When I was still on Facebook, I was getting endless ads for Fin, an “assistant” that claimed to use AI plus humans to take care of any task you can imagine. My favorite thing about the ads was the comments people would leave. People who tried it were pretty unhappy. But let’s warm you up first.

Their website opens with a video (blocked by my ad blocker) from a guy telling Fin he loves Phil Knight’s biography. Send copies to the homes of two friends. Just use my credit card.

Disclaimer: given the price of this service, I haven’t tried it. I’m going by what I see and reviews others have left.

Whoa. Wait.

My immediate impressions are:

  • This isn’t going to use much AI. A human is going to listen to that and take care of it. So it’s like having a human assistant take care of stuff for you.
  • They have a lot of saved info about you. Credit cards. Phone numbers. People’s addresses. They better have the best server security company on the planet. They’d better heavily background check everybody who works there to make sure they’re not doing the wrong things with this information.
  • Are they logging into MY Amazon account to do this? They have my usernames and passwords?

The video and website show more examples.

The video also shows a woman asking Fin to please schedule her day tomorrow. She wants to take a SoulCycle class, take the dog to the groomers, find a quiet place to take a call from 11-11:30am, and drop by City Lights Bookstore (so SF!).

Definitely not AI. Definitely humans.

The website has a live feed of people making requests.

Still not AI. Someone is going to call Amazon. They have enough information to have (legal) access to your credit card account. And they will cancel your credit card. This sure assumes a LOT of set up work.

Let’s not leave out this guy.

He has an idea he wants on his to do list. He wants to do a flash mob of well-to-do millennials marrying refugees who have been banned from entering the US… and he wants to create a Tinder-style app to “matchmake” the refugees.

What? I’m choking on the hipsterness of this. He has HOW much spare money to build this app, deploy it, and get refugees to some non-US location (since they’re not allowed in the USA) to be mass married. And why?

Who is this guy? In another video example, he’s building a table. Table builder and part time flash mob imagine-er?

So what were people’s comments and reviews?

All negative. Except some VCs. They love it! Wow!

Let’s talk about what it costs since that’ll help explain negative reviews. As of when I’m writing this (and assuming they haven’t closed down yet) they are charging $1 per minute used by the assistants with a minimum of 120 minutes per month. Their free trial is 60 minutes free.

How long does it take to get things done? Let’s go to the reviews.


Because it’s not AI.

If you asked Fin to do things that AI could understand, it would be cheaper. But you also wouldn’t need it. Take the guy who wants to make the flash mob. Does he have to tell Fin this? Can he voice record his idea into his phone for free? Can he ask Google to set a reminder for him to do this in a month?

If you need a real assistant, you need a human. And if you need a human, can you pay less than $1 per minute? Well of course you can. There are many areas of the USA where people would find $12/hr to be an amazingly paid job. Let’s say we really like these assistants and we pay them $15/hr.

Get a few of your hipster (or non-hipster) friends together and give an American human a job. That’s about $31K per year. Split it among 5 of you. You each pay around $6000 per year to have a semi-dedicated personal assistant. Need it because you’re a busy businessperson? Write it off.

Think of it as an assistant time share. You get 20% of the assistant’s time, not quite 90 min a day since he or she will work Mon to Fri, 7 hours a day with a paid lunch break. 🙂

Therein lies the typical startup problem.

We have a solution but what’s the problem? Who is the solution for? It is only for rich VCs who aren’t counting how long this took or what it cost? Then it’s very elite and that’s your user base.

What if we dropped the whole AI part of this? If I want AI, I’ll talk to Google. What if this were truly YOUR personal assistant. Timeshare Personal Assistant. Fin hires Americans in America for fair wages. Marks them up so they make some money but not to $60/hr.

Because there is a critical mass point at which people don’t want to pay for tasks to be taken off their plate if they see tasks happening inefficiently or expensively. If it takes 32 minutes for an assistant to make your dentist appointment partially due to what they need to “learn” to do this, does $1/minute make sense? What about $25/hr (and Fin pays workers $15/hr). That 33 minute tasks now cost you $13.75. Now is having the assistant do it worth it?

Get UX research involved.

A company like this needs more and better UX research to determine things like:

  • Target audience. Who is willing to pay to have a “piece” of a live personal assistant? Who are these people? What are their typical needs, their typical tasks? How much of their personal information (credit cards, SSN, etc…) are they willing to give up?
  • Where is the most value? What does the assistant need to take care of that would make the customer feel like the money spent was worth it? Clearly setting up a dentist appointment might not be worth it. UX research would help us learn how people value their own time. What is taking care of a task worth financially?
  • What might someone budget for a week or month or assistant help? How can we price this so it makes sense to our target audience? Sure, it can be the private jet for VCs but those people also have money to just hire assistants. Personas will help us determine who this is for, why they need it, and what they’ll spend without feeling ripped off.

Make sure there is a real product market fit.