'Do one thing, but do it well': Seeking feedback on my AI app

[theguy22]

Hi all,

I wrote a web app that generates a conversation starter for a match on a dating app.

The way it works is you upload screenshots of your match's profile/bio/prompts, and the app generates a tailored conversation starter based on the profile.

It's currently free and in its beta stage so would be nice to get some feedback.

Check it out here – dately.ai

I was kind of inspired by roomGPT which is an app that lets you upload your living room picture and it transforms it for you.

It made me realise the value of just solving one narrow problem, very well.

I thought I could do this for dating apps and apply the 80/20 rule to just focus on conversation starters.

Would greatly appreciate feedback. And if you know someone that could use it, please do share!

Cheers

 

Dislike ads? Remove them and support the forum: Subscribe to Fastlane Insiders.

[rwhyan]

Too little information to warrant a signup.

Let users try it for free before asking signup. Give some images or example openers, etc.

 

[circleme]

Social Signups are a BIG NO-NO in this kind of area (dating) imho. I would never want to connect my real name with a service like this. But that's just me.

Why do you need a Sign-Up anyways?

Every additional step that prevents someone from using your product/service/software/etc. that is not necessary is a bad step. Even if YOU want the contact information (e-mail, name, etc.) from your customer doesn't mean you have to gather that information. Except, your business model revolves around it.

My approach would be something like this:

Visit Site > Upload it directly via Drag & Drop (no signups, no forms, etc.)

That's it. Visit site, upload the file. It's like converting an .JPG to .PNG with an online tool.

I mean, everything depends on your business model and also on things like backend/hosting-costs as well as API-costs, but this way you would get way more users than with the current version with social signup.

Just my 2 cents.

Btw: I like the idea a lot. I'm not a big fan of B2C SaaS, but still, it's a cool idea which 100% solves a problem a lot of people have.

 

[Michal__]

Have you tested that these conversation starters work?
If you demonstrated that it works and created a freemium revenue model it could work.

 

Dislike ads? Remove them and support the forum: Subscribe to Fastlane Insiders.

[theguy22]

View attachment 51204
Social Signups are a BIG NO-NO in this kind of area (dating) imho. I would never want to connect my real name with a service like this. But that's just me.

Why do you need a Sign-Up anyways?

Every additional step that prevents someone from using your product/service/software/etc. that is not necessary is a bad step. Even if YOU want the contact information (e-mail, name, etc.) from your customer doesn't mean you have to gather that information. Except, your business model revolves around it.

My approach would be something like this:

Visit Site > Upload it directly via Drag & Drop (no signups, no forms, etc.)

That's it. Visit site, upload the file. It's like converting an .JPG to .PNG with an online tool.

I mean, everything depends on your business model and also on things like backend/hosting-costs as well as API-costs, but this way you would get way more users than with the current version with social signup.

Just my 2 cents.

Btw: I like the idea a lot. I'm not a big fan of B2C SaaS, but still, it's a cool idea which 100% solves a problem a lot of people have.

Thanks for the feedback! Interesting take, all the big dating apps have social sign-ups so I thought it might not have been an issue?

I do see your point about sign-ups in general adding friction to the product but I think I need it for the business model.

I was thinking to use a freemium model such that they they're allowed to start conversations with 3 people for free every week but need to upgrade to use it unlimited

I think I kind of just copied RoomGPT , you have to be signed in even to use the free version

How would you have approached the entire thing differently if you were me? And any advice on how to go about marketing this?

 

[theguy22]

Have you tested that these conversation starters work?
If you demonstrated that it works and created a freemium revenue model it could work.

I've tested and used them personally. They're decent but I'm still only using GPT3.5 on the backend rather than GPT-4. GPT-4 would do a much better job but I haven't got access to the API yet. Hence I thought I'd release this version as a Beta. The plan is defo to use a freemium model. Any suggestions how to go about it?

 

[theguy22]

Too little information to warrant a signup.

Let users try it for free before asking signup. Give some images or example openers, etc.

Thanks! That's fair feedback, I need to add some more clear information and perhaps a how it works page with a demo.

 

Dislike ads? Remove them and support the forum: Subscribe to Fastlane Insiders.

[circleme]

How would you have approached the entire thing differently if you were me? And any advice on how to go about marketing this?

1. not using registration at all
2. monetize it with ads only

I have serious doubts about the reusability of your software, to be honest. Perhaps it seems 'cool' for the first 1 or 2 use cases, but beyond that, people will get the hang of it. So, freemium or a free trial won't likely work. Even if you were to attempt to monetize it with a subscription model or something similar, people will eventually realize that it's just an API request at the end of the day.