How Otkritie Bank’s AI-powered app finances invoices with a scan
Omnitive Extract customer story

With Omnitive Extract’s intelligent document processing solution, Otkritie is allowing customers to submit and process invoices, all with a quick snapshot of the document on their phones.

Large financial institutions are increasingly capitalising on the sizeable small and medium enterprise (SME) market—which represents about 90% of the business population. For this reason, banks see a huge opportunity to diversify their portfolios and offer a wider range of services to meet SME’s financing needs.

Access to finance is one key bottleneck on SME growth. In fact, securing bank loans and completing transactions via invoice financing can take them one to a few weeks to complete. 

A study by G2 showed that it can take up to 25 days and between US$12-30 to process a single invoice. On average, a medium size company processes 1,000 invoices a month, imagine the strain!

The main challenge? Invoice financing requires a great deal of human effort—for both customers and bank administrators. Invoices and supporting documents to finance payments are read and validated manually, which requires a long turnaround time to complete. It translates to significant cost, energy and expertise, that impacts cash flow and can create a poor customer experience.

With Omnitive Extract, invoice financing is 1.5x more effective for Otkritie Bank

We worked with one of Russia’s largest banks, Otkritie, to intelligently automate this complex process. On Otkritie’s client-facing app, users can now directly take a snapshot of their invoices and watch the app automatically pull out the necessary details such as payment type, amount, and the bank account information for payer and recipients. With a click, they can then immediately complete the B2B transaction.

Invoice recognition solution for Otkritie using TAIGER's Omnitive Extract

Invoice recognition solution for Otkritie using TAIGER’s Omnitive Extract.

Once completed, the extracted structured data is stored securely within the bank’s application, so their customers can experience quick and smooth transactions with a peace of mind.

In a benchmark test involving over 20,000 users, the app feature set an unprecedented benchmark with 1.5 times more successful transactions than a similar solution implemented by a leading competitor. Each invoice took as little as 10 seconds to be captured and extracted, helping SME clients complete transactions 99% faster than the standard week-long process.

1.5 times

more successful transactions that a leading competitor in a benchmark test

10 seconds

to capture and extract details from each invoice on the banking app

The app is powered by AI algorithms that can read human language

On the backend, the technology powering the app is Omnitive Extract, an Intelligent Document Processing (IDP) tool. What IDP does is use a host of technologies such as Optical Character Recognition (OCR) and artificial intelligence (AI) to capture data from all kinds of documents such as invoices.

Enterprise Grade IDP Solution TAIGER

Enterprise-grade intelligent document processing workflow.

Key to Omnitive Extract’s success in processing each invoice swiftly is the use of language-led AI such as Natural Language Processing and Knowledge Representation. These technologies help to capture and extract a document’s contents by understanding the semantic meaning of the text within the document.

Think about how invoices can come in a range of formats, and can use different ways to express the same parameters, such as a payment purpose and type. 

That’s variability which exists in unstructured data. It’s a main challenge for training machine learning models, which perform incredibly well but only when data sets are consistent.

However, language-based AI technologies can successfully extract information from free text. For example, Omnitive Extract uses an ontology, aka knowledge graph, which is uncommon in today’s IDP tools. 

An ontology contains a detailed dictionary of common entities used in a specific domain, like SME B2B invoices in Otkritie’s case. This helps the bank’s invoice recognition system identify parameters hidden in free text more accurately, lending to better performance despite the documents possessing high variability.

For the case of Otkritie, Omnitive Extract was also built to read documents in the Russian language. Today, TAIGER’s AI has successfully dealt with various languages such as English, Spanish and Vietnamese.

The secret is ‘hybrid AI’

While individually powerful, AI technologies used in combination gives an additional edge to software performance. This combination is referred to as hybrid AI, which blends the strengths of symbolic AI like NLP, and statistical AI like Machine Learning.

Hybrid AI blends statistical and symbolic AI branches

TAIGER’s ‘hybrid AI’ merges the strength of non-symbolic (statistical) and symbolic AI.

The resulting technology has both human-like logic and the ability to aggregate volumes of data to find meaningful patterns—symbolic and statistical capabilities respectively.

Intelligent document processing is making work easier in the back office too

TAIGER’s extraction solution has also been deployed across other business operations within the bank. It’s presently used to help employees automatically capture data from passports to complete SME onboarding for their 200,000 large client base.

Otkritie web interface for SME passport processing

Web interface for passport recognition for SME onboarding using Omnitive Extract.

With Omnitive Extract integrated into as a web interface, employees are now automatically processing passports at over 92% accuracy, and four times faster than the traditional process.

92%

accuracy rate at passport processing

4 times

faster than the traditional passport recognition process  

Find out more about the full customer story by downloading or case study paper here.

See how Omnitive Extract automated

SME onboarding and financing for Otkritie Bank

Russia was a new geography for TAIGER, but a clear market fit

Though relatively new to the Russian market, the Singapore-based company’s expansion into the geography was a fitting opportunity.

As shared in an exclusive interview with CNA, the Russian finance and banking industry is at an inflection point. According to the EY Global Fintech Adoption Index, the fintech services penetration rate in Russia was 82% as of 2019, largely exceeding the global average of 64% or the rates of Western countries.

Leading news television channel CNA speaks with us about our new expansion into Russia with finch solutions.

Mark, our SVP of Business Development also shared with CNA how rapidly the market is moving towards automation.

With Moscow now part of ESG’s Global Innovation Alliance network, TAIGER is hoping to catch another wave of growth in the region.

“We saw a great market opportunity in Russia. Russia has over 300 banks, a lot of insurance companies. And the majority of them want to introduce more automation in terms of document processing right now, so we want to tap on that.”

With Moscow now part of ESG’s Global Innovation Alliance programme, TAIGER is hoping to catch another wave of growth in the region.

About TAIGER’s Omnitive Extract

Omnitive Extract is TAIGER’s intelligent document processing (IDP) tool designed for structured to highly unstructured documents with high variability. The tool integrates various branches of AI, including Natural Language Processing and Knowledge Representation through ontologies, to effectively read data in the form of natural language.

What Omnitive Extract achieves in greater automation scalability for complex documents enterprises face in the business world.

Speak with one of our solutions managers today on how Omnitive Extract can be applied in your organisation today.

Sorry

Your browser is unable to display our site correctly.
It is best viewed on a modern browser, such as Chrome, Safari or Firefox.