Which AI workflow should you automate first in your ecommerce business?(Hint: It’s probably not what you’re thinking)Slobodan Manić just dropped some serious knowledge on the 3 AI topics every ecommerce professional needs to master:
- The ONE workflow that gives you instant ROI
- How to spot AI tools worth your money vs. expensive distractions
- What AI adoption actually looks like for smaller teams (spoiler: it’s simpler than you think)
Plus, if you want to go deeper into agents and production-ready workflows with n8n, Sani is hosting a workshop
At the micro level, we need 100% accuracy so that we can safely focus on the trends and the themes at the macro level, without worrying about precision.
🚀 Exciting Opportunity! 🚀
I'm looking for short-term #Nextjs help for our new startup project. 🌟
Flexible location (European timezone preferred). 🌍
If you're interested or want more info, let's chat! 💬
#RemoteWork#Startup
✨ New [ 👗 Try On 2.0 ] on https://t.co/lAyoqmSBRX
Finally it works REALLY WELL:
- pick any clothing, from a screenshot or laid out photo, upload it
- click take photo, you can prompt the location of model, light (sunset), time, anything like normal etc.
- AI model will wear the clothes you uploaded
- textured prints now work REALLY well too, not perfect, but very close
- it drapes almost any clothes properly to the body now, still buggy photos but 60-80% is good enough
I think it's good enough for Shopify sellers now cc @tobi, if anyone here has clothes they want me to try on Photo AI's models, reply a pic and I'll render them and reply to you
You can take 1000 pics for $39/mo on Photo AI so it's very economical ~$0.04 per photo, with error ratio maybe $0.06 per good photo you can use for your store without having to do a model photo shoot anymore
Render time is about 1-2 minutes per photo but you can run them in parallel of course
Last year try on mode wasn't precise enough and it couldn't do textures well, it would detect the clothes, prompt for them, and use the photo as ControlNet reference but that didn't actually work so well as people made it out to be, now it's 1000x better and actually works
Nothing is perfect of course, but I'm trying!