AI for Genealogy: FamilySearch Full-Text Search - Find Ancestors in 30 Seconds Using AI (2+ Billion Records Unlocked!)

What if you could search through 2+ BILLION genealogy records in 30 seconds instead of spending hours browsing page by page?

In this complete tutorial episode, you'll discover FamilySearch Full-Text Search—a free AI-powered tool that's revolutionizing how genealogists find their ancestors. This game-changing feature uses artificial intelligence to read handwriting in over 7,757 collections, making previously unsearchable records instantly discoverable.

🎯 WHAT YOU'LL LEARN:

Getting Started (No Tech Skills Required!):

• How to access Full-Text Search with your free FamilySearch account
• Step-by-step walkthrough of the search interface
• Understanding AI-generated transcripts and summaries
• When to trust AI results vs. when to verify original images

Search Techniques That Get Results:

• Using quotation marks, wildcards, and operators like a pro
• Finding ancestors as witnesses (never indexed before!)
• Searching by occupation, location, or any word on the page
• DGS numbers: How to search specific collections for targeted results

Advanced Strategies:

• The multi-collection sweep method
• Reverse searching by location instead of names
• How to find spelling variations automatically
• Using AI-generated English summaries for foreign language records

Real Success Story: Hear how Full-Text Search helped crack a 150-year-old Irish immigration mystery in just 30 minutes—finding a first marriage, three unknown children, and a county of origin that traditional searches never revealed.

💡 WHY THIS MATTERS:

FamilySearch has over 5 billion digitized images, but only a fraction were searchable—until now. Full-Text Search uses AI handwriting recognition to create searchable transcripts of deeds, probate records, court documents, church records, and more. Instead of browsing 800 images hoping to spot your ancestor's name, you can search the entire collection in seconds.

Current coverage includes: ✓ 7,700+ collections (growing weekly) ✓ 1.5 billion+ record images ✓ English, Spanish, and Portuguese records (more languages coming) ✓ Revolutionary War pensions, land deeds, wills, court records, and more ✓ 100% FREE with your FamilySearch account

🎓 YOUR HOMEWORK ASSIGNMENT:

Beginner: Set up Full-Text Search and practice one search Intermediate: Use wildcard operators to find name variations Advanced: Do a multi-collection sweep and document 3 new findings

⚠️ REMEMBER THE GOLDEN RULE: AI is your research assistant, not your researcher. This tool helps you FIND records faster, but you still verify, analyze, and build proof like the good genealogist you are.

🎧 NEW TO THE PODCAST? Start with Episode 1 to learn the foundations of using AI safely in genealogy research. We cover ChatGPT, Claude, Gemini, Perplexity, and other AI tools specifically for family history—all with practical examples and copy-paste prompts you can use immediately.

Connect with Ancestors and Algorithms:

📧 Email: ancestorsandai@gmail.com
🌐 Website: https://ancestorsandai.com/
📘 Facebook Group: Ancestors and Algorithms: AI for Genealogy - www.facebook.com/groups/ancestorsandalgorithms/

Golden Rule Reminder: AI is your research assistant, not your researcher.

Join our Facebook group to share your AI genealogy breakthroughs, ask questions, and connect with fellow family historians who are embracing the future of genealogy research!

New episodes every Tuesday. Subscribe so you never miss the latest AI tools and techniques for family history research.

Transcript

Ok, imagine this. You've been searching for your great-great-grandmother in county deeds for, let's say three hours, maybe four. You're clicking through image after image of handwritten documents, squinting at cursive that looks like it was written by a spider having a bad day. Your eyes are crossing, your coffee's gone cold, and you're starting to question all your life choices. Now, imagine finding that exact same record in 30 seconds. Not 30 minutes. 30 seconds. That's not genealogy fiction, my friends. That's FamilySearch Full Text Search. And today, I'm going to show you exactly how to use this absolutely game-changing tool to unlock almost 2 billion records that were essentially invisible until now. Let's dive in. Welcome to Ancestors and Algorithms, where family history meets artificial intelligence. I'm your host, Brian, and if you're new here, welcome to the lab. We're not just talking about using AI and genealogy. We're actually doing it together. Learning what works, what doesn't, and how to use these tools safely and effectively. Today's episode is going to be a step-by-step tutorial. I mean, truly step-by-step. We're going to walk through every single click, every search technique, and every feature of FamilySearch's full text search tool. And here's what makes this so exciting. This tool uses AI to read handwriting in over 6,000 collections. That's almost 2 billion record images that were basically unsearchable until now. Quick reminder before we start. This episode builds on what we've learned in previous episodes about using AI responsibly. If you're brand new to the podcast, you might want to start with episode 1, where we cover the foundations. But don't worry, you'll still get tons of value from today's tutorial, even if this is your first episode. So, let's get into it. Before we dive in, I want to mention something quickly. The names, locations, and some specific details in today's stories have been slightly altered to protect people's privacy. So, if a county name doesn't quite match up, or an ancestor's name seems off, or you're thinking, wait, that doesn't sound right for that region, you're probably catching one of those changes. The research techniques and the core genealogy challenges, those are absolutely real and based on actual brick walls that genealogists face. But I've tweaked identifying details because, well, not everyone wants their family mysteries broadcast to the world. The AI tools and methods I'm showing you will work just as well on your real ancestors as they do on my slightly fictionalized examples. So, let me tell you about a brick wall that had haunted me for months. My third great-grandfather, William Henderson, we'll call him Will for short, supposedly owned land in Tennessee in the 1840s. I knew this because family stories mentioned it and I'd found one deed where he sold property in 1847. But where did he get that land? Who did he buy it from? Was he maybe mentioned in other deeds as a neighbor or a witness or a boundary reference? Now, here's the problem. The county's land deed books for that period, they're on family search. All digitized. Hundreds and hundreds of beautiful, clear images. But, and this is a big but, they weren't indexed. Not the grantors and grantees, not the witnesses, not the property descriptions. Nothing. So, my options were, browse through literally 800 deed book images one by one, or... Well, that was basically my only option. I'd try the traditional index search on FamilySearch. Nothing. I'd check the grantor-grantee indexes at the county level. Just that one 1847 deed. And you know what? I started browsing. I really did. I made it through about 50 images before I realized this was going to take me days. Maybe weeks. And, I might still miss him if his name was misspelled or if the handwriting was particularly rough. Or, if I just got tired and my eyes glazed over at the wrong moment. That's the brutal reality of unindexed records. They're there. We know they exist. We can see them. But, actually finding anything in them? It's like looking for a needle in a haystack. While blindfolded. And the haystack is on fire. But, here's where the story gets good. Fast forward to about nine months ago. I'm at RootsTech 2025 and there's buzz about the FamilySearch's full text search getting major upgrades. I've tried it when it first launched in 2024 and honestly, it was okay. Promising but limited. Only a few collections. Results were hit or miss. But now, now they're saying it covers thousands of collections, almost 2 billion images, and the AI has gotten way better at reading handwriting. So, I pull up my laptop right there in the convention center. Sign in to FamilySearch, navigate to the full text search experiment, and I type in William Henderson and Tennessee and 1840s. 30 seconds later, and I mean this literally, I timed it because I couldn't believe it. I'm looking at three deed records I've never seen before. One from 1842 where Will witnessed a property sale. One from 1844 where he's mentioned as an adjacent landowner. And one from 1846 where he bought the property he later sold in 1847. Three records. Three pieces of evidence that would have taken me days or possibly weeks to find manually. Found in 30 seconds. Now, before we get into exactly how to use this tool, I need to be really clear about something. And this is important, so stick with me here. AI didn't do my research for me. That's not what happened. The AI read the handwriting and made those records searchable. But I still had to evaluate each record, look at the original images, verify the transcriptions, and figure out how these documents fit into Will's story. This is why I always say, AI is your research assistant, not your researcher. Full text search is a finding aid. An incredibly powerful finding aid, but it's still just showing you where to look. You still have to do the actual genealogy work. The analysis, the verification, the proof building. But man, oh man, does it make that work so much easier and so much faster. Okay, let's talk about what full text search actually is and how it works. Alright, so what exactly is FamilySearch full text search and how is it different from the regular FamilySearch search we're all used to? Let me break this down as simply as I can. You know how when you search FamilySearch normally, you're searching through indexed records. Those indexes were created by actual human volunteers who looked at each record and typed in the names, dates, and places into specific fields. That's traditional indexing and it's incredibly valuable, but it's also incredibly time-consuming. FamilySearch has over 20 billion digitized images in its collection. That's billion with a B. But only a small fraction of those images have been indexed. The rest, they're just sitting there as images. You can browse them if you know exactly where to look, but you can't search them until now. Full text search uses artificial intelligence, specifically handwriting recognition technology, to read every word on every page of certain record collections. Not just the names. Not just the dates. Every single word. The AI creates a full text transcript of each image and then you can search through those transcripts just like you'd search through a book or a PDF. Want to find records that mention blacksmith? You can search for that. Want to find documents where someone's described as a widow? Search for it. Want to find land descriptions that mention Maple Creek? You got it. Now, you might be thinking, wait, isn't that just OCR? Optical character recognition? Well, sort of, but not exactly. FamilySearch is using really advanced AI models that can handle old handwriting styles, different languages, different languages, faded ink, and all the quirks of historical documents. It's way more sophisticated than the OCR you'd use on your scanner at home. So, here's what makes this such a game changer. Let's take marriage records as an example. Traditionally, when volunteers index marriage records, they typically capture the bride's name, the groom's name, the date, and maybe the location. But, what about the witnesses, the officiant, the parents' names that they're mentioned, the location where the ceremony took place? All that information is in the record, but it's not in the index. So, if you're trying to find your ancestor as a witness to their brother's wedding, The traditional search won't find them. But full-text search, it sees everything on that page. Every name, every relationship term, every and it makes it all searchable. Okay, let's talk about what's actually available right now. As of November 2025, so this is current as of recording this episode, full-text search includes over 6,000 collections. That's up from about 4,000 in August, so they're adding collections really quickly. And we're not talking about tiny collections. We're talking about almost 2 billion individual record images. That number keeps growing, so by the time you're listening to this, it might be even higher. The collections include deed records, probate records, court records, church records, vital records. Basically, a lot of the unindexed collections that genealogists have been wanting to search for years. Now, here's something important to know. The AI models work best with English, Spanish, and Portuguese right now. Those are the languages FamilySearch developed first. They're working on models for Chinese, French, German, Dutch, Italian, and other languages, but those aren't fully rolled out yet. ...but those aren't fully rolled out yet. So if you're researching in those other languages, you might have mixed results at the moment. But even with that limitation, the coverage is incredible. Major collections from the United States, Latin America, the UK, Ireland, Australia, Canada... …but there's a ton here to work with. Now, let me tell you about some features they've added that are honestly mind blowing. When you pull up a record in full text search, you don't just see the image. You get: 1. A full AI generated transcript of the document. Every word the AI could read is typed out for you. No more squinting at impossible handwriting. 2. An AI-generated summary of the document that pulls out the key information. Names, relationships, dates, what the document is about. This summary is in English even if the original document is in Spanish or Portuguese. 3. The system highlights your search terms right on the image in yellow so you can instantly see where your ancestor's name appears on the pages. 4. You can download the image, the transcript and even a pre-formatted citation all from the same screen. And 5. There's a feature called Fact View that extracts and displays all the people, places and dates the AI recognizes in the document. Now, before you get too excited, and I know this is exciting, we need to talk about accuracy. 5. Because here's the thing: The AI is really good, but it's not perfect. It makes mistakes. 6. Sometimes it misreads handwriting. 7. Sometimes it gets confused by faded ink or weird spellings. 7. Sometimes it hallucinates words that aren't there or misses words that are. That's why, and I can't stress this enough, you always, always, always need to look at the original image yourself. 8. Use the transcript as a guide. 9. Use the summary to quickly figure out if a record is relevant. But then verify everything against the actual historical document. Think of it this way: If you had a research assistant who read through 800 deed books for you and said, "Hey, I found three that mention your William Henderson," would you just trust them blindly? "No!" You'd say, "Great, show me those three records." And then you'd examine them yourself. That's exactly how you use full-text search. 1. Alright, so that's what full-text search is: An AI-powered finding aid that makes previously unsearchable records searchable by creating full-text transcripts. It's experimental, it's still improving, improving, but it's already incredibly powerful. Now, let's get into the practical stuff. Let me show you exactly how to access this tool and get started. Okay, I'm going to walk you through this like you're looking over my shoulder in my computer screen. We're going to do this together, one step at a time. And don't worry if you can't follow along right this second, you can always pause this episode and come back to it when you're at your computer. First things first, you need a FamilySearch account. The good news is FamilySearch accounts are completely free. If you already have one, you're golden. If you don't, go to FamilySearch. org and look for the sign-in button in the upper right corner. Click that, then click "Create a free account" and follow the prompts. Takes about two minutes. Got your account? Great! Now sign in. Here's the easiest way to access full-text search. Once you're signed in at the top of the screen, you'll see "Family Tree", "Search", "Memories", "Get Involved", "Activities". Click on "Search" and the second one down should say "Full Text". Click on that. Now, in the middle of your screen, you should see the search interface. There are several search fields. First is the "Keywords" field. This is super powerful. You can search for any word or phrase. Below that is the "Name" field. You can put "Ancestor" names here. Then there's "Place", pretty self-explanatory. "Year" with the "Range" option, so you can search like 1840 to 1850. And finally, "Image Group Number", which we'll talk about in a bit. That's the "DGS" number. Okay, let's do a practice search together. I'm going to use a fairly common name, so you can see how the results work, but not so common that we get overwhelmed. Let's search for someone named "Sarah Collins" in Virginia in the 1850s. Now, I don't actually have a "Sarah Collins" in my tree. I'm just using this as an example, but follow along with the process because you'll use the exact same steps with your own ancestors. In the "Keywords" field, I'm going to type "Sarah Collins" in quotes. Notice I put that in quotation You can see that in quotation marks that's important. Without the quotes, full text search would look for records with the word "Sarah" or the word "Collins" anywhere on the page. Not necessarily together. With the quotes, it looks for that exact phrase. In the "Place" field, I'm typing "Virginia". In the "Year" field, I'm going to put "1850" in the first box and "1860" in the second box. That gives me a decade to work with. I'm leaving the "Name" and "Image" group number field blank for now. Now, I'm clicking the big blue "Search" button. And...okay, results are loading. In the "Lear" field, I'm looking at the "Markin" field. Here's what you're looking at. Each result shows: A snippet of text from the document with your search terms highlighted. The collection name, like "Virginia deed books" or "Virginia County court records". The date range of the collection. The specific place, county level usually. And a little preview of where in the document your search term appears. At the top of the results list, you'll see a "Filter" option. You can filter by "Collection", "Year", "Place", or "Record" type. These are incredibly useful when you're getting hundreds or thousands of results like we just did. There's also a toggle in the upper right called "Fact View". If you turn that on, instead of seeing text snippets, you'll see structured data, the people, places, and dates the AI extracted from each record. Now let's click on one of these results to see what a full record looks like. Okay, I've clicked on a result that looks promising. It's from a deed book. Here's what I'm seeing now. On the left side of the screen is the original document image. I can see it's a handwritten deed from 1852. The writing is in that old-style cursive that's always a bit tricky to read. On the right side is the AI-generated transcript. It's typed out line by line trying to match what's on the page. Above the transcript there's a button that says "Summarize the document". If I click that, the AI generates a summary in English, even if the original document was in Spanish or another language. The summary might say something like "This is a deed of sale. Sarah Collins sold land to John Mitchell, witnessed by Thomas Brown, dated March 15, 1852". Super helpful for quickly figuring out if this is the record you need. Now, notice on the image on the left, my search terms are highlighted in yellow. I can see exactly where "Sarah Collins" appears on the page. That's really handy when you're dealing with multi-page documents. At the top of the page, I've got several options. The first is "Names", "Transcript", "Information", and then a paper clip if I click on says "Attach to Person and Family Tree" or "Save to Source" or "Save to Source" box. All really useful features. But here's what I want you to do every single time. Look at that original image and compare it to the transcript. Read through it yourself. In this example, I can see that AI did a pretty good job. The names are correct. The date is right. But I noticed it transcribed March as "Mock" in one spot. Small error, but it shows you why verification matters. This is good genealogy practice anyway, right? We always want to look at original sources, understand context, evaluate evidence. Full-text search gets you to those sources faster, but it doesn't replace that critical evaluation step. Alright, you've now searched full-text search and examined a record. That's the basic workflow. But there's a lot more we can do to make your searches more effective and find exactly what you're looking for. Let's talk about search techniques and strategies. Okay, now that you know how to access full-text search and run a basic search, let's talk about how to actually get good results. Because there's a big difference between searching and searching well. I'm going to teach you some search operators and techniques that will transform your results. Think of these as power moves for full-text search. First up, quotation marks. We touched on this earlier, but let me explain it more thoroughly. When you put a phrase in quotation marks, like, quote, William Henderson, end quote, you're telling the system, find this exact phrase in this exact order. Without quotes, if you type William Henderson, the system searches for any record that contains the word William or the word Henderson, you'll get William Smith and Mary Henderson and all sorts of results you don't want. With quotes, you only get records where those two words appear together as a phrase, much more precise. Here's a pro tip. Try searching both ways. Sometimes your ancestor's name is written Henderson comma William or William H. Henderson or W. M. Henderson. The quoted phrase might miss those variations. So, do one search with quotes for the exact name, then do another search without quotes, but using some of the other operators I'm about to teach you. Second, the plus symbol. Let's say you're looking for John Smith, which, okay, good luck, right? There are thousands of John Smith, but your John Smith was a carpenter in Ohio. Here's what you do. In the key word field type, quote, John Smith, end quote, plus symbol carpenter. The plus symbol means must include. So, you're saying, find me records with the phrase John Smith that must also include the word carpenter somewhere on the page. You can use multiple plus symbols, like John Smith plus carpenter plus Ohio plus 1840s. Now, you're really narrowing it down. This is incredibly useful for common names or when you need to filter through a lot of results. Third, the minus symbol. This is for exclusion. Let's say you're searching for Elizabeth Johnson, but you keep getting results for a different Elizabeth Johnson who lived in the same county. Maybe the one you want never lived in Springville, but the other one did. You could search, quote, Elizabeth Johnson, end quote, minus symbol Springfield. That minus sign means exclude any records that mention Springfield. Really helpful for filtering out results you know aren't your ancestor. Fourth, the question mark. This one's for spelling variations. Historical records are full of creative spelling, especially of names. Is your ancestor's name spelled Delaney, Delaney, Dulaney, or Dulaney? Here's the trick. The question mark replaces a single character. So, if you search D, question mark, L, A, N, question mark, Y, the system will find Delaney, Delaney, Dulaney, Dulaney, Dulaney, any spelling where those question marks can be any single letter. This is super useful for names that have multiple spellings or when you're not sure if there's an extra letter somewhere. Fifth, the asterisk. The asterisk is a wild card that can represent multiple characters. Let's say you're looking for someone named Smithson, but they might be recorded as Smith, Smithson, Smithers, or Smithfield. Search for Smith's asterisk mark. They'll find anything that starts with Smith and has any combination of letters after it. Now, be a little careful with this one. If you search for something too broad like S asterisk, you'll get thousands of results for every word that starts with S. But used strategically, it's really powerful. You can also use it at the beginning or in the middle. Like, asterisk S-O-N. We'll find Johnson, Henderson, Jameson, etc. Here's how I used this recently. I was looking for an ancestor whose first name I'd seen written as both Francis with an I and Francis with an E. I wasn't sure which one was correct. So, I searched F-R-A-N-C asterisk Henderson. Got results for both spellings, plus I also found a Frank Anderson, which turned out to be the same person going by a nickname. Six, using the keyword field strategically. The keyword field is more powerful than you might think. You're not limited to names. You can search for occupations, blacksmith, farmer, teacher, merchant, relationships, widow, orphan, heir, son of, daughter of. Legal terms, indenture, last will and testament, deed of sale. Geographic features, near the creek, adjoining lands, along the road. Anything else that might be mentioned in records. Let me give you a real scenario. Let's say you know your ancestor owned land near something called Maple Creek in a certain county, but you don't know exactly when or in which deed book. You can search keywords, quote, Maple Creek, end quote, place, your county or state, year, your date range, and boom. You'll get every deed, survey, or other record that mentions Maple Creek, and you can look through those to find where your ancestor's land is referenced. Or, let's say you're trying to find witnesses to events. Marriage witnesses, will witnesses, deed witnesses. Search keywords, witness, plus symbol, your ancestor's name. You'll find any document where someone witnessed something involving your ancestor, or where your ancestor witnessed something for someone else. This is especially powerful for collateral research. Finding your ancestor's associates, neighbors, and relatives by seeing who they witnessed for and who witnessed for them. Seventh, searching by collection using DGS numbers. Okay, this one's a little more advanced, but it's incredibly useful. Every collection on FamilySearch has something called an image group number, or DGS number. It's like a unique ID for that collection. Why would you want to search a specific collection? Well, let's say you know your ancestor lived in a specific county during a specific time period, and you want to search all the deed books from that county, or all the court records, or all the church records. Instead of searching across all 6,000 collections, which might give you thousands of irrelevant results, you can narrow your search to just that one collection. Here's how. First, you need to find the DGS number. There are two ways. Way one, go to the regular FamilySearch catalog, find the collection you want, and look at the URL. You'll see a number. That's the DGS. Way two, from the full text search page, click Browse All Collections, and find the collection. The DGS number will be displayed. Once you have the DGS number, put it in the image group number filled when you search. Now, your search only looks through that specific collection. I use this all the time. I'm doing a comprehensive search of an ancestor in a particular county, I'll pull up the list of available collections for that county, and then I'll search each major collection individually. Deeds, wills, court minutes, marriage records, using their DGS numbers. It takes a little more time up front, but the results are way more focused and relevant. Eighth, use filters effectively. After you run a search and get your results, don't forget about those filter options at the top of the results page. You can filter by collection, narrow down to just deed records or just probate records. Year, tighten your date range if you're getting too many results. Place, drill down to a specific county or city. Record type, focus on a specific type of document. These filters are cumulative, so you can apply multiple filters to really zero in on what you're looking for. Let's say you search for Thomas Brown in Virginia and get 500 results. That's too many to go through efficiently. So, you apply filters. Year, 1850 to 1860. Place, Albu Merrill County. Record type, probate records. Now, you're down to maybe 20 results, and that's way more likely to be relevant. Much more manageable. Alright, those are your basic search techniques. Quotation marks for exact phrases. Plus and minus symbol for including or excluding terms. Question marks and asterisks for spelling variations. Strategic use of keywords for occupations and other details. DGS numbers for collection-specific searches. And filters to narrow your results. Master these techniques, and you're going to find records. Fully trained records, you never would have before. I promise you that. Okay, I want to share a success story that really shows the power of full text search when you use it strategically. This happened about two months ago, and I'm still excited about it. I've been researching my third great-grandmother, Margaret O'Brien. She married Patrick Sullivan in New York in 1858, and then the couple moved to Pennsylvania sometime before 1870. I have their marriage record. I have them in the 1870 census in Pennsylvania, but I had absolutely no idea where Margaret came from originally. No parents' names, no location in Ireland, nothing. This is a classic Irish genealogy brick wall. Common names, pre-civil registration in Ireland, limited records. I'd searched everything I could think of. Naturalization records, death records, church records in New York. Nothing gave me Margaret's parents or her place of origin. So, I decided to try a full text search strategy focusing on their early years in Pennsylvania. My thinking was, maybe they knew other Irish immigrants. Maybe there were witnesses to events. Maybe Margaret had siblings who also immigrated. I went to full text search and searched Pennsylvania deed records for the county where they settled. First search, Patrick Sullivan for 1860 to 1880. Got about 40 results. Started going through them. Found Patrick buying land in 1869. Great! That confirms when they arrived. Found him selling part of that land in 1874. Nothing groundbreaking yet. Then, I searched for Margaret Sullivan in the same collection, same date range. Third result down was a deed from 1872. I clicked on it and I'm looking at this handwritten document and the AI summary says, Deed of sale. Margaret Sullivan, widow of Timothy O'Brien, sells land to John Murphy. Wait! Margaret Sullivan? Formerly Margaret O'Brien? Married to Timothy O'Brien? I'd assume Patrick was her first husband, but no. She'd been married before. Now, I'm excited. I searched for Timothy O'Brien in the same county. Find a probate record from 1857. Timothy O'Brien died, leaving widow Margaret and three children. Then, I searched Timothy O'Brien in New York records around the same time. Find a ship passenger list from 1852. Timothy O'Brien, age 28, traveling with wife Margaret, age 24, from County Cork, Ireland. Exact ages match with later census records. In 30 minutes of searching, 30 I went from having no idea about Margaret's origins, to having her first husband's name, her approximate birth year, her departure port in Ireland, three children from her first marriage that I didn't know existed, a connection to County Cork. And here's the thing. None of these records were indexed in a way that would have connected them. The deed didn't mention Patrick Sullivan, so searching for him wouldn't have found it. The passenger list wasn't linked to her second marriage. Full text search, let me search for different name combinations in multiple collections quickly, and find documents that traditional index searches would never have surfaced. Now, do I have 100% proof that this is the same Margaret? Not yet. I still need to verify, build the case, maybe find more documentation. But I went from no leads to multiple solid leads in one afternoon. And that's the power of having AI as your research assistant. Not your researcher. I'm still doing all the analysis and verification work, but as an assistant that can help me find needles in those haystacks. That's what I want for all of you. Those breakthrough moments where suddenly a brick wall starts to crumble. Alright, let's wrap this up. We've covered a lot of ground today, and I want to make sure you walk away with clear action steps. Just a quick pause and update of the end of this episode on full text search collection numbers. I watched a family search webinar yesterday, and they said there's 7,757 collections as of November 20th, 2025. Okay, my friends, we have covered a ton of information in this episode. Let's recap the key points before I give you your homework. FamilySearch Full Text Search is an AI-powered tool that makes over 2 billion previously unsearchable records searchable by creating full text transcripts of handwritten documents. It's free with your FamilySearch account. It includes over 7,757 collections

Strategic keywords searching for occupations, relationships, and other details. DGS numbers for collection-specific searching. Filters to narrow your results. And remember, always verify AI-generated transcripts against the original images. AI is your research assistant, not your researcher. Use full-text search to find records faster, but still apply good genealogical practices for evaluation and proof. Alright, homework time! I'm giving you three levels. Pick the one that matches where you are in your comfort level. Beginner. If this is your first time with full-text search, your assignment is simple. Set up access to the tool. Go to FamilySearch. org. Make sure you're signed in. Then click on Search and go to Full Text. And do one search for a known ancestor. Search for someone you already have records for, just to practice the interface. Look at one record. Compare the transcript to the image and get comfortable with how it works. Intermediate. If you're ready to dig deeper, use the wildcard operators, question marks and asterisks, search for an ancestor whose name might have multiple spellings. Try at least three different spelling variations and see if you find any new records. Download one record you find and add it to your research files with proper citation. For those of you who are ready to really put this to work, do a multi-collection sweep of one ancestor in one location. Search at least three different collection types, deeds, court records, and probate records, using DTS numbers for each collection. Document three new findings about your ancestor or their community. Bonus points if you find them as a witness to someone else's transaction. And please, please, share your results. Come over to our Facebook group, Ancestors and Algorithms, AI for Genealogy. And tell us what you find. Did you break through a brick wall? Find an unexpected record? Discover a new connection? We want to celebrate with you. As always, if you have questions, you can email me at ancestorsandai at gmail.com. You can also visit our website at ancestorsandai.com. If you're enjoying the podcast, I'd be so grateful if you could leave a rating and review wherever you listen to podcasts. It really does help other genealogists find the show. And hit that follow or subscribe button so you don't miss the next episode. Thank you so much for spending this time with me today. Thank you for being part of this journey as we figure out together how to use these amazing new AI tools responsibly and effectively in our family history research. Remember, AI is your research assistant, not your researcher. Use these tools to work smarter, but never stop being the genealogist who evaluates evidence, verifies sources, and build solid proof. Until next time, this is Ancestors and Algorithms. I'm your host, Brian. Happy searching. And may your brick walls start to crumble.