Show up to Thanksgiving with something only you can bring (and its not a pie)

A 10-minute writing exercise that turns census records into a story your family has never heard.

Imagine showing up to Thanksgiving with something beyond the usual pumpkin pie.

Not the family stories everyone’s heard a hundred times, but something new. A story about an ancestor’s Thanksgiving that you wrote yourself.

And imagine reading it aloud while the sweet potato casserole makes its way around the table.

Most of us have names and dates. Birth certificates, census records, maybe a faded photograph tucked in a drawer. What we struggle with is seeing them as real people who sat at actual tables, passed actual dishes, and lived through actual November days.

That’s where this 10-minute exercise comes in.

The writing prompt that opens doors

Here’s the writing prompt: Describe your ancestor’s Thanksgiving Day.

Simple question. But it opens something most genealogy work doesn’t touch. It forces specificity. It demands imagination grounded in research. It turns a name in a family tree into a person at a table.

And with a little help from AI, we can ground that imagination in actual historical reality. No guessing required.

How the writing exercise works

Step 1: Pick one ancestor

Choose someone from roughly 1900-1950. This range gives us enough historical records to work with while still feeling distant enough to require imagination.

Write down:

• Their name

• Where they lived (city and state)

• The specific year you’re imagining

• Their occupation

Step 2: Let AI do two jobs

Open Perplexity.ai (it’s free) and ask it two questions:

1. Historical weather lookup: “What was the weather in [city, state] during Thanksgiving week [year]?”

2. Occupation decoding: “What did a [occupation] in [city] in [year] actually do day-to-day? What would their income and social circle have been like?”

These two pieces of information transform the exercise.

Weather grounds the scene in reality. Not every Thanksgiving was crisp and perfect in my state. Some were rainy. Some were unseasonably warm. Some had early snow that kept people home.

Occupation tells us about resources, social circles, and what was actually accessible. A seamstress in Pittsburgh had different food access than a railroad worker in Nebraska. A schoolteacher in rural Georgia sat at a different table than a factory foreman in Detroit.

Step 3: Write for 10 minutes

Set a timer. Don’t overthink it.

Write about what might have been on their table. Who might have been there. What the day felt like. Let the weather and occupation details guide you, but let your imagination fill in the rest.

Don’t aim for perfection. Aim for specifics. Aim for one true detail that makes this Thanksgiving theirs, not a generic historical reenactment.

Why this works

This exercise works because it combines three elements that rarely come together in genealogy:

Research (the facts you already have)

AI assistance (the historical context you can find quickly)

Creative writing (the imaginative leap that brings it all to life)

Most of us get stuck doing only research. We collect facts but never transform them into stories. Or we write stories that feel too generic because we’re missing the specific details that make a moment real.

The weather and occupation details (things AI can help us find in seconds) give us the frame. Then our 10 minutes of writing hangs the story on that frame and makes it sound like us, and not AI.

An example: my grandmother Mildred in 1930

Here’s what I started with from the 1930 census:

• Mildred John, 26 years old, single, never married

• Lived with her mother and sister at 155 East Ninth Street, Conshohocken, Pennsylvania

• Worked as a telephone operator for Bell Telephone on Fayette Street

• Brothers were married and lived in the same town

• Parents were first generation Americans whose parents immigrated from Prussia (Germany)

That’s helpful, but it doesn’t tell me what Thanksgiving felt like for her.

The research that made the difference

I asked Perplexity about the weather first. Within seconds, I had this:

Twenty sources synthesized into one answer: Thanksgiving week 1930 in the Philadelphia region (which includes Conshohocken) was cold with temperatures in the 30s and 40s. There had been early snow in mid-November, but Thanksgiving Day itself was partly cloudy.

That detail about the early snow that had already come and gone tells me something. The ground would have been cold. The air would have had that bite that makes you want to stay inside near the stove.

Then I asked about telephone operators in 1930:

Another twenty sources summarized for me (thanks Perplexity!) gave me context: Telephone operators in 1930 worked long shifts connecting calls manually. The job required patience and precision. Most operators were young, unmarried women. The pay was modest but steady, important during the Depression. The work was isolating in some ways (headset on, focused on the board) but deeply social in others (hearing every conversation, knowing everyone’s business).

Suddenly, I could picture Mildred. Not just her name and age, but her hands on the switchboard cords, her voice saying “Number please,” the way she’d come home tired from talking all day.

And if you want to do the research within ChatGPT, I have prompts for that too. You can get them here:

Ready to do deeper research on your ancestor's Thanksgiving?

Turning research into story

Now with those details, I wrote a rough draft directly into ChatGPT. Not even in complete sentences, just the ideas I wanted to capture. Then I asked it to polish the draft while keeping my voice and holding it to 500-800 words.

The first version it produced was this (only the first two paragraphs for space):

Good start, but the narrator’s voice wasn’t quite right. I adjusted it a few times, asking for different perspectives and tones, and landed on this version:

Much better. This one feels like Mildred: a woman who’s capable, a little tired, glad for the day off but already thinking about Monday’s shift.

What comes next

My next steps are to read this aloud and revise it so it sounds natural in my voice (not ChatGPT’s). Then I’ll add a photo of telephone operators from the 1930s, an image of the house on East Ninth Street if I can find one, and maybe a simple map showing Mildred’s walk to work on Fayette Street.

The whole process (research, drafting, and polishing) took less than 20 minutes. And now I have something specific enough to feel real.

Want to try it yourself? I’m sharing the exact prompt I used.

Get the Thanksgiving Day in the Life Writing Prompt here.

Your turn

You already know the stories you’ll hear around the table this Thanksgiving. The one about Uncle Jim’s terrible driving. The time Grandma burned the rolls. The year it snowed three feet and nobody could leave.

How about throwing in a new one?

Pick an ancestor. Open Perplexity. Look up the weather and decode their occupation. Set a timer for 10 minutes and write.

That’s it. That’s the whole exercise.

And if this kind of prompted, AI-assisted story-making feels like exactly what you’ve been looking for, share it with a friend because they probably are looking for it too.

Happy chronicling!

—Denyse

P.S. Set the timer right now. Ten minutes. One ancestor. See what shows up on the page. Use the AI prompt to help you walk through it and support you. You don’t have to work alone at this.