Use unstructured data.

Transform the three most important data sources of your company.


Procurement

From order to billing, we help you to use the data in your supply chain. Check inbound and outbound information for plausibility, reduce the time for internal coordination processes.

Distribution

Automatically create highly customized offers based on existing contracts of potential customers. Recognize sales opportunities based on documents such as exposés or annual reports.

Customer service

Ultimately, your success depends on the satisfaction of your customers. Analyze your customer communications, solve customer problems, identify patterns and thereby best practices.

Customer Success Stories

Konfuzio, for example, helps 1:1 Assekuranzservice AG as a subsidiary of WWK Versicherungen, smoothly process insurance policies and premium invoices. The video shows the final presentation at WERK1, the center of the digital insurance industry in Germany.

Intelligent software

Mine your documents.


Assign

Gain time by classifying documents using our software. Konfuzio even works for complex processes and assigns requests to the right people.

Understand

Automatically transfer information in unstructured texts to your ERP system. Avoid typos and update the information stored in your database in real-time.

Validate

Check the content of inbound and outbound unstructured texts. Recognize deviations of already processed documents and reduce internal coordination efforts.

Kessler In IBAN: Fisch z.H. Schmalweg 88709 LIEFERSCHEIN Sehr vielen und Hiermit Leider 3342345 Wir Mit Ihr Datum 22.01.2019 22.01.2019 22.01.2019 22.01.2019 13.02.2019 13.02.2019 der 13.02.2019. senden freundlichen Kessler Joachim geehrter DE89370400440532013000 Geiger Meersburg Kehl Dank konnten Zoologiegroßhandel übersenden und 5 Ihnen 16, für 4925903245 Zoologiegroßhandel Fischfabriken Geiger Herr Artikelnummer 35413131 A001-323423 HONDA-555-023 Fischer-ABEW 23409234 A003-234234 wir 67122 Ihre diese Grüßen Geiger, wir die Bestellungen Altrip Ihnen von leider Artikel Gesamtbetrag GmbH Ihnen folgende nicht DE Menge gesondert GmbH & 500 bestellten Co 2582 vom 14 50 1 8 5 sofort & KG 22.01.2019 Artikel: Co 67138 Betrag 5.268,85 1.232,56 9.164,83 (HRB 15.03.2019 nach. 359,20 872,92 500,00 816,37 liefern. KG Artikel 60993)

1. Preprocessing

Before the Konfuzio algorithm can be applied, OCR is used to extracted the entire unstructured text of a document. Since many of our customers already use their own OCR software, we offer the flexibility to replace the Konfuzio OCR with another solution.

Using the example of a delivery note: Based on a new delivery, we would like to introduce you to our process. In the first step, e.g. a warehouse employee takes a picture of the enclosed delivery note (see picture) with your mobile phone.

Kessler In IBAN: Fisch z.H. Schmalweg 88709 LIEFERSCHEIN Sehr vielen und Hiermit Leider 3342345 Wir Mit Ihr Datum 22.01.2019 22.01.2019 22.01.2019 22.01.2019 13.02.2019 13.02.2019 der 13.02.2019. senden freundlichen Kessler Joachim geehrter DE89370400440532013000 Geiger Meersburg Kehl Dank konnten Zoologiegroßhandel übersenden und 5 Ihnen 16, für 4925903245 Zoologiegroßhandel Fischfabriken Geiger Herr Artikelnummer 35413131 A001-323423 HONDA-555-023 Fischer-ABEW 23409234 A003-234234 wir 67122 Ihre diese Grüßen Geiger, wir die Bestellungen Altrip Ihnen von leider Artikel Gesamtbetrag GmbH Ihnen folgende nicht DE Menge gesondert GmbH & 500 bestellten Co 2582 vom 14 50 1 8 5 sofort & KG 22.01.2019 Artikel: Co 67138 Betrag 5.268,85 1.232,56 9.164,83 (HRB 15.03.2019 nach. 359,20 872,92 500,00 816,37 liefern. KG Artikel 60993)

2. Classification

In the first step, the Konfuzio algorithm assigns the document to one or multiple categories. In addition to the classification of documents Konfuzio also allows the classification of text passages.

Using the example of a delivery note: The algorithm assigns the category “delivery note” to the picture, taken in the last step. In addition, the Konfuzio algorithm detects a conspicuous text passage that is not normally included.

Kessler In IBAN: Fisch z.H. Schmalweg 88709 LIEFERSCHEIN Sehr vielen und Hiermit Leider 3342345 Wir Mit Ihr Datum 22.01.2019 22.01.2019 22.01.2019 22.01.2019 13.02.2019 13.02.2019 der 13.02.2019. senden freundlichen Kessler Joachim geehrter DE89370400440532013000 Geiger Meersburg Kehl Dank konnten Zoologiegroßhandel übersenden und 5 Ihnen 16, für 4925903245 Zoologiegroßhandel Fischfabriken Geiger Herr Artikelnummer 35413131 A001-323423 HONDA-555-023 Fischer-ABEW 23409234 A003-234234 wir 67122 Ihre diese Grüßen Geiger, wir die Bestellungen Altrip Ihnen von leider Artikel Gesamtbetrag GmbH Ihnen folgende nicht DE Menge gesondert GmbH & 500 bestellten Co 2582 vom 14 50 1 8 5 sofort & KG 22.01.2019 Artikel: Co 67138 Betrag 5.268,85 1.232,56 9.164,83 (HRB 15.03.2019 nach. 359,20 872,92 500,00 816,37 liefern. KG Artikel 60993)

3. Extraction

We combine expert rules and NLP Rules dynamically to extract the information from a document in a structured way.

This new approach makes Konfuzio self-learning. This means that the algorithm starts with expert rules to learn more abstract rules over time. Thereby the algorithm can even recognize information if no expert rule existed.

Using the example of a delivery note: The quantities and the article numbers are recognized. Although for the Honda article number no expert rule was provided, the algorithm finds those article numbers.

Kessler In IBAN: Fisch z.H. Schmalweg 88709 LIEFERSCHEIN Sehr vielen und Hiermit Leider 3342345 Wir Mit Ihr Datum 22.01.2019 22.01.2019 22.01.2019 22.01.2019 13.02.2019 13.02.2019 der 13.02.2019. senden freundlichen Kessler Joachim geehrter DE89370400440532013000 Geiger Meersburg Kehl Dank konnten Zoologiegroßhandel übersenden und 5 Ihnen 16, für 4925903245 Zoologiegroßhandel Fischfabriken Geiger Herr Artikelnummer 35413131 A001-323423 HONDA-555-023 Fischer-ABEW 23409234 A003-234234 wir 67122 Ihre diese Grüßen Geiger, wir die Bestellungen Altrip Ihnen von leider Artikel Gesamtbetrag GmbH Ihnen folgende nicht DE Menge gesondert GmbH & 500 bestellten Co 2582 vom 14 50 1 8 5 sofort & KG 22.01.2019 Artikel: Co 67138 Betrag 5.268,85 1.232,56 9.164,83 (HRB 15.03.2019 nach. 359,20 872,92 500,00 816,37 liefern. KG Artikel 60993)

4. Plausibility

In addition to the classification, the algorithm offers the possibility to perform special plausibility checks. Manual checks can be automated cost-effectively.

Using the example of a delivery note: The algorithm checks amounts of the line items and compares it to the total amount. It recognizes a significant deviation of € 24.00.

Your success is our success

Transform your organization's top three data sources and any processes that support them.


Industrials

The more diverse the services and the larger the company, the more communication is recorded. The Konfuzio algorithm recognizes information even in e-mails, faxes, PDFs or photos.

Banks

Several clients are using the Konfuzio algorithm to classify and read relevant key data from a variety of financial documents.

Insurances

Insurance contracts of insurance companies use a variety of different layouts and formulations. With the help of Konfuzio, the key data is reliably read even in contracts of more than 100 pages.

Operation of the software

We offer three different operating options.


Cloud

We run our software in the cloud and provide you with you custom document service via the internet.

Our prices are based on the number of text pages read out.

On-Premise

Operate our software on your own hardware, if necessary in your own data center without internet.

Our prices are based on the requested functionality.

CLI

Use Konfuzio in your software and stay flexible on how to use our functionality.

Our prices are based on the number of developers.

Information Security

Handling of your data

We set the highest standards both when creating the software and when processing your data. Both when creating the software and when processing your data we set the highest standards. Checking for the latest security vulnerabilities of our own processes and software always leads us to find security holes. Thereby, we were able to point out improvements to the following companies:

  • Deutsche Bahn AG (Bahn Navigator App)
  • BASF SE (basf.com, incl. being awarded as “Hero of BASF”)
  • Handelsblatt GmbH (wiwo.de)
  • Axel Springer SE (bild.de)
  • Celonis SE (celonis.de)

Try our beta version

Apply now as a Data Scientist or Software Engineer to test the Konfuzio Python Package.


konfuzio.py
                              

from glob import glob from konfuzio.models import bills from konfuzio.ocr import fast_ocr from konfuzio.classify import deepclf

for filepath in glob.glob (path): doc = fast_ocr (filepath) if deepclf (doc) == ‘Invoice’: model = Bills () doc_info = model.extract (doc) print (doc_info)

We offer to join our beta version for free. If you are interested, please apply by writing us a short e-mail and answer briefly the following questions:

  • For which case do you want to use the Python package?
  • Which operating system and which Python version do you use?
  • At which company do you work?
  • How many developer licenses do you need?

Become part

Come talk to us


IT partner

Do you develop software yourself and are currently looking for a solid and fast solution to classify documents and extract information from unstructured texts? We are are always interested in exciting applications and eager to create synergies.

Employees and students

Can you explain the content in a way that makes it fun to get people excited about your ideas? You want to to work more flexible than a 9to5 working day? You want to work as a student and gain work experience and take responsibility? Please contact us!

Mentor

You know a use case from your day-to-day business that still has hundreds of documents manually processed per day? Together with you, we are happy to evaluate if Konfuzio should include the case of application in the core software.

News

News from partners, competitions and other events


InsurTech Hub Munich at WERK1

Top 10 of 160 applications by other Start-Ups

We are now part of the Isuretech Hub in Munich. The InsurTech Hub Munich is an entrepreneurial platform that attracts, inspires and organizes key players and disruptors across industries and technologies to collaborate on groundbreaking, innovative insurance products and services – revolutionizing the future of Insurance. Find out more on their website, click here to read more.

Register Here, to be informed by e-mail about all further news.

LMU EC Accelerator

Stay tuned

We were admitted to the LMU EC Accelerator. “Empowering Entrepreneurs” is the vision of the LMU Entrepreneurship Center (LMU EC), which has the goal of to train the entrepreneurial managers of the future, to actively promote the foundation of successful companies and to support and promote a culture of entrepreneurial thinking and action

Andrássy Universität Budapest

Presentation: Transformation potential through deep learning

Content: The inner engine of digital business models are algorithms that control business processes based on the information available. However, many companies have information in the form of unstructured data, such as PDF, fax or e-mail. The still manual access to information increases the unit costs and makes innovative products and services for many companies not appear economical. The ability to automatically evaluate unstructured data through deep learning and extract information using algorithms offers competitive advantages and transformation potential. The presentation shows how companies use unstructured data to serve innovative products and services even in mature corporate structures. Read more

Cooperation Payyxtron

Perfection in payment transactions

The Payyxtron GmbH develops “Perfection in payment transactions” in dialogue with their customers. The customer portfolio of Payyxtron GmbH covers a wide variety of industries and ranges from small to large medium-sized companies to corporations and banks.

In our cooperation, we have developed an SAP Add-On that empowers employees in their daily work in payment transactions and cash management. Read more

Sparkassen Hackathon

A 1st and 2nd place at Symbioticon

Based on the Konfuzio algorithm, we developed an application for small and medium-sized companies A company that intelligently identifies and settles receipts for employee expenses.

As part of the hackathon we won, on the one hand, an implementation budget worth a total of € 100,000 and on the other hand the special prize in the category of corporate customers worth € 5,000. Read more

[Disclaimer: Video is in German language]

Initiative for Industrial Innovators

Joins forces for startup excellence in Europe.

The “Initiative for Industrial Innovators” sees itself as a program for the pre-seed financing of innovative projects with a focus on industrial applications.

Among other things, we used the financial support to meet requirements for B2B use cases: in the processing of documents through a dynamic combination of expert rules and abstract NLP Rules .

XPRENEURS

Incubator for high-tech start-ups in early stages

XPRENEURS is a 3-month full-time incubator program based in Munich, which supports start-ups around the world, turning technology-driven business ideas into scalable businesses.

The very personal support of the XPRENEURS team helped us to increase the number of our operation models and to further improve our compliance with GDPR. Watch the video of our final pitch.

ContractGuru

1st place at the Bankathon 2018

After 2015, we again presented a new concept to the Bankathon in 2018. An app, that summarizes key data of contracts with the help of Konfuzio and let users sign the contract with just one click using Verimi. This signature is legally binding even under German Law. In addition the process is also accepted by the German Bundesdruckerei. In reference to the Frankfurt based start-up Finanzguru we named our concept ContractGuru.