How to use AI for genealogy

Essential AI prompts for family historians

Carole McCulloch writes:

As we begin to get comfortable with our conversations with AI tools, we very quickly learn the importance of effective prompting.

Here are seven tried-and-tested prompts that every family historian should master:

1. Recreate historical context:

‘Based on these known facts [list evidence], describe my ancestor’s daily life including housing conditions, food, and social customs that would have affected their family.’

2. Build character profiles:

‘Given their documented choices and circumstances, what personality traits and motivations might my ancestor have possessed? I want to understand them as a real person.’

3. Create life timelines:

‘Help me create a chronological timeline incorporating these documented events, and suggest what might have happened during the gaps based on typical life patterns.’

4. Generate interview questions:

‘Create thoughtful interview questions for my elderly relative about our family history, focusing on daily life details that wouldn’t appear in official records.’

5. Suggest story angles:

‘I have these documented facts about my ancestor. Suggest five different storytelling approaches that would engage modern readers while remaining historically accurate.’

6. Write vivid scenes:

‘Using these verified historical facts, write a historically accurate scene showing my ancestor in their documented environment.’

7. Find story titles:

‘Based on this family history, suggest engaging titles that capture both the personal drama and historical significance.’

See examples of Carole McCulloch's family history slides, created using AI tools, via her SubStack

Tips about prompts:

 

• Build your own ‘prompt library’ - to store your most frequently used prompts. Start with the selections above. Remember, these prompts are conversation starters, not rigid formulas.

Use PromptCowboy.ai to turn your prompt into a really effective prompt for pasting into your AI tool of choice, such as ChatGPT etc.

• Use PromptCowboy.ai to help you create a more refined and effective prompt that you can paste into your regular Gen-AI tool (eg ChatGPT, Perplexity and so forth).
• Generate research logs & research plans - Carole McCulloch's prompt ideas above are particularly useful for when working on family history narratives. Carole is also very keen that AI is used to create more efficient research processes too - ask your AI tool to create a template research log for you, for instance, or to generate a research plan. 
• Think of it as a brainstorming partner that turns vague intentions into actionable pathways.

Try this: create a three-step research plan prompt

1. Frame a clear research question.
 Start broad and enter your ancestor’s name, place, and time period into a conversation with your favourite AI tool and request a research plan.

2. Add known evidence and ask AI to suggest next steps.
 Refine with detail, such as known records, to uncover gaps and patterns. Request archives, repositories, or databases relevant to the case.

3. Anchor in history.
 Ask AI to frame your ancestor’s life within larger historical events. Generate name variants for search flexibility.

The workflow blueprint for AI-assisted genealogy

When working with AI in genealogy, the key is not to overcomplicate the process. A simple, repeatable workflow ensures that you save time without sacrificing quality. Think of it as a rapid blueprint – one you can adapt for any task, whether drafting an ancestor profile, building a locality guide, or shaping a family biography. 

1. Define your goal
 Begin by clarifying what you want to achieve. Is your focus on a single ancestor, a migration story, or a community snapshot? A clearly stated goal prevents you from drifting into endless tangents and keeps your AI prompts sharp and targeted.

2. Feed AI accurate, concise data
 The quality of the output depends on the clarity of the input. Summarise your records – names, dates, places, occupations – before sharing them with AI. Think of this step as preparing a research bundle that AI can transform into structured content.

3. Request structured outputs
 Avoid vague prompts. Ask the AI for specific formats: a narrative profile, a locality guide with section headings, or a bullet-point research plan. Structured outputs are easier to verify, adapt, and reuse later.

4. Verify and enhance with archival sources
 AI accelerates the drafting process, but credibility still comes from the archives. Cross-check AI’s suggestions against parish registers, census records, and trusted databases. This stage is where human expertise ensures accuracy and depth.

5. Repurpose into other formats
 Once verified, don’t stop at one version. Expand into a detailed blog post, condense into a newsletter, or craft a visual social media snippet with an image or map. Each output reaches a different audience while drawing on the same research foundation.

Consistently apply these steps for reliable, shareable genealogy outputs.

2 key AI family history pitfalls

When using AI in genealogy, it’s easy to fall into subtle traps that can undermine your research if you’re not careful. Two in particular show up time and again.

Pitfall #1: The Speed Trap

• It’s tempting to be dazzled by how quickly AI can generate text. One click, and suddenly you have a draft profile, a locality guide, or even a whole article.
• But speed can be misleading. Just because AI delivers information in seconds doesn’t mean it should be accepted at face value. Think of these rapid outputs as possibilities, not facts.
• The real work begins when you step back, slow down, and apply your genealogical eye, checking whether the narrative actually matches the evidence.

Pitfall #2: The verification shortcut

• Perhaps the most dangerous pitfall is the urge to skip fact-checking.
• AI is convincing; its language flows smoothly, and it often sounds right. But genealogy demands accuracy. If you don’t return to parish registers, census data, or trusted archives, you risk embedding errors into your family history. Verification is not optional – it is the foundation that keeps your work credible and trustworthy.

Next steps: Practical implementation

Try it for yourself:

1. Choose one task this week – an ancestor profile, a locality guide, or a research plan – and let AI take the first pass.
2. Start small by applying the sample prompts to one ancestor, building both confidence and a foundation for future work.
3. From there, experiment with creating your first multi-format content series and establish a simple verification checklist to keep accuracy at the forefront.
4. Over time, refine your process by developing template prompts, building a tailored prompt library, and designing reusable content formats. 
5. As your confidence grows, explore advanced strategies such as combining multiple AI tools in a single workflow, sharing your methods within the genealogy community, and measuring both time saved and quality improved.
6. Each step builds on the last, ensuring your AI-assisted genealogy practice is not only faster but also richer, more reliable, and deeply rewarding.
    

Keep learning about AI for family history

• Visit the Coalition for Responsible AI in Genealogy website: CRAIGEN.

Be sure to explore the 'Education' resources on the CRAIGEN website for advice to help us all use AI in a responsible manner within the field of genealogy.

• Join a specialist community on social media platforms, such as Carole McCulloch's group: www.facebook.com/groups/ancestryintheageofai

John Beaumont writes:

If you’ve been researching family history for a while, you’ve already lived through one revolution, the shift from physical to remote research.

Not so long ago, being a family historian meant spending hours in record offices, ordering CDs, making trips to cemeteries, and carefully handling fragile parish registers. Those of us who lived through it remember the thrill of discovery, but also the long waits, travel planning, and frequent dead ends.

Then the internet changed everything.

The first leap: how the internet gave us remote research

As more records became digitised and indexed, the foundations of genealogical research began moving online. Within a matter of years, you could browse the 1911 Census from your kitchen table, order birth, marriage and death certificates through the General Register Office website, and access decades of newspapers via the British Newspaper Archive.

For many, it was transformative. For others, especially those who’d honed their craft in physical archives, it was initially met with skepticism. Could clicking through a database really compare with the rigour of handling original documents?

It turns out, it could, if you did it right.

As online tools matured and coverage expanded, “remote research” became the norm. Newcomers to genealogy today often start with Ancestry, Findmypast, or FamilySearch before they ever step into an archive – if they ever do.

But for all the gains, remote research brought new challenges:

• Incomplete or poorly indexed records
• Fragmented records
• Monetisation of sources 
• Proliferation of unverified information

So, while the internet has democratised research in some ways, we have also inherited new problems. And that brings us to where we are today, on the edge of the second great leap forward in family history.

The second leap: assisted research

In the last couple of years, Generative AI (Gen-AI) has become mainstream.

Tools such as ChatGPT can answer questions, write summaries, suggest strategies, translate documents, and even assist in drafting family narratives.

These systems should not be used to replace your skills as a genealogist. Instead, they should be used to amplify them.

ChatGPT is your assistant, not an expert. Just like the internet, when used effectively, it will transform the way we approach our family history research.

I like to refer to this new phase as assisted research; it is the next logical step from remote research. It’s not about handing over the hard work to a machine; it’s about having a capable helper to speed things up, brainstorm ideas, and free us up for more analysis and interpretation.

What can Gen-AI do for a family historian?

Here are some real-world examples of how genealogists are already utilising Gen-AI, along with some straightforward prompts.

• Transcribe and analyse:
  [Prompt] “Please transcribe this 1856 marriage record from Lincolnshire and tell me what information it contains about the bride’s family.”
   
• Provide historical context:
  “What was life like for a coal miner in County Durham in the 1880s? I’m writing about my great-grandfather who lived there.”
   
• Translate and identify:
  “Translate and explain this Latin phrase from a Church of England burial record in 1792: ‘sepultus est in coemeterio parochiali.’”
   
• Summarise a document:
  “Summarise this three-page probate document from 1870, Somerset. I’m especially interested in the heirs and land mentioned.”
   
• Help with a narrative:
  “Turn these research notes into a short family story about my 3x great-grandmother, who moved from Belfast to Liverpool around 1850.”
   
• Brainstorm a brick wall:
  “I can’t trace Mary Ann Pearson after the 1871 Census in Yorkshire. She was born in 1843 in York. Can you suggest where to look next?”
   
• Create a research plan:
  “Create a step-by-step research plan for tracing my ancestors in the East End of London between 1800 and 1850.”
   
• Check and review your work:
  “Review this family history summary for clarity and historical accuracy. Does anything seem inconsistent or unclear?”
   
• Provide DNA analysis:
  “I share 215 cM across 12 segments with a match on Ancestry. What is the likely relationship, and how should I approach verifying it?”
   
• Review and analyse images:
  “Please describe and date this family photo taken in Hull. Can you suggest clues from the clothing or setting?”
   
• Create images:
  “Create a photorealistic image of a Victorian midwife in Manchester around 1890, standing outside a working-class terraced home.”

But it has limits…

Gen-AI: it’s just a guide

And this is important. Although some new versions can execute limited internet searches within a conversation, Gen-AI tools are not search engines.

They don’t “look up” information in the way that Google or Bing does; they generate it based on patterns in their training data in a similar way to your mobile phone generating predictive text.

That means they:

• Sometimes invent sources.
• Are often 100% confident even when they are wrong.
• Have no concept of true or false.
• Can reflect bias from training data or even your own prompt wording.
• Produce images that are interpretations, not historically accurate.
• Raise questions of ethics, privacy, and potentially copyright.

In short, AI is not a source; it’s a guide and assistant.

Use it to brainstorm, explore, and summarise, but always check its suggestions against original sources.

And remember, it always remains your responsibility to be accurate, not Gen-AI’s.

Common mistakes (and how to avoid them)

Even experienced users can trip up. Watch out for:

• Accepting results without verification, even if the information seems plausible, it may be incorrect or fabricated.
• Mistaking fluency for authority, a beautifully written paragraph doesn’t make it accurate.
• Assuming it knows everything, its knowledge is limited to the training data and doesn’t update in real-time.
• Thinking one answer is final, try rephrasing your prompt; you may get a better response, or a very different one.
• Using it as a shortcut to “proof”. Gen-AI can guide your thinking, but it can’t replace primary sources or traditional analysis.

How genealogy websites use Artificial Intelligence

[Advice in this section from Family Tree.]

In addition to the AI websites we can use ourselves, which can have numerous uses not just related to genealogy, the major websites, software and app providers, and countless other family history products and services also offer AI tools.

For example, popular genealogy website MyHeritage offers AI tools including:

• MyHeritage Deep Nostalgia™ - video reenactment technology to animate the faces in still photos and create high-quality, realistic video footage.
• LiveMemory™ - allowing you to turn photos into video clips with AI, and bring memories to life.
• AI Time Machine™ - which gives you the chance to create an AI avatar of yourself and travel through history
• DeepStory — technology which actually makes your family photos speak!

Meanwhile, Ancestry offers the AI Assistant, which (2025) is described as 'a new beta tool provided by Ancestry that uses artificial intelligence to assist you with Family History and DNA-related questions and research. It can answer general Family History and DNA-related questions, offer suggestions, and guide you through your research based on information from our help content.'

Of course, some of these tools are more of a novelty and may not directly impact our research, but they provide good examples of just what artificial intelligence can do, and how the technology is quickly evolving.

A new era for genealogy

Gen-AI isn’t replacing good research; it’s helping us do it better. And just like the early days of the internet, some will hesitate. That’s okay. The key is to experiment, learn what it’s good at, and stay in control.

Because this second great leap, from remote to assisted research, isn’t about changing our goal.

It’s about removing the friction, so we can focus on the parts that matter: analysing records, building stories, and understanding our families.

Summary thoughts from John Beaumont on using AI for genealogy

We’ve gone from lifting bound volumes in record offices to browsing scanned images from home, and now to having our own personal digital assistant at our side. Each step has brought us closer to more efficient, inclusive, and thoughtful research.

So, if you’re feeling curious, give Gen-AI a try.

Ask it a question.

Test its advice.

Let it help you shape your narrative.

Just don’t forget: the records are still out there waiting to be found, and you, not Gen-AI, are the one uncovering the story.

The second great leap is here. Let’s use it well.

Have you tried FamilySearch's Full-Text Search?

Family Tree advice:

Above the advice by John Beaumont and Carole McCulloch has largely concentrated on the AI large language models. AI is increasingly woven into the family history websites we use as well. Below is coverage of FamilySearch's Full-Text Search.

No longer will we be relying on indexes, restricted to specific fields. With a full-text search, genealogists will increasingly be able to search on whatever words and people they wish. Read on to see how the FamilySearch Full-Text Search works ...

Why is FamilySearch Full-Text Search tool useful?

By and large, the current state of genealogy searching is that we have the functionality to search on the defined fields. And, due to the historically laborious nature of indexing, only certain details would be indexed - and thus searchable. With the FamilySearch AI powered Full-Text Search you may search on any names or words of interest to you. You are no longer dependent on the data fields defined by an indexing project.

How to try out the FamilySearch AI Full-Text Search:

• Go to https://www.familysearch.org/en/search/full-text/ - and log in (using your usual FamilySearch log in)
• Then simply enter names and or keywords of interest to you, and define the date parameters of your search as you wish. If you know it, you can refine the records you wish to be searched further by including the Image Group Number reference.
• You are able to search nearly 2 billion records on FamilySearch using the Full-Text Search.

 

---

About John Beaumont

John Beaumont's article 'Genealogy is changing: don't get left behind' was first published in the September 2025 issue of Family Tree. He presented on this topic (June 2025) to Family Tree Plus members and the recording of his presentation is available in the Family Tree Plus members' library. The material by John Beaumont is an original work written by John Beaumont with assistance from ChatGPT, ensuring it is based on original thoughts, thorough verification, and robust research.

John Beaumont combines years of genealogical research with a focus on Generative AI, helping others learn how to use new tools to enhance traditional methods through classes, talks, and online videos. More about John at www.beaumont-genealogy.com

About Carole McCulloch

The material by Carole McCulloch is original work, created with the assistance of generative AI tools, and was first published in her article 'How to get started with AI for genealogy - confidently' in the November 2025 issue of Family Tree, and 'Speeding up family history with AI for genealogy - safely and strategically' (Family Tree, December 2025).

Carole McCulloch is an experienced educator and enthusiastic explorer of AI for genealogy.
Through her Essential Genealogy Academy, she designs masterclasses and self-paced courses that guide family historians to use AI with confidence in both research and ancestral storytelling.
She regularly shares practical skills and fresh approaches as a presenter and course creator. Learn more at coachcaroleonline.thinkific.com.
 

The 'How genealogy websites use Artificial Intelligence' was written by Family Tree (published September 2025 issue, in a sidebar on John Beaumont's article). The FamilySearch Full-Text Search wording was written by Family Tree and first published on the Family Tree website.

Web page last updated 24 February 2026.