While doing outreach for a test prep industry meet-up I’m hosting early next month in Boston, a Massachusetts-based tutor began pitching me on a platform he’d been building. I explained that we’d also built our own platform and looked forward to chatting about them and exchanging ideas at the event. After a bit of back and forth, he sent me this:
“While poking further around your site, I noticed that your colleague Thomas likely built your internal tool?”
He then inquired as to whether Thomas would be attending the event. In a Herculean display of restraint, I replied:
“Nope! I built it myself :)”
I had already introduced myself to this person as the founder of a tutoring company based in NYC who sits on the board of the National Test Prep Association. And yet, he went looking for the builder of the platform when he was already talking to her. The natural next step in our conversation would have been to ask how the platform came together. Instead, he supplied the answer.
After sending the most passive-aggressive smiley emoji to ever grace my outbox, I received an apology explaining that he’d assumed based on the bios on my website. Which, by the way, for anyone who’s keeping score, I also vibe coded entirely on my own.
He understood that I was the person running the business. (Not a huge stretch given that it’s literally named Kate Fisher Tutoring, and I am Kate Fisher.) But when it came to the technology powering the business, he assigned credit to a man on my team. (Thomas did empower me to learn all I have learned about AI and vibe coding, so if you’re reading this, Tom, hi, thank you, you’re the best.) I can’t say how much of that assumption was about our professional backgrounds and how much was about gender. I can tell you that he clearly respected my business and wanted my feedback, yet went looking for a man to discuss the technology I built. I do wonder how many of my male colleagues have been asked to introduce someone to the woman who presumably built their software.
Between LinkedIn and email, I receive approximately fifteen unsolicited messages a week from people offering products for tutors and tutoring companies. I can barely tell the pitches apart from each other anymore, and I am honestly tired of being approached as a prospective user of someone else’s solution to the problems they think I have as the founder of a tutoring and test prep company.
Some of these products promise that they will triple my tutors’ capacity, asserting that most tutoring businesses still use “manual lesson planning, paper-based tracking, and one-size-fits-all content.”
(Sir. It’s 2026. No one’s used paper-based tracking since long before Covid. Did you even ask what we’re currently using?)
Aside from the assumption that we are likely technologically unsophisticated and haven’t heard of Claude Code and Codex ourselves let alone used them, the problem with this premise is that individualization is presented as an element the software will provide rather than the very obvious cornerstone of our entire business model. I’m sure we’d all love to save a few hours a week on lesson prep. What remains unclear is how that generates the capacity for three times as many students per tutor. As far as I can tell, the platform, the apparent purpose of which is to increase student volume, cannot magically make these students available during the day.
Other platforms promise practice tests that “closely mirror the real exams” while offering homework tools and progress analytics. These solicitations inevitably close with, “All I’m asking is 30 minutes of your time, so we can show you our platform.” Excellent test-like questions delivered in an environment that mimics the real thing are genuinely useful. Evaluating the quality of those questions and whether the analysis reveals anything useful, however, takes time. Each of these solicitors inevitably wants 30 minutes of my time. I guess I can use the time freed up by the first platform to take on this endless parade of 30-minute meetings.
Yet another product has promised it will double student improvement, save me ten or more hours per week, and help me scale without hiring additional tutors. (It slices, it dices!) How, you ask?
“Imagine giving your students a 24×7 AI Avatar of yourself, turning boring PDFs into interactive spaces & monetizing your expertise.”
We are in the uncanny valley, folks. Alongside those promises to save me time, this platform offers to create an AI version of me that answers questions in my tone and brand. This assumes that my value as a test prep provider or tutor is about how I sound. And that’s part of it, sure. I am nothing if not a personality hire. But what I doubt this Sims 4.0 version of me can replicate is the split-second judgments that happen in a tutoring session based on the tutor-student relationship. A lot of tutoring—especially test prep—is informed by data. Excellent tutoring and test prep is informed by the relationship between the tutor and the student.
So. Here’s what I actually built and why I built it. I was using another platform (one whose founders I respect immensely and whose success I’m rooting for) and while it did a lot of what I wanted (organized content, allowed me to push homework), there were certain features I wanted that it didn’t have and others it had that my practice neither wanted nor used. I’ve also realized that I have very specific aesthetic sensibilities when it comes to software that this platform wasn’t meeting. Building my own also allowed me to incorporate other tools I’ve built like our study guide generator and score report builder. So: over approximately six weeks over the summer, I built an internal learning management system (LMS). There is an absolute joy in creating your own software and being able to immediately implement feedback rather than running it through a help desk. Sync to local time zones? Sure. Add a feature to time assignments per question rather than total? No problem. Study guide generator came back with an error code? Fixed.
As I’ve shared before, I don’t know how to code. And this is by far the wildest, coolest thing I’ve built to date. A colleague asked how much I actually had to write to execute it. According to Claude’s count of our sessions, I’d typed approximately 38,000 words across 577 messages. Most of those messages were short: “the reading keys look wrong,” “no, the other one.” Some were instructions; others were corrections. Over and over, I had to decide what I wanted, inspect what it produced, and explain what needed to change. I didn’t write the code, but I spent weeks making decisions about what the software should do.
So. I built software for my business without knowing how to write the code. Why do these pitches so rarely seem to account for that possibility? I could build this because AI handled coding I didn’t know how to do, while I supplied the decisions about what we needed and whether it worked. What that required was the understanding of what features are needed and how to implement them. Or knowing when a tool doesn’t work as efficiently as it should or is honestly extraneous to the goal. Or understanding how users will actually use it and being able to revise it when their behavior differs from your expectations. These pieces are much more aligned to how an educator already thinks. So why is an educator’s knowledge treated as background information while a presumed technical contributor receives authorship?
And, as it happened, in the opening anecdote of this piece, the educator was a woman and the presumed technical contributor was a man.
For those of you reading this in EdTech who are trying to build and scale a product aimed at tutors and tutoring companies, here’s what I’m asking about your pitch before deciding whether to give you 30 minutes of my attention.
- Is it addressing a problem that actually exists in my practice?
- Is it working with or against the resources we already use?
- Is it reducing the amount of work needed to keep everything functioning?
- Is it supporting the decisions we need to make about students on an individual basis?
- Does it offer enough value to justify adopting and maintaining another tool in an oversaturated market?
If you’re telling me you’re going to give me time back, here’s what I want that recovered time to go to: thoughtful preparation, noticing patterns, following up, and concentrating on the student in front of me. Also, more time with my kids and a nice spa day occasionally.
Want new essays on testing, admissions, and learning in your inbox? Subscribe on Substack.