Remember those early months of the pandemic when the world seemed to grind to a halt? Between sourdough starter experiments and home workouts, my wife and I discovered solace in streaming series from around the world. Pretty soon, we needed a system to keep track of what we’d watched and what we hoped to watch next. Enter the humble Google Sheet—our digital log of viewing pleasures.
It began as a simple way to avoid the “What should we watch tonight?” conversation that can drag on longer than an episode itself. We’d jot down titles, the service we watched them on, and whether we liked them enough to recommend. As months went by, our list grew to encompass everything from the sweet British show “Last Tango in Halifax” to the Danish noir “Borgen” to the quirky Israeli thriller “Fauda” to Italian gems like “Imma Tataranni.” It was our personal archive, a testament to the endless hours spent exploring the worlds of others while we sheltered in place.
Eventually though, the list got a little stale. We’d plowed through most of the shows our friends had suggested. We’d looked at the “best of” list on news sites and social media. But many of the recommendations were generic and often completely off-mark.
I wanted something smarter—recommendations that acknowledged my love for international crime dramas and historical epics but didn’t default to whatever was currently trending. My tastes may be eccentric, but they are mine. That’s when I thought: why not ask ChatGPT?
First Contact: The Generalist List
I took the plunge and pasted my sprawling Google Sheet into a prompt. The question was simple: “Based on these shows and movies I’ve enjoyed, what else should I watch?” ChatGPT returned with an impressively eclectic list that spanned continents and genres. It suggested contemporary thrillers like “The Diplomat” and “Bodyguard,” period dramas like “Alias Grace,” and even dark comedies I hadn’t considered.
What amazed me was the nuance. The chatbot didn’t just rehash whatever was trending; it found shows with similar sensibilities—a blend of deep character studies, political intrigue and dark humor. It recommended several titles I had already seen but not recorded in my spreadsheet (such as “Mindhunter” and “Bosch”), which validated that it had indeed captured my taste. It also surfaced lesser-known gems I’d never heard of. By the end, I had a list of about 20 titles that felt both fresh and eerily aligned with my viewing habits.
I shared this new list with a friend, and he had the obvious next question: “These look fantastic, but where can we watch them?” And that was the kicker.
The Availability Quandary
On paper, ChatGPT’s list was a dream. In practice, however, some of the shows weren’t actually available in the United States. Others required subscriptions to platforms I didn’t have. So I asked ChatGPT: “For each of these recommendations, where can I stream them?” It tried to answer, but the responses were hit-or-miss. In some cases, it correctly identified that “Narcos” was on Netflix and “Goliath” was on Prime Video. In others, it noted that a series wasn’t currently streaming anywhere in the U.S. But for more obscure titles, the information was outdated or incomplete. Some shows hadn’t been licensed in the U.S. at all.
I didn’t want to invest time hunting for a series only to discover it wasn’t available. My friend’s comment made me realize I needed to adjust my approach. Instead of asking ChatGPT to imagine the best possible matches across the entire entertainment universe, I needed to constrain my request: “Based on my original list, recommend shows currently available on Netflix” (or Amazon Prime, or Apple TV, or MHz). By narrowing the query to a specific service, I could ensure that anything recommended would be accessible to me.
The Light Bulb Moment: Constraining the Query
So I tried the new plan. First up: Netflix. I asked ChatGPT to recommend shows based my list that were similar in tone and were currently available to stream on Netflix in the U.S. The results were almost immediate and strikingly relevant. Gone were the unavailable British imports. Instead, the suggestions included “Mindhunter,” “Ozark,” “Narcos,” “Kleo,” “Top Boy,” “Alias Grace,” and “Giri/Haji.” Each recommendation was accompanied by a short description, an explanation of the logic behind the recommendation, and verification that it was indeed on the platform.
I repeated the process for Amazon Prime Video, Apple TV+, and MHz Choice, narrowing the recommendations with each prompt. For Prime Video, I got “Sneaky Pete,” “Bosch: Legacy,” “Patriot,” and “Jack Ryan”—all series that matched my love for crime procedurals and espionage. For Apple TV+, I was directed toward “Slow Horses” and “Bad Monkey,” both of which scratched my itch for dark humor and spy stories. And on MHz Choice, a platform I hardly touched before, the recommendations included Nordic noirs like “The Bridge” and “Bordertown,” as well as Italian mysteries like “Murders at Barlume.”
What took some time (and a lot of patience) initially became streamlined. Instead of sifting through entire catalogs, I asked for specific matches that aligned with my tastes and were available on services I already had. It was like having my own personal streaming concierge.
The Payoff: A Curated Watchlist That Works
Was it a perfect process? Not quite. It required multiple refinements. Sometimes a show that I expected to be available had recently left a platform. Other times, new licensing deals meant a show was unavailable in my region. But by cross-referencing ChatGPT’s suggestions with each platform’s current listings, I ended up with a watchlist that was far more tailored and, more importantly, watchable.
The final lists were robust. Netflix offered a mix of international espionage and time‑bending mysteries. Prime Video delivered gritty procedural dramas and action-packed series. MHz Choice became my go‑to for European crime. And Apple TV+ surprised me with top-quality original thrillers. The cross-service approach felt like a mix of recommendation engine and research assistant, delivering better results than what I usually see on social media or aggregated “Top 10” lists.
Why This Approach Matters
Most recommendation algorithms assume a one-size-fits-all model, and streaming services rarely work together to provide a unified watchlist. In addition, the catalogs of Netflix or Prime are vast—much of what might be interesting is buried. But by using a combination of my own data (in that trusty Google Sheet) and the large language model’s ability to parse patterns in my preferences, I created a self-curated ecosystem. It’s like training a digital sommelier but for TV.
The key lessons?
- Keep your data handy. Knowing what you already love helps a recommendation system pick up on nuances in genre, tone and pacing.
- Constrain your query. Ask for shows available on the platforms you actually use. There’s little point in daydreaming about the perfect show if it isn’t on your streaming menu.
- Validate availability. Even ChatGPT sometimes pulls from outdated licensing information. Cross-check recommendations with official listings or tools like JustWatch.
- Don’t be afraid to iterate. It may take a few rounds to get it right, but the payoff is a list that’s far more satisfying than a generic “Best of” article.
By the time I was done, I’d not only discovered a slew of new favorites—I’d also refined a system that makes recommendations personal, practical, and incredibly rewarding. It’s a reminder that with a little creativity (and the right AI tool), even something as mundane as scrolling for a show to watch can become a thoughtful, data-driven experience.
So the next time you’re staring at a menu of endless options, consider giving this approach a try. Curate your own list, feed it into ChatGPT, and let it filter the world for you based on what you already know you love. Just remember: it’s not about finding the most famous shows. It’s about finding the right ones—and making sure they’re ready to watch tonight.