I’ve got a favour to ask! If you enjoy this, please double check that you’ve liked the video and subscribed to the channel! Thats a small way you can help me carry on doing this ❤🙏 really appreciate you! 🩷 Subscribe: ua-cam.com/users/DamoBird365 ☕ BuyMeACoffee: buymeacoffee.com/DamoBird365 🔗 LinkedIn: linkedin.com/in/DamoBird365 🐦 X (Twitter): x.com/DamoBird365 📱 TikTok: tiktok.com/@DamoBird365 🙋♂ Feedback: forms.office.com/r/4EqE7VHVfHua-cam.com/users/sgaming/emoji/7ff574f2/emoji_u1f64f_1f3fe.png
@@paolovr1970 definitely lots of amazing use cases now that we have multimodal. The fact that you could extract a tracking id from an image, turn that into structured output and then look up a system via api, very exciting. I believe this will tie in nicely with agents too.
Using Azure AI Service is a lot cheaper, or was until recently. I haven't checked the new AI Foundry tool yet! $250 a month is a huge cost way out of range of almost everyone including businesses. Azure AI service is 0.01 per parsed document max 12 pages. The tokens allowed in Azure AI Service was a miserly 4000 or so(= 12 pages of text in a pdf). I will try a comparison, myself, soon.
I covered docs in Azure vs AI Builder here ua-cam.com/video/fLHmEwcg8Jo/v-deo.html but recognise there are many ways to entry. Just because it’s cheaper, doesn’t mean it’s more accessible. Some orgs will be sitting on included AI Builder credits through premium learn.microsoft.com/en-us/ai-builder/credit-management#number-of-ai-builder-credits-by-license others will be heavily invested in Azure. It’s definitely a consideration but the inclusion of multimodal in AI Builder is fantastic to see parity between low/pro code tooling. Feel free to report back your findings. It’s valuable to have these options explored.
I’ve got a favour to ask! If you enjoy this, please double check that you’ve liked the video and subscribed to the channel! Thats a small way you can help me carry on doing this ❤🙏 really appreciate you!
🩷 Subscribe: ua-cam.com/users/DamoBird365
☕ BuyMeACoffee: buymeacoffee.com/DamoBird365
🔗 LinkedIn: linkedin.com/in/DamoBird365
🐦 X (Twitter): x.com/DamoBird365
📱 TikTok: tiktok.com/@DamoBird365
🙋♂ Feedback: forms.office.com/r/4EqE7VHVfHua-cam.com/users/sgaming/emoji/7ff574f2/emoji_u1f64f_1f3fe.png
Thanks very Interesting I think you can use it for the flow of goods control before loading on the truck for shipping.😊
@@paolovr1970 definitely lots of amazing use cases now that we have multimodal. The fact that you could extract a tracking id from an image, turn that into structured output and then look up a system via api, very exciting. I believe this will tie in nicely with agents too.
Using Azure AI Service is a lot cheaper, or was until recently. I haven't checked the new AI Foundry tool yet! $250 a month is a huge cost way out of range of almost everyone including businesses. Azure AI service is 0.01 per parsed document max 12 pages. The tokens allowed in Azure AI Service was a miserly 4000 or so(= 12 pages of text in a pdf). I will try a comparison, myself, soon.
I covered docs in Azure vs AI Builder here
ua-cam.com/video/fLHmEwcg8Jo/v-deo.html but recognise there are many ways to entry. Just because it’s cheaper, doesn’t mean it’s more accessible. Some orgs will be sitting on included AI Builder credits through premium learn.microsoft.com/en-us/ai-builder/credit-management#number-of-ai-builder-credits-by-license others will be heavily invested in Azure. It’s definitely a consideration but the inclusion of multimodal in AI Builder is fantastic to see parity between low/pro code tooling. Feel free to report back your findings. It’s valuable to have these options explored.
I get text and finished reason but structured output does not exist (red underline). Did you set sth specific?
br
It’s in my demo 👍 you’ll need logic to handle this.
ua-cam.com/video/ONjioJG7YQQ/v-deo.htmlsi=RB3PHR0KI7VGGuAj&t=560 watch from here