the ones that matter most for Addi, each with the operator who best stands for them:
| Who | Best example | What they are stuck on | How well Addi fits |
|---|---|---|---|
| The volume operator who is scaling | Kyle Karges (30-plus doors, eight states, buys five to six houses a month) | Tools that do not talk to each other, property management outside his own network, lenders that open and close every few months | A test of whether Addi is good, and a rival. Not a paying customer. If Addi does not do what he can build himself in thirty minutes, he builds it. |
| The mission-driven late-career operator | Marianne, Naples (senior women's housing, "Her Haven") | The words in the reports, fear of touching the calculator, no cash, nine to twelve months to land a nonprofit partner | The strongest case for us doing the underwriting for people. Also the person the tool has to explain itself to. |
| The AI-curious solo operator with several kinds of property | Logan Reinard (flips, rentals, RV parks, moving into senior living) | Understands the ideas but cannot get the tools installed; has to rebuild his underwriting for every kind of property | The high-end customer. A small group, maybe five to ten percent of the room. |
| The warm lead in the trial funnel right now | Terry Smith (Colorado Springs), Mike Perdue (Kansas City, selling his landscaping business to buy his first co-living house) | Speed and what-if scenarios (Terry); underwriting from scratch (Mike) | High. Both were given free fourteen-day trials on camera. |
| The financing specialist between every deal and the bank | Fernando Corona, Julian Rizov | Takes on the operator's problem and says "I don't have a solution for that" at the edge of what the market will lend | Not a customer. His rules are the rules Addi should have built in. |
| The multifamily converter | Clara Arroyave, Linda Holtz (204 units next to a university), Adam Catledge (24 units, New Orleans) | Nobody has a calculator for converting part of a building | Nobody is serving this yet. Clara's session reads like a description of the feature. |
| The flipper crossing into co-living | Derrek Hooyman, Ricky Steele, Michael Riordan | Money locked up for years; the co-living math is new to them | Strong. Michael already built his own deal intake. |
| The realtor or bird dog opening a new market | Jason Crouch (Cincinnati), Nathan Williamson (20 years old, Virginia) | Zoning in a town nobody has tried; whether to trust the report | Strong for zoning. Nathan checked Addi's value against his own comps and it matched. |
| The capital partner who got burned | Jose Avila | Put in money, could not see what was happening on the property | The reader's own note: probably the most willing to pay of anyone in the set. One or two people. |
| The passive-income late-career operator | Jenny Trejo, Mishawn Jackson | "I don't want another job"; drawn to master leases | Needs different words than the mission-driven group. |
| The large portfolio holder who was never pitched | Benjamin Zonis (Baltimore, 120-plus properties) | Unknown; he was in the room for months and nobody showed him Addi | Unknown. A missed conversation. |
| The first-timer stuck in analysis (from last year) | Susan Smiley (November 2025) | Structuring a creative deal, "I'm really gonna need hand-holding" | Her problem is not Addi's; she belongs to the coaching side. Kept because that "hand-holding" line keeps coming up. |
| The operator who needs a text-first tool | Andy Ungert (deaf entrepreneur buying remotely) | Anything that assumes audio | A design rule, not a group of customers. |
One more thing about the room itself
In the seven coaching calls from May to June 2026, five people were told about Addi or given access. About seven others said out loud that they had a problem Addi answers, and nobody told them. David is careful on that call ("I'm not here to share what we built to push our stuff"), which is right for the room. But it means roughly forty percent of the warm leads in front of us never hear about the product. Kyle's own suggestion on June 11, to livestream underwriting with Addi, is the fix the research backs.3. The evidence
Ranked by how many of the fifteen batches of notes the problem shows up in, then by how bad it sounded. Where we say we checked something against the recording, someone other than the reader found the exact words at a named line of the transcript. Every count comes from the research. Nothing here is estimated.
This came up in all 15 of 15 batches. Nobody escapes it. Four questions pile up before anyone can make an offer. What does the town's rule say about unrelated adults, boarding houses and what counts as a family? Does anyone actually enforce it? What happens if you close and then find out you were wrong? And why are the rules different a few miles over?
David told the story on the May 28 demo, and we checked it against the recording: "It's 9,500 square feet, and then there's another little house next to it. That's another 1,100 square feet. 18 bathrooms in this house, plus another 2 in the other house, 20 bedrooms. And nine and a half bathrooms. And then we looked at the zoning. And you can have four unrelated people in the whole house. And the whole deal fell apart." He calls it "a several hundred thousand dollar mistake", on top of the twenty to fifty thousand he says such cases usually cost.
Same session, also checked against the recording, David on Hickory, North Carolina. A city planner three years into the job told him "look, I've been here for three years. No one's talking about it, and I'm not aware of anyone, any of those rules ever being enforced." Two weeks later his team ran the property and found "185 different enforcement actions on it."
Tom Paschke physically drove to city hall to answer one question. That tells us something about what people would pay for. He said on the June 11 webinar: "I have been to the city of Shakopee and asked that. Basically, it's 4 unrelated in general, unless it's in a specific commercial slash residential zone, and then it can be up to 12."
The community's answer changed over the year. In May 2025 the advice was "call the city clerk." By September 2025, after Purvi Shah's trouble in Glendale, it became "don't ask them, work with them." By April 2026 experienced operators were underwriting two or three backup uses per house (sober living, student housing, long-term rental) as insurance against the zoning answer. That last habit is a feature request in disguise. See Section 6.
This came up in 12 batches. David said on the May 28 demo, and we checked it against the recording: "I was spending 2, 4 hours minimum per property before I even made an offer. Just trying to figure out, okay, how do I know if it's legal or not?" On June 11 he put the top end at "4 hours to 2 days." Kyle Karges, Purvi Shah, Rayce Barnes and Dr Tami Romriell all describe the same cost. Terry Smith's reaction to the live demo on June 25: "It's crazy how fast it is." This is the problem Addi most clearly already solves, and the thing people praise most in every demo.
This came up in 13 batches. There are no comparable sales for a nine-bedroom house. Appraisers use a single-family comp or walk away. Every financing problem further down the line starts here.
William Denyko said it in the first real line of the May 28 demo, and we checked it against the recording: "It's really hard to try and uh, lock down, you know, these values for these co-living places. Especially when you've got nothing to compare to. You've got no comparables."
From the lender's side, Fernando Corona told us on June 12 last year: "9 bedroom 9 bath doesn't exist. No living room, right?" And later: "I'm in the middle of trying to help this guy who has a 12 bedroom, 12 bath, and she's like Fernando man, I would love to do this appraisal, but there's literally nothing I can do."
The workaround is now standard practice. On the April 9 coaching call David described a refinance where the appraiser wrote up the keypad locks, mini fridges and coin laundry, six lenders dropped the file and the seventh closed. (That one is from a batch of calls we read together, so treat the wording loosely.) The standing advice now is to strip the locks and door numbers and leave the last bedrooms unwalled until the appraisal is done. That is word-of-mouth knowledge. No tool does it.
This came up in 12 batches. DSCR loans are the investor loans that qualify on the property's rent rather than the borrower's income. They are underwritten on ordinary long-term rent, and a co-living house beats that by a wide margin the lender cannot see. Lenders open up to co-living for a few months and then close. Somewhere between four and eight individual room leases on one house, the property becomes impossible to refinance.
Fernando Corona, June 12 last year: "You want to go 20% down, but the market rent sucks, you can't buy DSCR because the DSCR sucks." In the same session he said he had gone through seven lenders in eighteen months who opened to co-living for two or three months and then cut him off. And: "If somebody has 8 individual leases, I don't have a solution for that. Maybe I could try with 4 leases." Fernando also described how the closers already turn the bedroom count like a dial: "we need this property to be a 6 bedroom. We will literally communicate with the seller and say, buyer needs this property to be a 6 bedroom."
Nathan Yerkes on a 2025 coaching call: "I've called like 15 different ones right now, and haven't found anybody that's willing to take that." (From a batch of calls we read together, so treat the wording loosely.)
This came up in 11 batches. Listing sites, the county map, a rent tool, a room-rent marketplace, a spreadsheet calculator, a CRM and a project board. None of them know what the others know. This is the problem David's pipeline idea speaks to, so it gets a closer look in Section 6.
David laid out his own set of tools on the February 19 coaching call, when Karina Ramirez asked whether the CRM could also manage projects. Leads come into the CRM. His own AI tool does the zoning and market research. Then a manual look. One project board through closing. A second board for managing the property after. Karina's reply, naming the CRM she had hoped would do it all, was flat: it "doesn't do that." Kyle Karges on February 12: "do you know how many things I probably paid for. Like, to do the right thing, and just wasn't organized enough about it, and like, couldn't find it or access it or whatever. And then probably just paid for something else again." (From a batch of calls we read together, so treat the wording loosely.) That is a thirty-door operator admitting he pays twice because he cannot find things.
The same split shows up inside Addi. Mufaro said on the June 11 debrief: "the challenge is having two user interfaces. I'm trying to solve, to make it much easier to spend more time in the GUI interface." People talk to Addi in one place and work the deal in another.
This came up in 10 batches. Managing a co-living house (chores, roommate conflict, cleaners, screening at volume) does not exist as a product outside an operator's own network. In the October to December 2025 calls it blocked a deal on 8 of 12 calls. David cancelled his own Wilmington deal on December 17 last year because his management network was not ready five hours from home. Kyle launched his own management company on December 18 for the same reason. David told Marianne on June 18: "you can take the property manager's job and divide it into pieces. And maybe you would do a piece and we could do a piece." Addi does not solve this. It is here because it is the next thing the customer hits after Addi says yes.
This came up in every Addi demo and in the latest stretch of coaching calls. Elayne Kohan's Phoenix address on May 28 was rescued with "we can add it tonight." Tom Paschke's Minnesota parsonage on June 11 came back at 436 square feet on a 4,000-square-foot house, and he said "Yeah, we know that's wrong." That is the moment trust flips. Karel Johnson learned Florida was not covered on April 30. The other side of this is worth knowing on its own. In the May to June 2026 calls David volunteered the gap three times without being asked ("7,500 of 25,000", "that specific jurisdiction's not in our database"), and the room trusted him more, not less.
18 of the moments we wrote down were about what stops people using the tool, plus the whole Naples session. Marianne on June 18: "if I knew what a GIS was, that's good. Now I have that. Geographic something, something, I don't know. I'm guessing." And on the calculator: "This would make no sense to my brain. I'd be like, I don't know what I just did." Terry Smith, a confident operator, on June 25: "I can see me getting lost for 3 days on this. I want one property. That would be bad." The answer the research points at is Kyle's own habit. David read it from his notes on June 25, and we checked it against the recording: "He has three levels of underwriting. Stage 1, he only checks 8 or 10 things. And they start making offers at the end of stage one. Once it's under contract, his team's job is to convince Kyle why they should cancel the contract." Kyle accepts that one deal in ten loses money because the other nine more than cover it.
This one is not fading. It keeps coming up, and it is spreading. Six named operators hit it in the February to April 2026 calls. The June 25 call spent forty-five minutes on it, now split into foster-care, veteran, sober-living and disability markets. Kathryn Carlson on March 19, checked against the recording by the second reviewer: "the people that I know that have done some of these things, it takes them 9 months. Some people have done it for over a year. To get in the door and get the relationship with these non-profits." The problem is about relationships, so no software has touched it. The research calls it the strongest market nobody owns in the whole set, and cannot say how big it is.
Three smaller patterns worth knowing. Operators running several kinds of property have to rebuild their assumptions for each one (4 batches; Logan's list of new senior-living line items is the example). Insurance carriers cap how many co-living policies they write per ZIP code, and operators find out at renewal (5 of 12 calls, October to December 2025). And a capital partner with no window into the property (Jose Avila: "I know you guys are experienced, and I'm just bringing capital, but... We need to move things around").
4. The map
The research matched 49 problems against where Addi stood in early July. The July features brief (the five ideas David asked for on June 29) is folded in as the right-hand column, so the two pieces of work read as one.
| Problem | What Addi does today | Honest state | Where the July features brief lands |
|---|---|---|---|
| Is it legal, is it enforced | Zoning report with the unrelated-adults cap; a separate check of the enforcement history | Answered, in covered towns only. Coverage was North Carolina, Texas and part of Wisconsin at the end of April; 7,500 of 25,000 towns by mid-June. | "Full zoning read" (Addi reads the actual ordinance for new-build, planned-development and rezone questions) goes deeper on this |
| Hours per property | The whole report in about ten minutes | Answered, and proven on camera | |
| What is it worth | Comparable-sales report, with the condition read from photos | Partly. The Minnesota failure was still live in mid-June. A fix shipped July 1 that the research could not check | "Underwriting auto-fill" moves the numbers into the calculator without typing them again |
| Will a bank lend on it | Several financing scenarios in the calculator | Partly. The scenarios exist. There is no per-property "will a lender take this" check, no warning about lease count, no bedroom-count slider | Not in the brief |
| Five tools, typed in twice | Public submissions flow straight into the pipeline; a Drive folder per deal | Partly, and only for our own team. The pipeline the market would use is the one David raised on the August 24 check-in | "Lead generation" (listings that fit the buy box arrive on their own) and "workspace agent" (move a deal or assign a task by asking) both sit on this |
| Property management | Nothing | Not answered, and not Addi's job; a separate business (the "decomposable" model) | |
| Coverage gap | David says it out loud | Not answered in the product: no public coverage map, no "add my town" button, no promised turnaround | |
| Too much on screen, words people do not know | The full calculator | Not answered. The three-level quick answer, plain-English explanations next to each number, and a glossary are all on the roadmap. None have shipped | |
| Nonprofit partner | Nothing | Not answered. Addi could show which programs exist in a town; it cannot build the relationship | |
| Multifamily conversion | Nothing | Not answered. It would need a separate calculator with the ordinary-building case and the co-living case side by side | |
| A window for the capital partner | Nothing | Not answered. It would be a partner's view of the deal | Partly covered by the pipeline idea |
| Investor deck | Nothing | Terry called a printable investor summary "priceless" on June 25 | "Pitch-deck generation" |
| Our own way of judging deals | Generic scoring | Not answered | "Buy-box memory" |
Two things the second reviewer insisted go on the front page
On the June 25 webinar, with David presenting, he edited a rent figure, came back, and the edit was gone. The background process had overwritten it. David's ruling on the spot: "If a human overrides and manually enters something, then the system does not then go back and change it." That is not a feature request. It is a trust bug, and it has to be fixed before any beta user hits it without warning. The second thing: the link between the analysis findings and the underwriting calculator was, in Mufaro's own words on the June 11 debrief, half wired. The same weakness caused the rent overwrite.5. The test
The second analyst's job was to break the first analyst's conclusions. Here is what changed:
Nothing was found to be made up. All five of the quotes the study leans on hardest were found word for word at named lines of the transcripts, by two separate checkers.
6. The new work
Everything below comes from the evidence above. Where a line is our guess rather than something the research says, it says so.
The research's rule for how we present Addi applies to every post: lead with the operator's problem, in the operator's words, and let the technology stay invisible until someone asks. The lines below are written for David or Lacey to say out loud to camera. Each one rests on a quote from Section 3.
| Problem | The line | Why it works |
|---|---|---|
| Is it legal | "Twenty bedrooms, nine and a half baths, and the city said four unrelated people. Here's the question I ask before I ever drive to see a house." | The Greensboro story is the origin story. It comes up in three sessions, and David regretted forgetting it on the June 11 webinar. |
| Does the town enforce it | "The planner told me nobody enforces that rule. The records showed 185 times they did. Never take the verbal answer." | Checked against the recording. The single most quotable number in the set. |
| Hours per property | "I used to spend four hours to two days per house before I could make an offer. Here's what those hours were actually spent on." | David's own line, repeated by four other operators. |
| No comparables | "Your appraiser has never seen a nine-bedroom house. Three things to do before the appraisal so the file doesn't die." | Fernando's "9 bedroom 9 bath doesn't exist"; David's seven-lender story; the strip-the-keypads advice is something the community already passes around by word of mouth. |
| Financing does not fit | "Why the market rent on a co-living house is the number that kills your loan, and what the closers do about bedroom count." | Fernando's "the market rent sucks" and "buyer needs this property to be a 6 bedroom." |
| Five tools | "I run lead intake in one tool, research in another, a project board in a third and asset management in a fourth. Here's what I'd consolidate first." | David's own set of tools from the February 19 call; Karina's "doesn't do that." |
| Too much on screen | "Kyle only checks eight to ten things before he makes an offer. Here they are." | Checked against the recording; Kyle's three levels. Our guess, not something anyone said: this will also do best because it is a list. |
| Property management | "The deal I cancelled because my property manager was five hours away, and how I'd split the job today." | David's Wilmington cancellation; the "divide it into pieces" line. |
| Coverage | "We've read 7,500 of 25,000 towns' rules. Here's how to find out if yours is one, and how to get it added." | David's unprompted honesty earned trust in the room. The research recommends making that a habit in public. |
| Nonprofit partners | "Everyone leaves the conference expecting a $5,000 master lease. It takes nine months to get in the door. Here's month one." | Kathryn Carlson, checked against the recording; the forty-five minutes on it on June 25. |
| Multifamily | "A normal apartment building yields three to six percent. Convert a third of the units to co-living and Clara says seven to fifteen. The math." | Clara Arroyave's cap-rate line; Linda Holtz's 204-unit question. |
| Solo operator | "'How many people work on your team?' 'Eight.' 'Why do I join your team.' The analyst you can't afford to hire." | Ricky Steele, July 17 last year; the research's own recommended line for how to pitch Addi. |
Two notes for the content plan. First, the audience for the technical story is real but small. Logan Reinard and a handful like him (the research guesses five to ten percent of the room) want to know how it works under the hood. The recommendation is a separate technical page for them, never the main feed. Second, Kyle's suggestion to livestream underwriting a real property with Addi is the one format the research explicitly backs. It closes the gap where seven warm leads in seven weeks described Addi's problem and never heard Addi's name.
Each idea is described as what a person sees and does. They are ordered by how much evidence is behind them, not by how big they are to build.
Idea 1. The deal pipeline, for the market, with the analysis in it
What the person sees: every address they run lands on a board. They drag it through stages they name themselves (looking, offered, under contract, closing, owned). Each entry carries the zoning answer, the enforcement history, the rent numbers, the financing scenarios and the notes, so that entry is the whole file. They can invite a bird dog to drop addresses in, and a partner or lender to look at a deal without editing it. They open it every morning because that is where their deals are.
David's words on the check-in: "if Addie had the pipeline management, that would make it more sticky. If we gave the pipeline management piece to the market instead of just leave it internal, then more people would run more properties through it, and then be consistently referencing back to it every single day. Therefore it's more integrated into their daily process."
What the evidence says for it. The "five tools, typed in twice" problem is in 11 of 15 batches of notes, and the tool David cannot consolidate today is exactly this: a board through closing plus a second board after. Between May and September 2025 the coaching calls shifted from "how do I do this" to "how do I do this at scale." Rayce Barnes named his acquisition pipeline as his bottleneck. Michael Riordan built his own intake (a form, a chat channel, an automation, three underwriters) because nothing existed. Kyle's admission that he pays twice for things he cannot find is a filing problem, and filing is what a pipeline does. Addi's own two-screens problem, which Mufaro named as the biggest gap, is the same problem from the product side. Putting the deal board and the analysis on one screen fixes both. The capital partner (Jose Avila) and the bird dog (Nathan and his father's fifty doors) both want to look at the same deal, and a pipeline gives them that for free. And the public submission form already feeds our internal pipeline, so the intake half is live.
What the evidence says against it, or does not say. No operator in the recordings asked Addi for a deal board by name. The 62 wishlist items include contract generation, an investor summary, a coverage map, a master-lease model and a lender list. Not a tracker. The demand is our guess, built from the too-many-tools problem and the shift toward pipeline questions, and we mark it as a guess. "They would touch it every day" is something to test. The research has no data on how often the ten beta users open Addi, and we could measure that before building (our suggestion, not something anyone said). And Kyle is the warning. A deal board is the kind of thing he builds himself in thirty minutes, so a plain board earns nothing. The board only matters if the analysis lives on it, and no tool he can build in thirty minutes has that.
The answer: of all the ideas, this one has the most evidence behind it. But only if it is built as the analysis and the deal on one screen. A plain deal tracker on its own would not be worth building. And daily use should be measured on the beta group first.
Idea 2. The quick answer (Kyle's three levels)
What the person sees: the first screen on any property is eight to ten checks, each with a plain yes, no or watch out, and one line at the bottom: make an offer, or walk. Nothing else until they say "go deeper." Once under contract, the second screen is the kill list: what would make you cancel. The full calculator is the third screen. Every number has a "why this number" link in plain English.
Evidence: Kyle's method, as David read it out on June 25, checked against the recording. Terry Smith's "lost for 3 days." Marianne's fear of the calculator. Tom Paschke's "doesn't make sense to me that a hard money loan would give you a higher cash on cash." The 51-field calculator was called overwhelming on both the June 11 debrief and the June 25 demo. Addi's speed is already proven; clarity is where the demos risk losing people. One condition: the rent-overwrite bug gets fixed first, or the quick answer inherits the distrust.
Idea 3. "What else could this house be"
What the person sees: at the top of the zoning report, next to the co-living answer, a short list of the other uses the same house is allowed in that town (sober living, senior residential care, student housing, boarding house, long-term rental), with a yes or no beside each. They are buying a house with two exits instead of one.
Evidence: by April 2026 the community had taught itself to underwrite backup uses. Rayce Barnes' switch to sober living. Kyle on October 30 last year: "Can you pivot to also underwrite it? As a recovery home". Bridgette Ruiz's Oxford House waiting list across 44 states. Kathryn Carlson's foster-care route. Marianne's adult day-care idea on top of hers. The research's point is that people are already quietly working out their way out of a house before they buy it, and nobody is helping them do it.
Idea 4. "Will a bank lend on it"
What the person sees: a "will a bank lend on it" panel on every property. Which kinds of loan this house clears today (ordinary rent, short-term-rental income, established co-living income, bank statements). A warning if the plan has more than four individual leases, with what that does to the refinance spelled out. A bedroom-count slider that shows how the income and the loan change from five bedrooms to eight. A note when the deal will need a two-contract structure.
Evidence: 12 of 15 batches of notes. Fernando's four quotes in Section 3. Nathan Yerkes' fifteen lenders. Julian Rizov's "there is no 120% financing on the DSCR loans". Terry Smith finding out on June 25 that commercial lenders will not underwrite on collections. The research says building Fernando's rules into Addi is the best chance in the whole set to capture what an expert knows.
Idea 5. The coverage map and the "add my town" button
What the person sees: before they type an address, a map of what is covered. If their town is grey, one button asks for it and shows a written turnaround (the June 11 debrief discussed promising 48 hours and aiming for 24). When a report cannot be built, the screen says which of three things went wrong: the address, missing coverage, or a failure on our side.
Evidence: every demo hit a gap. Tom Paschke's second question of the call was "What's the timeline for other areas being integrated?" Mufaro's own debrief note that the user "is left to guess" which failure happened. And David's spoken honesty earned trust, which the research says has to move into the product before Addi meets a cold audience with no David in the room.
Not a feature idea, but named because the evidence keeps pointing at it
Doing the underwriting for people, with the fee priced into closing costs (Marianne's conditional yes). Two people. The research's ruling stands: five to ten interviews with the mission-driven group before any engineering.| Gap | Why it matters | What would close it |
|---|---|---|
| Nobody outside our community | Every voice is a graduate, a member or a regular. We do not know which problems hold outside our world. | Ten to fifteen conversations with operators from another community (the Denver co-living conference list, room-rental marketplace forums), asking the same questions. |
| Willingness to pay is almost blank | Nobody on the coaching calls named a budget or a price. The only firm number is a competitor's $88-a-month advisory membership. Marianne's yes was at closing, not monthly. | A pricing conversation with the ten beta users and with Jose Avila, the one person the readers flagged as most likely to pay. |
| Seven newer recordings not yet read | The research stops at June 25. Since then: Addi webinars on July 23, July 30 and August 6, coaching calls on July 16, July 23 and July 30 (all in the transcript archive), and the August 20 webinar recording (edited, not yet in the archive). Two months of the launch era are missing from this profile. | Read those seven the same way and check whether the ranking in Section 3 moves. This is also how we find out whether the trust bug and the coverage answer landed. |
| Daily use is an assumption | David's pipeline argument rests on people "referencing back to it every single day." The research has no usage data. | Pull how often each of the ten beta users opened Addi in July and August before deciding the pipeline scope. (Our suggestion, not something anyone said.) |
| The done-for-you service rests on two people | Marianne and Susan. Everyone else is our guess. | Five to ten interviews with Marianne, Tami, Rikki, Kathryn, Jenny Trejo, Angela Lavezzo, Rosaisela Rowan, Lizeth Picuasi, Nick Fialdini. |
| The nonprofit market cannot be sized | The strongest market nobody owns in the set, and no way to tell if it is small or large. | Ten to fifteen interviews with the mission-driven operators. Ask what they would pay for an introduction, not for software. |
| Voices we only heard second-hand | Terry Smith, Cliff Johnson and the conference specialists are quoted only as David repeated them. | Talk to them directly before any decision that hinges on their words. |
| The people never pitched | Benjamin Zonis (120-plus properties) and the seven warm leads from May to June who were never told. | One message each. The cheapest research in this document. |
| Multifamily demand is a count of two | Linda Holtz and Adam Catledge. Whether there are fifty or five hundred converters in our audience, we do not know. | A one-question poll on the Thursday call. |
| Groups of one | Logan (the AI-curious solo operator) and Andy Ungert (text-first) are each one person. | Do not size anything from them. Keep the design rules they imply. |
7. Sources
Search for newer analysis: files modified since 2026-08-22 under the features workspace, the platform knowledge folder and Downloads were listed and checked for mentions of the webinars, Kyle, pain points or profiling. None contained new analysis of this research; the July work is the base, and the only additions are the check-in quotes above. The five quotes the study leans on hardest were located at named lines of the 2026-05-28 live demo (line 99), the 2026-06-25 webinar (line 449), the 2026-06-19 Logan session (line 349) and the 2026-06-18 Naples session (line 1621).