Representative appearances

• VoiceVision RAG - Integrating Visual Document Intelligence with Voice Response — Suman Debnath, AWS

  Custom tool for converting images to LLM prompts.

• Inside Reducto: From Pivot to Fortune 10 Customers, How to Build AI Products | CEO Adit Abraham

  a much harder problem than sort of the canonical traditional OCR.

• JustPaid - AI-native B2B billing and reporting automation platform

  we have our AI and our OCR and that OCR basically parses all this

• Better RAG Throught Better Data with Reducto CEO

  traditional OCR has always been horrible at.

• Reshaping Workflows with Lisen Kaci, CEO & Founder @ Discrepancy AI

  It's using a technology called OCR. So our initial product was how do we improve OCR document processing. for OCR, every single type of bank statement requires its own model you have to fine-tune. And that's really where our use case shines is that we can remove all the this 80% human we can remove we can save 80% on human costs for reviewing these documents and you don't need to have a separate AI model for every document. So with our customers, yeah, they've seen a reduction in costs for reviewing these documents by over 80%. So for yeah, for income verification, we've seen over 80% decrease in cost. And that's really where our use case shines is that we can remove all the this 80% human we can remove we can save 80% on human costs for reviewing these documents and you don't need to have a separate AI model for every document. So with our customers, yeah, they've seen a reduction in costs for reviewing these documents by over 80%. So for yeah, for income verification, we've seen over 80% decrease in cost. So like you're replace you're not replacing another software tool, you're placing labor time in essence. So right now they're using both. That's the interesting thing that right now they're using the software tool, but because the accuracy of it is so low that for the average hit rate, so that is is like under 10%. So like they have the AI models and then on top of that 90% of these documents still have to be reviewed by a person because they're so different because real world documents like someone takes a screenshot like their thumbs in it when you're dealing with real world documents like that's where OCR sucks really and that's why these new omni channel models are better. your AI native is that what you're displacing in terms of like the budget for the customer correct it's both so current implementations are these clunky OCR tools plus like human employees required to supplement it so right now they can replace both of those with just our tool and maybe retain just 10% of the people that the past implementation had. So it's really replacing both and yeah we're seeing like 80% cost reductions for our customers for all for their document processing workflows.