Every Building Is a Balance Sheet: Spatial Memory for Real-World Assets — Jordi Spranger, Spranic

AI Engineer · 10 min · 102 sentences · from YouTube's caption track

Each timecode opens YouTube at the start of that sentence. Line anchors (#s42) are the cue ids in the WebVTT, and every line carries its start and end seconds. All transcripts has every talk, and the whole corpus as one file.

  1. 00:01Hello.
  2. 00:02I'm very excited to give this talk today at AI Engineer NYC 2026.
  3. 00:07Um this is a collaboration between Sprana and Neo4j.
  4. 00:10Sprana, I am the founder of.
  5. 00:12It was founded in June 2026.
  6. 00:14Neo4j is a graph database in based in Sweden, Malmö.
  7. 00:19And um a special thanks here at Zaeem, who is the collaborator on Neo4j's side.
  8. 00:27This is about every building is a balance sheet and the spatial memory for real-world assets.
  9. 00:34The largest asset class on Earth is real estate and it's about 400 trillion globally in size.
  10. 00:40And um this is more than global equities, debt, and gold combined.
  11. 00:45If we look on the right, you see a lot of squares.
  12. 00:48Each square is about 1 trillion uh in worth and you can see the residential part is the biggest.
  13. 00:54It's around 300 trillion.
  14. 00:56Then you can see the commercial is around 60 trillion and then agriculture around 50 trillion.
  15. 01:02And uh today we will focus more on the commercial side, but all of this um real estate is impacted by essentially a gap.
  16. 01:11And the gap is that our financial books uh where we have lending and insurances and um mortgages, all of this is book kept um and this is updated every quarter.
  17. 01:26But the actual underlying assets, which is the real estate that they are based on, um they're not updated every quarter.
  18. 01:34In fact, they're only updated maybe every few years, maybe just before the financial instrument for the for the real estate is being created.
  19. 01:45And also maybe again once it needs a modification or a new financial instrument on top of it, maybe in in year five is a very re- realistic time frame.
  20. 01:58But between year zero and year five, you really don't have an update on the building state or the real estate state.
  21. 02:07So that means that while the book closes so many times, there's really a black box or maybe something that is based on some pictures and some PDFs about the building.
  22. 02:17And this is a real gap.
  23. 02:19Um and you can see it.
  24. 02:22So according to Kroll, um they did a research with 1,500 properties.
  25. 02:29Essentially, 70% of the buildings are underinsured.
  26. 02:32And why exactly that is, um I don't exactly know.
  27. 02:37It didn't conclude.
  28. 02:38But my read is that the the record is quite old and it's maybe also not as detailed.
  29. 02:44So the insurance is not very highly specified and really fulfilling the needs of the real estate.
  30. 02:51And
  31. 02:51[snorts]
  32. 02:51this impacts around 4 trillion US commercial mortgages.
  33. 02:56And um yeah, every one of them rests on the building.
  34. 03:00Uh so the question, what would it take to make a building as queryable as a ledger?
  35. 03:05And a ledger here can be give you an answer in milliseconds.
  36. 03:09But a building today is giving you only an answer after a site visit.
  37. 03:14And even then, maybe in like some notes, PDFs, very dispersed and intransparent, not unified.
  38. 03:23Um so this is where the idea comes in, where every walk becomes a memory.
  39. 03:26And how this works is you just take a off-the-shelf sensor, you walk the space, and you map it.
  40. 03:32And you can see it up there actually being created.
  41. 03:35That was me uh going through a conference hall.
  42. 03:38And the whole walk or the whole scan here took literally just walking through it.
  43. 03:42And a minute later, you had the scan.
  44. 03:45Um so you walk through the property, whoever that is, property management, etc., can be equipped with it.
  45. 03:51They map it.
  46. 03:52And [clears throat] as they map it, you get all the assets in the building being recognized and stored in a database, which is building below.
  47. 04:01This is the graph database.
  48. 04:03And so, this is then has a semantic meaning, which means we can really understand the space.
  49. 04:09And as we walk it again, both walks are kept.
  50. 04:12So, walk on Monday and a Monday next year or the next quarter.
  51. 04:16And these are now aligned on a time axis.
  52. 04:19So, we have a space, we know what's in the space, and we know the evolution of the space according to every every time we walk it and we measure it.
  53. 04:29And this is all kept as a graph.
  54. 04:31Graph have a bunch of technical benefits, but one of them is they're extremely flexible in what information you want to gather and how you want to connect it.
  55. 04:40And so, this can be hyper personal personalized to each asset to each to each property.
  56. 04:47Um yeah.
  57. 04:48And so, now what is it that a record has to answer?
  58. 04:53So, these are not vision questions, they are ledger questions.
  59. 04:57You really want to know in a property or in the real estate, what has changed?
  60. 05:02So, two walks of one space compare object by object and and surrounding by surrounding.
  61. 05:09You also want to know when it has changed.
  62. 05:10So, every observation carries and captures where it came from.
  63. 05:14And you also want to know who needs to know information or deltas that have happened that we have now observed.
  64. 05:21So, when you find something with evidence, for example, structural damages or maybe for fire insurance that fire extinguishers are off off Yeah, they have expired, then you need to know that and you need to be informed and counteract on it.
  65. 05:41Or maybe the insurance can know and they can increase their premiums by the performance of the upkeep of the property.
  66. 05:49So, hyper-individualized insurance.
  67. 05:53Um and here's really how it looks.
  68. 05:55So, this is just me walking through a kitchen captured twice, once on the left and once on the right.
  69. 06:01And you can see you just walk the space, you directly get the map, and um and as you walk the same space, they're being recognized as the same space,
  70. 06:12and you can align them between um the same space in the time period A and the time period B, and you have a timeline.
  71. 06:20And once you have a map, you can now use it as an AI-ready asset, which means that you can chat with it, you can talk with it, you can ask questions against it.
  72. 06:31For example, here I ask, "Where can I heat myself?"
  73. 06:34And in the map, the answer is being given to the radiator.
  74. 06:37Or, "Is there a glass or mug to drink from in a kitchen?"
  75. 06:41It It points you to the glasses and mugs in the kitchen.
  76. 06:45Um so, this is very powerful because it makes the map from pixels into something we can work with.
  77. 06:53Um and now also the change detection.
  78. 06:55So, this were two passes we made of this kitchen or of this space, and we can really now connect these two and ask, "What is the delta between them?
  79. 07:03What has changed in this environment?"
  80. 07:05And in this case, um these three chairs in the middle have been displaced, and you can see when you chat with the with the asset, with your data, you can actually identify the delta and get the ground source reference to the actual capture, so you have proof of it changing.
  81. 07:24Um so, what does this all mean?
  82. 07:26It means that you have now a building that is an AI-ready data asset.
  83. 07:31So, there's one record of what it what is there, and you can connect it to any downstream task.
  84. 07:37So, on the left we have this conference scan that I made, and then you can now check over time any repair that needs to be happening or some maintenance
  85. 07:47or some compliance checks or insurance or maybe interior design changes or maybe even you have some robotics that you want to place there and it's like a digital twin for it to use.
  86. 07:59So, let's check the insurance.
  87. 08:02When you walk the building, when the policy starts, you have like a first initial state of the property and on based on this you can do a hyper-personalized insurance policy.
  88. 08:14And as you walk it again at renewal, you now see the delta and the upkeep and so on and the insurance policy insurance company can actually measure the performance
  89. 08:23of the property for relevant facts against its policies and give a very hyper-personalized insurance policy, which means a very efficient premium to insure such property and insure it in the right way.
  90. 08:40Uh yeah, and these are the numbers behind that.
  91. 08:42So, the business case here for the insurance for example would be that the insurance pays for what it can verify and so currently what insurers know is that one in three commercial buildings are on the books with the wrong construction class.
  92. 08:5835% of the buildings in the specific case for example for the hurricane premium can actually get a deduction of of such percentage if it's a certified building.
  93. 09:10Certified building means it has certain compliance standards and that you can check from such an AI digital asset of the of the asset.
  94. 09:20And you also see that 70% 73% of the insurers agree that they claim fewer claims for hurricane damage if they have been checked and certified against the building.
  95. 09:35And just this is just one such certification you can do automatically with AI plus a lot of other certification and inside of the building that qualifies maybe for such programs.
  96. 09:48But yeah, all three of them essentially depend on somebody knowing the building and and the state of the building.
  97. 09:55Um and that's it.
  98. 09:56That's essentially the talk.
  99. 09:58I think that the real estate class is very big and it really needs to be transparent and there's a huge inefficiency happening day that I think with the new insights of what AI is capable to do
  100. 10:11we can make much more transparent and create a really new hyper-customized market around around assets.
  101. 10:20Thank you very much.
  102. 10:22Um if you want to find out more, please visit my website spranik.com and thank you for your attention.