WEBVTT 00:00:00.000 --> 00:00:02.560 When you receive a detailed message from a client, 00:00:03.080 --> 00:00:05.339 it feels great. Oh, absolutely the best feeling. 00:00:05.559 --> 00:00:07.459 Right. They list their budget, their timeline, 00:00:07.500 --> 00:00:09.859 and exactly what they want. Two sec silence. 00:00:10.539 --> 00:00:13.539 You feel this immediate, comforting sense of 00:00:13.539 --> 00:00:16.420 absolute clarity. You start to believe you actually 00:00:16.420 --> 00:00:20.559 see the entire picture. But, well, that feeling 00:00:20.559 --> 00:00:23.379 of certainty is often a dangerous trap. Yeah, 00:00:23.460 --> 00:00:26.000 it really is. That detailed message is secretly 00:00:26.000 --> 00:00:29.059 hiding the most important missing pieces. It 00:00:29.059 --> 00:00:32.359 is a total illusion. We humans absolutely love 00:00:32.359 --> 00:00:35.039 a complete narrative. Our brains just naturally 00:00:35.039 --> 00:00:36.979 fill in the blank spaces. All right, we do it 00:00:36.979 --> 00:00:39.280 without even realizing we are doing it. Exactly. 00:00:39.439 --> 00:00:41.659 Well, welcome to the deep dive. Today, we are 00:00:41.659 --> 00:00:44.960 unpacking a very specific lesson on real estate 00:00:44.960 --> 00:00:47.840 AI workflows. It is a highly practical one, too. 00:00:47.939 --> 00:00:50.280 It really is. We are taking a raw, seemingly 00:00:50.280 --> 00:00:52.920 perfect buyer inquiry. We're going to use Claude 00:00:52.920 --> 00:00:55.820 to build a practical bulletproof plan. It's all 00:00:55.820 --> 00:00:58.259 about preparing for that crucial first discovery 00:00:58.259 --> 00:01:00.439 call. Yeah, this is such a fascinating process. 00:01:00.619 --> 00:01:03.560 We are moving from raw data to a highly structured 00:01:03.560 --> 00:01:06.700 conversation. It's an absolutely vital skill 00:01:06.700 --> 00:01:09.920 for any modern professional. So let's look at 00:01:09.920 --> 00:01:12.480 this illusion of a complete inquiry. We have 00:01:12.480 --> 00:01:14.719 a great case study from our source material today. 00:01:15.400 --> 00:01:18.049 Her name is Sarah Johnson. Okay. She wants a 00:01:18.049 --> 00:01:20.590 three -bedroom home in Austin. Her budget is 00:01:20.590 --> 00:01:25.590 between $450 ,000 and $500 ,000. That is a very 00:01:25.590 --> 00:01:27.650 solid starting point for an agent. It gives you 00:01:27.650 --> 00:01:29.609 some real hard numbers to work with right away. 00:01:29.709 --> 00:01:32.790 Yeah. She also has a 90 -day timeline. She says 00:01:32.790 --> 00:01:35.030 her family simply needs more space right now. 00:01:35.430 --> 00:01:38.569 It's a classic, straightforward real estate scenario. 00:01:38.730 --> 00:01:41.189 On paper, that looks like an absolute dream lead. 00:01:41.390 --> 00:01:44.129 Most agents see that and immediately start pulling 00:01:44.129 --> 00:01:46.670 active listings. Right. They want to show her 00:01:46.670 --> 00:01:49.150 houses this very second. It's exactly like looking 00:01:49.150 --> 00:01:51.549 at a fresh puzzle box. You see the cover art 00:01:51.549 --> 00:01:53.170 perfectly clearly. It looks like a completely 00:01:53.170 --> 00:01:55.890 finished picture. But if you dive in, you're 00:01:55.890 --> 00:01:59.010 missing half the pieces inside. You try forcing 00:01:59.010 --> 00:02:01.569 edge pieces together that just don't fit. You 00:02:01.569 --> 00:02:03.890 end up damaging the pieces working off bad assumptions. 00:02:04.170 --> 00:02:06.510 Right. And that ruins the entire client relationship. 00:02:06.709 --> 00:02:08.449 The critical gaps here are actually quite huge 00:02:08.449 --> 00:02:11.150 when you look closely. Yeah. We have no idea 00:02:11.150 --> 00:02:14.050 which Austin area she actually prefers. We don't 00:02:14.050 --> 00:02:16.490 know her other home priorities beyond just bedrooms. 00:02:16.789 --> 00:02:19.129 And we definitely don't know her current financing 00:02:19.129 --> 00:02:21.789 status. Has she actually talked to a local bank? 00:02:22.229 --> 00:02:24.490 Does she need to sell her current house first? 00:02:24.719 --> 00:02:27.500 We just don't know yet. Right. Without that context, 00:02:28.039 --> 00:02:30.379 property recommendations come way too early. 00:02:30.580 --> 00:02:33.060 She is what we call a warm buyer lead in the 00:02:33.060 --> 00:02:35.300 industry. Right. She needs a 15 -minute discovery 00:02:35.300 --> 00:02:38.159 call right now. She does not need specific neighborhood 00:02:38.159 --> 00:02:41.680 guidance quite yet. Why do so many experienced 00:02:41.680 --> 00:02:44.530 agents rush? to recommend properties so quickly. 00:02:44.750 --> 00:02:46.729 They're absolutely terrified of losing a hot 00:02:46.729 --> 00:02:50.150 lead, so they overcompensate with immediate action. 00:02:50.409 --> 00:02:52.909 Fear of losing the client triggers premature, 00:02:53.229 --> 00:02:55.430 overwhelming property pitches. It backfires almost 00:02:55.430 --> 00:02:57.689 every single time you try it. So to stop ourselves 00:02:57.689 --> 00:03:00.229 from immediately pitching houses, we need a hard 00:03:00.229 --> 00:03:03.330 boundary. Yeah. And that brings us to how we 00:03:03.330 --> 00:03:06.710 actually set up AI as that filter. We are continuing 00:03:06.710 --> 00:03:09.930 this process inside Claude projects. It's an 00:03:09.930 --> 00:03:12.210 incredibly powerful feature if you know how to 00:03:12.210 --> 00:03:14.460 use it. For those unfamiliar, Cloud Projects 00:03:14.460 --> 00:03:17.659 is essentially a saved workspace. So AI remembers 00:03:17.659 --> 00:03:20.780 your previous files and instructions. We're opening 00:03:20.780 --> 00:03:24.039 a specific project called the Austin Family Buyer 00:03:24.039 --> 00:03:27.939 Workflow. Sarah's confirmed context is already 00:03:27.939 --> 00:03:30.020 saved securely in there. This is where we shift 00:03:30.020 --> 00:03:33.180 from reacting to actually planning. I'll be honest, 00:03:33.460 --> 00:03:35.379 I still wrestle with wanting to jump straight 00:03:35.379 --> 00:03:38.900 to listing homes myself. It is so hard to resist. 00:03:39.120 --> 00:03:41.360 playing the hero right away. We all do that. 00:03:41.460 --> 00:03:43.819 We inherently want to be helpful to people, but 00:03:43.819 --> 00:03:46.699 AI is being used here to establish strict conversation 00:03:46.699 --> 00:03:49.759 boundaries. We aren't using Claude to write an 00:03:49.759 --> 00:03:52.740 email to Sarah. We are asking it to analyze her 00:03:52.740 --> 00:03:55.560 actual readiness. We tell the AI to generate 00:03:55.560 --> 00:03:57.699 a lead readiness analysis card. I'm assuming 00:03:57.699 --> 00:04:00.159 it acts like a sort of checklist. or is it doing 00:04:00.159 --> 00:04:02.879 something deeper? It acts as a very strict cognitive 00:04:02.879 --> 00:04:05.439 filter for you. It separates the confirmed facts 00:04:05.439 --> 00:04:07.860 from the glaring open questions. Right. More 00:04:07.860 --> 00:04:10.680 importantly, it explicitly identifies what not 00:04:10.680 --> 00:04:12.919 to present yet. So it gives you a solid boundary. 00:04:13.020 --> 00:04:15.039 How does it actually do that under the hood? 00:04:15.539 --> 00:04:18.279 You use negative constraints in your AI prompt. 00:04:18.720 --> 00:04:21.879 You explicitly tell Claude, do not output any 00:04:21.879 --> 00:04:24.629 property listings. You tell it. do not provide 00:04:24.629 --> 00:04:27.389 any pricing guidance. So by forbidding the AI 00:04:27.389 --> 00:04:29.970 from doing the easy stuff, you force it to analyze 00:04:29.970 --> 00:04:33.129 the gaps. It explicitly says, no listings in 00:04:33.129 --> 00:04:35.589 the output. Right. No listings, no pricing guidance, 00:04:35.870 --> 00:04:38.730 no financing advice whatsoever. It forces you 00:04:38.730 --> 00:04:41.149 to completely pause your normal sales process. 00:04:41.430 --> 00:04:43.990 You only move forward with what is actually confirmed 00:04:43.990 --> 00:04:47.689 fact. Two sec silence. What is the actual tangible 00:04:47.689 --> 00:04:50.490 risk of AI hallucinating market trends here? 00:04:50.649 --> 00:04:53.269 If Claude invents a neighborhood price trend, 00:04:53.750 --> 00:04:56.170 the client anchors their financial expectations 00:04:56.170 --> 00:04:59.230 to a complete fantasy. Hallucinated data does 00:04:59.230 --> 00:05:02.329 not just look bad. It actively bankrupts client 00:05:02.329 --> 00:05:04.649 trust. It really does. That's why that strict 00:05:04.649 --> 00:05:06.350 boundary is absolutely essential. Now we have 00:05:06.350 --> 00:05:08.850 to turn that missing information into actionable 00:05:08.850 --> 00:05:11.250 decisions. We know exactly what is missing from 00:05:11.250 --> 00:05:14.490 Sarah's profile. How do we extract it systematically 00:05:14.490 --> 00:05:16.970 without sounding like a robot? We use Claude 00:05:16.970 --> 00:05:20.269 to craft a discovery call preparation note. We 00:05:20.269 --> 00:05:22.529 strategically target the three specific gaps 00:05:22.529 --> 00:05:25.310 we found earlier. So we map our questions directly 00:05:25.310 --> 00:05:27.930 to those glaring information gaps. Yeah. Question 00:05:27.930 --> 00:05:30.629 one, what preferred Austin areas do you want 00:05:30.629 --> 00:05:33.649 to focus on or avoid? That clarifies the future 00:05:33.649 --> 00:05:35.970 search focus right away. You avoid showing them 00:05:35.970 --> 00:05:38.230 the wrong side of town completely. Right. It 00:05:38.230 --> 00:05:40.649 saves everyone hours of wasted driving time. 00:05:40.990 --> 00:05:43.129 Question two, what are your priorities beyond 00:05:43.129 --> 00:05:46.730 three bedrooms? That clarifies the deeper underlying 00:05:46.730 --> 00:05:49.250 buyer needs. Maybe they desperately need a quiet 00:05:49.250 --> 00:05:51.430 home office. Yeah. Maybe they need a big yard 00:05:51.430 --> 00:05:53.410 for a new dog. You have to know the lifestyle, 00:05:53.529 --> 00:05:56.110 not just the bedroom count. And question three, 00:05:56.430 --> 00:05:59.550 is your 90 -day timeline flexible? That prepares 00:05:59.550 --> 00:06:02.069 your appropriate next steps immediately. If it's 00:06:02.069 --> 00:06:03.889 a hard deadline, you have to move incredibly 00:06:03.889 --> 00:06:06.689 fast. If it's flexible, you can wait patiently 00:06:06.689 --> 00:06:10.649 for better inventory. Three questions. That feels 00:06:10.649 --> 00:06:12.629 incredibly brief for a real estate discovery 00:06:12.629 --> 00:06:16.029 call. Usually, agents have a checklist of 20 00:06:16.029 --> 00:06:19.069 things to ask. Three focus questions are plenty 00:06:19.069 --> 00:06:21.430 at this stage. You don't want it to feel like 00:06:21.430 --> 00:06:24.149 an intense interrogation. Right. The AI gives 00:06:24.149 --> 00:06:26.649 you the structural roadmap. The human remains 00:06:26.649 --> 00:06:29.769 completely responsible for listening and recording. 00:06:30.050 --> 00:06:32.569 Why limit ourselves to exactly three questions 00:06:32.569 --> 00:06:35.240 for this initial call? Three questions. Keep 00:06:35.240 --> 00:06:37.860 the call conversational anymore, and it feels 00:06:37.860 --> 00:06:40.240 like an aggressive interrogation. Keep it to 00:06:40.240 --> 00:06:42.740 exactly three questions to spark a dialogue, 00:06:43.240 --> 00:06:45.680 not an interrogation. It really opens the door 00:06:45.680 --> 00:06:49.819 for genuine human connection. Sponsor. We will 00:06:49.819 --> 00:06:52.079 be right back to the deep dive after a brief 00:06:52.079 --> 00:06:54.449 word from our sponsors. All right, we are back. 00:06:54.610 --> 00:06:56.550 We're structuring the journey from inquiry to 00:06:56.550 --> 00:06:59.209 action now. We're building the AI powered lead 00:06:59.209 --> 00:07:01.870 to discovery decision map. This is where I absolutely 00:07:01.870 --> 00:07:04.449 love using Claude artifacts. Instead of getting 00:07:04.449 --> 00:07:07.389 lost in a long chat thread, you use this feature. 00:07:07.430 --> 00:07:10.470 Okay. A dedicated side window where AI generates 00:07:10.470 --> 00:07:13.610 clean editable final documents. It keeps the 00:07:13.610 --> 00:07:15.829 messy brainstorming chat separate from the final 00:07:15.829 --> 00:07:19.370 product. This decision map covers four very specific 00:07:19.370 --> 00:07:22.420 logical stages. First, buyer inquiry received. 00:07:22.959 --> 00:07:26.060 Second, lead readiness analyzed. Third, discovery 00:07:26.060 --> 00:07:29.300 call prepared. And fourth, next action recorded. 00:07:30.180 --> 00:07:32.420 Whoa, imagine scaling this level of preparation 00:07:32.420 --> 00:07:35.060 to dozens of clients simultaneously. It would 00:07:35.060 --> 00:07:37.019 completely change the architecture of a real 00:07:37.019 --> 00:07:39.579 estate business. You would never lose track of 00:07:39.579 --> 00:07:42.319 a client's status again. It's like stacking legger 00:07:42.319 --> 00:07:45.180 blocks of data perfectly every single time. Every 00:07:45.180 --> 00:07:48.019 single stage of this map requires a short human 00:07:48.019 --> 00:07:51.199 review point. The AI organizes the journey beautifully. 00:07:51.720 --> 00:07:54.519 You personally verify the facts. We have to keep 00:07:54.519 --> 00:07:57.279 that next action recorded stage flexible, right? 00:07:57.519 --> 00:07:59.420 Imagine you can't carve the next step in stone 00:07:59.420 --> 00:08:01.920 yet. Yes, absolutely. You write down the planned 00:08:01.920 --> 00:08:04.579 invitation right now. But you update the document 00:08:04.579 --> 00:08:07.100 only after Sarah confirms her priorities on the 00:08:07.100 --> 00:08:09.600 live call. Why is it so crucial to keep that 00:08:09.600 --> 00:08:12.500 next action stage completely flexible? Because 00:08:12.500 --> 00:08:15.019 a live conversation might reveal an unexpected 00:08:15.019 --> 00:08:17.860 priority that changes your entire strategy instantly. 00:08:18.079 --> 00:08:20.100 Live conversations change priorities, so your 00:08:20.100 --> 00:08:22.839 next steps must adapt instantly. You absolutely 00:08:22.839 --> 00:08:24.779 have to listen to the human in front of you. 00:08:25.120 --> 00:08:28.019 Once this decision map is finally built, where 00:08:28.019 --> 00:08:31.089 does this data actually live long term? We have 00:08:31.089 --> 00:08:33.909 to prepare the final CRM handoff. Our sources 00:08:33.909 --> 00:08:36.230 use a platform called System for this, but the 00:08:36.230 --> 00:08:38.809 underlying logic applies universally. You desperately 00:08:38.809 --> 00:08:41.750 need a CRM handoff sheet. Right. It ensures your 00:08:41.750 --> 00:08:45.070 database reflects actual reality. It must reflect 00:08:45.070 --> 00:08:48.210 reality, not just your hopeful assumptions. I've 00:08:48.210 --> 00:08:50.850 seen agents lose massive commissions over this 00:08:50.850 --> 00:08:53.570 exact issue. Oh, it happens constantly in this 00:08:53.570 --> 00:08:55.710 industry. They assume a client has a massive 00:08:55.710 --> 00:08:58.750 budget because they drive a nice car. The CRM 00:08:58.750 --> 00:09:00.830 looks complete, but it's based on pure fantasy. 00:09:01.289 --> 00:09:03.490 The handoff groups information into three distinct 00:09:03.490 --> 00:09:05.990 important buckets. Bucket 1 is confirmed client 00:09:05.990 --> 00:09:09.470 context. Her specific lead type, location, budget 00:09:09.470 --> 00:09:12.350 range, and timeline. These are the hard, undeniable 00:09:12.350 --> 00:09:14.929 facts from her email. Bucket 2 records the current 00:09:14.929 --> 00:09:17.970 workflow position. She is categorized as a warm 00:09:17.970 --> 00:09:20.690 buyer lead. A short discovery call is currently 00:09:20.690 --> 00:09:24.110 planned. Bucket 3 is the open items. This is 00:09:24.110 --> 00:09:25.889 the stuff we still critically need to figure 00:09:25.889 --> 00:09:28.850 out. Preferred areas, home priorities, and current 00:09:28.850 --> 00:09:32.289 financing status. That final group is so incredibly 00:09:32.289 --> 00:09:35.289 easy to overlook, but it prevents the dangerous 00:09:35.289 --> 00:09:38.570 false impression of completeness. A CRM must 00:09:38.570 --> 00:09:41.309 clearly show what you still need to learn. If 00:09:41.309 --> 00:09:43.389 it looks complete, you naturally stop asking 00:09:43.389 --> 00:09:46.059 questions. Right. You think you have the whole 00:09:46.059 --> 00:09:48.639 picture already. If your CRM lacks financing 00:09:48.639 --> 00:09:51.440 data but looks finished, future follow -ups become 00:09:51.440 --> 00:09:54.480 completely irrelevant. You end up sending useless, 00:09:54.500 --> 00:09:58.440 annoying emails. Two -sex silence. How exactly 00:09:58.440 --> 00:10:01.539 does bad CRM data destroy a client's trust over 00:10:01.539 --> 00:10:04.450 time? You end up asking the same questions twice 00:10:04.450 --> 00:10:07.009 or sending irrelevant listings, which proves 00:10:07.009 --> 00:10:09.250 you aren't listening. Asking the same questions 00:10:09.250 --> 00:10:11.610 twice proves you simply are not listening to 00:10:11.610 --> 00:10:13.309 them. It completely destroys the professional 00:10:13.309 --> 00:10:15.409 relationship before it even starts. We're at 00:10:15.409 --> 00:10:18.220 the final stage of the workflow now. The vital 00:10:18.220 --> 00:10:21.000 human review checkpoint. This is the critical 00:10:21.000 --> 00:10:23.399 boundary between AI preparation and confirmed 00:10:23.399 --> 00:10:26.159 client guidance. There is an absolute unbreakable 00:10:26.159 --> 00:10:29.259 rule here. Never ask Claude to judge if its own 00:10:29.259 --> 00:10:31.820 work is accurate. You must manually compare the 00:10:31.820 --> 00:10:34.279 generated artifact against Sarah's original brief. 00:10:34.639 --> 00:10:37.399 Yes. You have to establish a strict pass condition. 00:10:37.759 --> 00:10:40.460 Does the data perfectly match the original source 00:10:40.460 --> 00:10:42.519 material? But wait, I have to push back on this 00:10:42.519 --> 00:10:46.230 a little bit. If the AI is smart enough to organize 00:10:46.230 --> 00:10:49.470 this whole complex map, why can't it just cross 00:10:49.470 --> 00:10:53.289 -reference its own output? That seems like unnecessary 00:10:53.289 --> 00:10:55.789 busy work for us. Because AI can suffer from 00:10:55.789 --> 00:10:58.870 a very real kind of confirmation bias. Large 00:10:58.870 --> 00:11:01.190 language models are designed to predict words, 00:11:01.350 --> 00:11:04.490 not necessarily check facts. It inherently wants 00:11:04.490 --> 00:11:07.289 to please you. It might subtly alter a fact to 00:11:07.289 --> 00:11:09.409 make the output look mathematically cleaner. 00:11:09.970 --> 00:11:11.909 So the human past condition is completely non 00:11:11.909 --> 00:11:14.350 -negotiable here. It's the only reliable way 00:11:14.350 --> 00:11:16.809 to catch unsupported property or pricing claims. 00:11:16.950 --> 00:11:19.610 You have to be the ultimate final filter. You 00:11:19.610 --> 00:11:21.870 personally correct anything that overstates what 00:11:21.870 --> 00:11:24.590 you actually know. Excellent use of AI depends 00:11:24.590 --> 00:11:27.809 entirely on that clear line. Preparation is strictly 00:11:27.809 --> 00:11:30.269 for the AI. Confirmed guidance is strictly for 00:11:30.269 --> 00:11:33.289 the human. B, where does AI actually fit into 00:11:33.289 --> 00:11:35.509 the legal liability of a real estate transaction? 00:11:35.889 --> 00:11:39.110 It doesn't exist legally. The licensed human 00:11:39.110 --> 00:11:43.029 agent carries 100 % of the liability. The licensed 00:11:43.029 --> 00:11:46.830 human agent owns all legal liability, never the 00:11:46.830 --> 00:11:49.850 AI. That is exactly why you verify every single 00:11:49.850 --> 00:11:52.409 tiny detail. Let's recap the big idea from this 00:11:52.409 --> 00:11:56.149 entire deep dive. We started with a raw, seemingly 00:11:56.149 --> 00:11:59.429 complete buyer inquiry. It had several hidden 00:11:59.429 --> 00:12:01.789 unknowns lurking just beneath the surface. We 00:12:01.789 --> 00:12:04.230 used Claude to build a strict readiness review. 00:12:04.490 --> 00:12:07.190 We mapped out three highly targeted questions 00:12:07.190 --> 00:12:09.549 for a discovery call. And we prepped a highly 00:12:09.549 --> 00:12:12.940 accurate CRM handoff. We separated confirmed 00:12:12.940 --> 00:12:15.720 facts from assumptions at every single step of 00:12:15.720 --> 00:12:17.980 the workflow. It's a phenomenal way to structure 00:12:17.980 --> 00:12:20.360 your daily thought process. It radically protects 00:12:20.360 --> 00:12:22.460 your business and it serves the client much better. 00:12:22.679 --> 00:12:25.039 Two -sex silence. Think about your own professional 00:12:25.039 --> 00:12:27.659 database right now. If someone audited your records 00:12:27.659 --> 00:12:30.220 today, how much of it is confirmed fact and how 00:12:30.220 --> 00:12:32.299 much is just a hopeful assumption waiting to 00:12:32.299 --> 00:12:34.720 break a deal? That is a genuinely scary thought 00:12:34.720 --> 00:12:36.840 for a lot of professionals out there. I highly 00:12:36.840 --> 00:12:39.720 encourage you to apply this mindset in your next 00:12:39.720 --> 00:12:42.019 interaction. Separate your confirmed facts from 00:12:42.019 --> 00:12:44.700 your open questions. See how it instantly changes 00:12:44.700 --> 00:12:47.200 the clarity of your conversations. Thank you 00:12:47.200 --> 00:12:49.620 for joining us on this deep dive of Yuji Haro 00:12:49.620 --> 00:12:50.000 music.