A new kind of assistant that runs in the cloud

#2 Product of the WeekNovember 28, 2017
Featured Embed

UPDATE: Since the time of this post, we have lowered prices 25% to 40% and dramatically improved the quality of service across all major categories.

➡️ Read more in THIS POST.

  • Jeremy Peronto
    Jeremy PerontoBusiness Development, Amazon

    Aspirational; Incredible demo.


    Expensive on a per task and overall basis; Poor definition of an "effective minute".

    I want to believe in Fin. I was turned off by the price, but was wowed by the demo video and took it for a spin. I was so excited by the potential that I even found a spot for it on my home screen.

    My first request was to send flowers to someone for under $50. The ETA given was 95 minutes. Three hours later I had gotten no feedback. Concerned that I going to end up spending more on the service than the flowers, I asked for a status. I got feedback that the task would take longer but I was only being charged for the active minute. My ETA increased to 10 hours. As I watched the cost balloon up to $38, I cancelled the task.

    My second request was to add a brewery I had just heard about to my "Restaurants to try" list. In my mind, this should basically be an AI task that would be used when I asked Fin to make future bookings. I assumed it would take next to zero time. Instead, it took 18 effective minutes.

    $56 later, I had a single restaurant added to a list.

    I believe in the utility Fin is selling. But after limited use, the vision and the execution feel very far apart. And as much as I want to be an optimist, the founder responses to user comments here on Product Hunt read as tone deaf and are in the process of converting me to a pessimist.

    Jeremy Peronto has used this product for one day.
    Comments (1)
  • Christopher Leach
    Christopher LeachProgrammer and Student

    Can basically get anything done for you


    Can be expensive depending on how you use it

    I find that the offer they give us hunters is the best way to use Fin. I used it when It had a monthly rate and loved it and love it now, but I save a lot of money because I don't request often but when I do it really helps. Don't know why so many detractors.

    Christopher Leach has used this product for one year.
You need to become a Contributor to join the discussion.
Johan Le Bray
Johan Le Bray@johan_le_bray · Full Stack Developer, Shopmium
Would love to try it, but it's way too expensive for a casual usage. Interesting to see a company take the AI assistant to the professional side.
Brandon McConnell
Brandon McConnell@brandon_mcconnell · Full-Stack Developer + UI/UX Designer
@johan_le_bray My thoughts exactly. $120/week for the lowest tier, "occasional" usage, just seems very impractical to me. Even for semi-frequent use, I'd be hesitant to pay $120 per *month*.
Will Wright
Will Wright@williamwright · Co-founder @Kindaba
@johan_le_bray @brandon_mcconnell "Most users spend between $500 and $5,000 per month." Yikes! 💸
Brandon McConnell
Brandon McConnell@brandon_mcconnell · Full-Stack Developer + UI/UX Designer
@johan_le_bray @williamwright news just came out that Fin is a collaboration project between Facebook and Venmo execs. No surprise there 😂
Samuel W. Lessin
Samuel W. LessinMaker@lessin
@williamwright but not all 😄 We like to give people a sense that if you use Fin very heavily costs add up... but plenty of people just do a key task or two a month and spend far less - the price is on demand / based on use!
kortinaMaker@kortina · Co-founder of &
@williamwright @lessin (also, cc @rrhoover re "what have you learned building Fin) A few of the biggest learnings I have had working on Fin involve time. 1. As a Fin user, I tend to really underestimate how long many tasks take. Often I'll talk to people who say things like, "it would only take me 30s to make this restaurant reservation if I did it myself" - sometimes I'll even think things like this. But at Fin we've measured all sorts of tasks like this thousands of time, and are beginning to understand all of the actual steps involved (eg, seeing if a restaurant is on opentable, finding available times when the exact time asked for was not available, perhaps calling the restaurant if they are not bookable online, adding an event to your calendar, notifying your guests, etc) and developing a much better understanding of all the real work involved and time it takes to do things like this. When you look at all of this work that goes into seemingly 'simple' requests, you can start to see how we often tend to grossly underestimate even the best case scenarios for how long they will take. 2. As someone who has worked many sessions as a Fin agent, it is remarkable to me what 'real' productivity feels like -- working an 8 hour day on the clock is really intense compared to a lot of desk jobs that I had when I was much younger. My sense is that there is a huge inefficiency with much of time spent at desks just waiting for work to arrive ( @lessin and I talk about this here ). I expect the labor utilization at many desk jobs is probably like 20% so a '40 hour week' is more like an '8 hour week.' The upshot of this is that if you look at the cost of a single request with both of these misperceptions, it may feel high. But, if you actually measure the work that happens closely (and see that it takes longer than you might think) and consider the potential utilization loss when full time workers are sitting idle, Fin is actually quite economical. Looking at myself as an example, I use about 10 hours of work per week at Fin, and I get more work done than I used to with a full time assistant at 40 hours per week. Something I've personally been spending lots of time on is trying to illustrate this sort of time accounting on our website / through our service / to our customers. (Finally, I'll mention that this is just a straight up $ cost comparison I am talking about -- there are other massive benefits to using Fin that you can't account for in this sort of analysis, eg, (i) Fin is totally on demand, so there is no sunk cost if you use only 5 hours in a light week and then scale up to 50 hours the next week and (ii) Fin eliminates hidden costs of context switching, so while there may be instances where I theoretically could do a task myself in 5 minutes that would take Fin 7 minutes to do, outsourcing non-critical work to Fin keeps me focused on the most critical work that uniquely requires my attention and creativity -- this is invaluable)
Derek Shanahan
Derek ShanahanPro@dshan · Three Hex, Cloud Unicorn
I've tried it three times, haven't been able to get value out of it. Last time I got a $160 bill for finding and booking an in-home dinner chef for a trip my fiance was taking with friends, for 8 emails and 20 mins on the phone - way too expensive for normal use.
Abadesi@abadesi · Head of Maker Outreach at Product Hunt
@dshan whoa that is pricey but I guess its about how you value your time. $160 to only spend 20 mins on phone vs an hour may be enticing for some. Who are your target audience @lessin @kortina? Are you aiming for HNWI?
Samuel W. Lessin
Samuel W. LessinMaker@lessin
@dshan @kortina @abadesi hey all... so, we bill at $1 per minute on demand, so you only use what you pay for. This is much less expensive than having a full time assistant in most cases, but isn’t cheap for sure (good help never is!) — in the case above my sense is that if it took us 160 min to do all those things it would have taken at least as long for anyone else!
Susan Ho
Susan HoHiring@susan_ho · Founder, Journy (
@dshan Yikes! That's also why I think it's beneficial to use the right service that specializes in what you're looking for. My company, Journy (, specializes in travel experiences and because we've done stuff like hire in-home chefs for travelers in 75+ destinations around the world, it'd take us a fraction of the time to help with something like that. Because we specialize in travel and reservations, our fees start at $25 per day of your trip (we too are updating our pricing, but the max you'd pay is $50 per day of travel plus an additional $5 per person above 4 travelers).
tommyent@tommyent · Freelancer
@dshan @kortina @abadesi @lessin how is this cheaper than having a full time assistant? At @40 a week fin is costing $115,200 a year. This seems to be on the high end especially considering it's virtual.
kortinaMaker@kortina · Co-founder of &
@dshan @abadesi @lessin @40 @tommyent per my comment above, it does net out cheaper because all Fin work is on the clock / you only pay for results, not idle time. I get more done in 10 hours of work per week of Fin service than when I had a full time 40 hr / wk assistant.
Ryan Hoover
Ryan HooverPro@rrhoover · Founder, Product Hunt
Ever since YC-backed Magic launched 3 years ago, there's been a wave of personal assistant services hitting the market. From my experience, most were either too expensive or unreliable. I used Fin during the beta ~1 year ago but it never worked into my habits and after several price changes, I was uncertain how much I would be charged for each request. Side note: to properly evaluate the cost you need to consider how much your time is worth. Services like this are ultimately selling time, removing trivial tasks (e.g. make reservations at a restaurant for at 6pm on Friday) to free up one's calendar. What's been your biggest learning over the last few years building Fin, @lessin and @kortina?
Samuel W. Lessin
Samuel W. LessinMaker@lessin
@rrhoover we have learned a ton Ryan... and written a bunch about it at
Samuel W. Lessin
Samuel W. LessinMaker@lessin
@rrhoover kortina wrote a particularly good 2017 roundup recently with a bunch of key insights at
kortinaMaker@kortina · Co-founder of &
@lessin @rrhoover See also my comment above ( ) for a few other key insights regarding time that are not captured in this post
james huang
james huang@weiluenhuang · Product @ Microsoft
Congrats on the launch! I've tried Magic before. I saw my engagement with Magic decrease with time. I recognized that it took a lot longer for them to do something than it would take for me, which meant I only used Magic for really taxing and time consuming endeavors. Eventually, I churned out of Magic because of the infrequency with which I interacted with Magic. If it's only really large tasks, it's never top of mind and eventually I forget about it. I'm wondering how this service is different than what Magic was offering previously or other scheduling services like clara or x?
kortinaMaker@kortina · Co-founder of &
@weiluenhuang I think there is a real frequency of use threshold to get you hooked on assistant service. I'm personally interacting with Fin 5+ times per day, which nets out to about 10 hours of work per week. Although you could use an assistant like Fin in a more occasional manner, where I think Fin really shines is in getting to know you as well as a human assistant that works with you for years -- we invest deeply in 'shared memory' tools for our team that make this possible, and I think enable a higher quality of service than you might get when you using something more infrequently: you can read a bit more about this shared memory concept here Re clara + x -- these seem like great tools for scheduling. One thing I personally like about Fin is that it will do pretty much anything for me -- this often involves picking up the phone and calling someone, or emailing someone or using a website on my behalf (for more on this, you can see the same blog post under "Humans are the universal API.")
Alan Kane
Alan Kane@alankane · Pure USA
@kortina what is the machine learning component?
Andrew Mason
Andrew Mason@andrewmason · CEO Descript, former Detour / Groupon
I think I've tried every personal assistant app in Fin is by far the best - it's the real deal, incredibly useful.
Samuel W. Lessin
Samuel W. LessinMaker@lessin
@andrewmason thx man! We are working hard at it!
Dharmesh Shah
Dharmesh Shah@dharmesh · Founder and CTO, HubSpot
@andrewmason Wow, dude. That's high praise. Running off to try it out now.