• I started using Instinct in October 2026. The useful part isn't only asking a question and getting an answer. It's being able to hand over a small job, give it the context it needs and stop carrying it around in my head. Here are some ways I've been trying it.

    Reminders that let me unload my brain

    A lot of my day is made up of things I need to remember later. Make a call. Be ready for a technician. Bring something along. I can text the thought to Instinct while it's fresh instead of hoping it comes back to me at the right time.

    The reminder doesn't have to stay where I first put it. I've moved a call from one time to another and then to a different day. When a reminder arrives and I'm not ready, a reply like "snooze 30 minutes" moves it out of the way without losing it.

    The longer-term reminders may be even more useful. I asked for one six months out that includes when a maintenance job was last done. I also set a monthly reminder about something stored in a closet. By the time those reminders arrive, I may have forgotten the whole conversation. The context comes back with the task.

    A packing list I don't have to remember all at once

    I had a trip coming up and asked Instinct to send my packing list when it reminded me to pack. Then I kept adding items by text as they occurred to me, using "pl" as a shortcut.

    That meant I could add a headlamp in one moment and a dishwashing sponge later without opening a list or starting over. When the packing time changed, I moved the reminder and the list together. The small things had somewhere to go as soon as I thought of them.

    A weather warning with a job attached

    I don't always want a weather report. Sometimes I want to know when the forecast means I should do something.

    I asked Instinct to watch for an overnight freeze so I could disconnect water hoses and move equipment indoors. I also asked for a different temperature threshold for another property, with no message if there was nothing to report. The useful part is tying the forecast to the action, rather than leaving me to remember why I checked it.

    From sailing screenshots to a map

    I had a sailing plan spread across screenshots, with stops and overnight times. I sent the screenshots and asked, "Make a map of this."

    Instinct turned them into one overview with numbered stops and arrival and departure times. I could see the sequence without going back through the screenshots. It wasn't a navigation chart or a safe course to follow on the water. It was a clearer picture of the plan.

    Getting past photo links that wouldn't open

    Someone preparing a report couldn't open the Google Photos links I'd emailed. I wasn't sure Instinct could help, but I asked it to find the original filenames behind the links so we could work from those instead.

    It recovered the filenames and descriptions, then we worked on the links themselves. The result was an email draft with a description followed by a usable link for each photo. The links were checked without signing in, and the email stayed a draft. What started as a frustrating photo problem became something ready for me to review.

    Understanding a trip from a shared spreadsheet

    I knew someone had emailed me a trip spreadsheet, but I remembered the sender wrong. I asked Instinct to find it and tell me what was on the itinerary that day. It found a likely sheet and checked with me before treating it as the right trip.

    Once that was settled, I asked about the local weather and for short videos of the day's sights. A schedule of places became a much easier day to picture. It also kept a useful distinction clear: the spreadsheet showed what was planned, not proof of where the travelers actually were.

    Catching a small mistake before a flight

    While helping with a flight booking, Instinct spotted a misspelled passenger name. I got it corrected before the trip and came back with the updated itinerary.

    That was more useful than simply repeating the flight details. I've also asked for check-in help, brought boarding passes back up by text and followed an inbound plane when I wanted to know how a flight was progressing.

    Finding dinner that fits the actual evening

    One restaurant question wasn't just "Where should we eat?" I wanted to know whether a place took reservations, whether we could sit outside and whether we could eat before sunset.

    Other times I've wanted a very specific food or a comparison of actual happy-hour food and drink prices. I've asked Instinct to check the menus and offers rather than stop at a general list of recommendations. The interesting part is giving it the whole situation instead of doing each little search myself.

    Turning email into something I can act on

    I've asked Instinct to find proposals in email and explain the differences together, rather than make me flip between messages. I've also asked what an invoice was for and what mattered in a conversation before deciding whether to archive it.

    When an email needs a reply, we can gather the details and work on the wording before sending. I still review it. Getting the information together is often as much of the job as writing the message.

    Giving old blog posts a clearer voice

    I asked Instinct to edit my older blog posts for clarity without changing what I meant. It showed me the wording first, applied the edits after I approved them and helped remove a duplicate post. The posts I kept stayed at their original dates and links.

    Location-based reminders

    I asked to be reminded when I reached or left certain places: arriving at a job site, leaving it, and getting home. All three reminders fired.

    And this post itself

    This started as a long numbered list. After reading it, I didn't think it put the interesting examples first or explained them well enough. I asked Instinct to tighten it, then to make it more story-driven. I also pushed back when reminders sounded like an afterthought, because unloading my brain is a big part of the value for me.

    Now I've asked it to review new uses each day and suggest additions without private details. I approve the wording before it updates the blog. Figuring out what belongs here, and how to explain it, has become another real example.

  • There are a few things I really want AI assistants to handle:

    1. Full voice control of my devices.
    2. The ability to interrupt the assistant while it’s talking.
    3. Everyday email tasks: read a message, archive it, draft a reply or delete it.

    Siri is frustrating to use and far too limited for what I need. With full voice control, I could get much more done while I’m in the car.

  • Google NotebookLM is blowing my mind. I’ve created several notebooks, and I keep finding more ways to use them.

    I added PDFs, web links and copied text, then asked questions about those sources. My favorite feature was Audio Overview. It turned the material into a very natural conversation between two speakers.

    I tried it with a legal filing and newspaper articles. One newspaper PDF also contained an unrelated article that I hadn’t cropped out. The audio summary started discussing that article too. I removed it from the sources and generated a new overview.

    That was a good reminder to check what I’m feeding into these tools. In my testing, changing the sources didn’t automatically update the audio. I had to generate it again. I could also download the overview as a WAV file.

    Listening to two speakers discuss a complicated subject made it much easier to follow. What a time to be alive!

  • Friends were planning a camping trip in Telegram, and the messages went on for weeks. Someone would bring a griddle. Someone else had a generator and a big tent. Others volunteered for meals. Arrival times were mixed in with everything else.

    I took screenshots of the conversation, then captured those screenshots together as one image in my iPhone photo album. I uploaded it to Claude and ChatGPT and asked, “Organize this information.”

    The results were amazing. Both tools grouped the plans by person, showing who was bringing what and when everyone was arriving.

    A long, scattered conversation turned into a plan I could actually follow. Very cool.

  • I was following a lawsuit and wanted a clearer explanation of the material. I uploaded a news article and a legal filing as PDFs in separate Claude chats, then asked it to summarize and simplify each one.

    It was amazing how much easier the documents were to follow.

  • I’ve been using Gemini to summarize YouTube videos before I watch them. It helps me decide whether a video is worth my time and gives me an idea of what it will cover.

    It’s been extremely helpful.

  • We ordered steel cut to several sizes. When it was ready for pickup, we needed to decide which truck or trailer to send.

    I copied the emailed order into ChatGPT and Claude. Both estimated that the steel weighed close to 4,000 pounds.

    That gave us useful information for choosing what to send. I love finding practical uses for AI like this.

  • I asked the support team for our fleet-tracking software a simple question: How do I log a truck’s oil change before it’s due? Instead of a direct answer, they sent me a 10-page PDF of the user manual.

    I uploaded the manual and my original question to ChatGPT. Within seconds, it gave me the answer I needed. I tried the same approach with Claude and got a useful answer there too.

    That’s the kind of support I wanted: a clear answer to my question, not just a PDF to work through. I’d like to see MyGeoTab offer that kind of help.