Instant Invoice Recognition: Whitepaper

 

Invoice recognition without learning


Most invoice recognition systems require machine learning (time and many invoice with the same structure) or human interactive. This system provide results instantly without learning.

It also provides meta data for recognized invoices, that helps accounting software's to save data with better quality.

Who We are

Filip — CEO adviser

Site.pro, CEO
B1.lt, CEO
ClientsDay.com, Founder
20 years in IT industry

Alex — CTO adviser

Site.pro, CTO
B1.lt, CTO
Startups Adviser
20 years in IT industry

Prototype

(The system is protected by a Lithuanian patent)


 

Problem

  • 99% invoices are inserted manually
  • Accounting Companies (AC) needs it for reducing the costs
  • Existing solutions: require human resources, require machine learning, can't provide instat results
  • Unable to work with unknown invoice structure

Whole Market Size:

400M+ SMB's x 50 invoices (per month) = 2T+ invouces (per months)

Potential Market:

2T x 0.1% = 1B invoices to recognize

Potential Revenue:

1B x €0.01ct = €10M / month

 

Channels:

  • Direct sales
  • Free version (Freemium)
  • Google SEO
  • Word of mouth

Audience:

  • Accounting softwares (1:many)
  • Enterprises

Coverage:

  • Market PotentialPenetration

Know How

  • Provides meta-data with invoice (not only data)
  • Started as patented technology (URL i patenta)
  • Learning helps, but not required
  • Able to recognize instantly

Technologies:

  • PHP, PostgreSQL, OpenCV, Teseract OCR
  • Patented technology
  • Inhouse development

Risks:

  • Google and Amazon totally change their algorithms

Investment Timeline:

  • 6 mon: €90K — launch MVP and start sales (lt, ru, en)
  • 6 mon: €120K — position recognition
  • 6 mon: €180K — better quality, 5 markets, more languages
  • 6 mon: €180K — break even
 

Revenue Projections:

  • 6 mon: €0
  • 6 mon: €20K
  • 6 mon: €60K
  • 6 mon: €120K