During my conversation with friends, acquaintances and candidates applying for a job, I have come to a realisation that people have difficulty in expressing their understanding of what Inkredo is even after reading the blogs. Some people say that we are a data transformation company, some say that Inkredo is a data analytics company, or a data visualisation app while there are few who still have a perception that we’re a lending company.
At the least they have heard us and they are on the right track. But their responses make me cringe! Recently, I have realised that the reason people make such guesses about their understanding of Inkredo is because of the complexity created by use cases of Inkredo within the financial services industry.
Through this blog, I’ll try to simplify what Inkredo does without oversimplifying our work.
Today, even the most efficient lenders (both corporate and retail), payment gateways and wealth managers make data-driven decisions through a long, tedious and human-dependent processes involving spreadsheets. They must regularly (as frequent as daily) gather a holistic picture of all their customers’ transactions but the process of collecting income information, account balances, fraudulent transactions and history of assets from all sources and files create a considerable work for those at the helm of it.
There is too much data hibernating in those files but there is too less information extracted because there are multiple custodians of data and there is lack of shared electronic data feed that makes such a task prone to manual error.
A TYPICAL SOLUTION
A typical solution involves manual transcription of PDF statements into a spreadsheet or some sort of accounting software. The scale of this manual transcription varies from a handful of people working in a wealth manager’s office to a large KPO employing hundreds of people.
The manual transcription from PDF costs money, is slow and error prone. How much and how efficiently can a human transcribe in 8-hours of work daily? The cost is substantial and it could be anywhere between USD 25–50 per statement and it could cost anywhere from couple of hours to a number of days. The cost depends on whether the process is done in-house or is it outsourced to a KPO. The best case outcome is that it works correctly and requires manual checks on both the client and vendor sides. How many abled-heads and their hours does one need to arrive at a consensus to fully approve the captured data?
THE INKREDO SOLUTION
Step 1: Accumulate Data
The first step is data accumulation from all sources. This usually comes from internal sources from the client side which is usually a data repository or a direct upload.
Step 2: Transform Data
The second step is data transformation. Although the data comes in electronic form, it is not machine readable. e.g. data stored in electronic PDFs or scanned statements cannot be read by machines.
Chances of getting machine-readable data directly from financial institutions varies from easy-to-tough-to-impossible depending on where your bank is located (easy in the US and impossible in Kenya) and whether you are talking retail or private banking (easier in retail and almost impossible in private banking). I will share the reasons for this inconsistency across geographies and banking types in a separate post.
Step 3: Analysing Data
Since we are dealing with bank transactions from structured files there is no need for cleaning before analysis. We create an executive summary supported by graphs and tables within few seconds to help decision makers make informed data-driven decisions.
BEHIND THE SCENES
Data Intelligence is a better way to make an important decision with data. Inkredo starts with collating all the data extracted from different sources. We then transform this throughput of data into intelligent insights that make the best decision obvious.
Inkredo handles every step of using data to make intelligent decision. From accessing to analysing and visualising data, Inkredo brings the entire decision-making process into one place for quicker, easier, more efficient data-driven decisions. Inkredo is optimising to take on the toughest of files (e.g. meta-data stored in investment banking statements) and ugliest of data systems such as statements from co-operative banks.
We have taught our system to extract high-quality machine readable data from PDF formats of different banks. What we have found is that reading text from PDFs is easy but it takes some effort to convert into machine readable and maintain the tabular structure of the data. Inkredo automatically detects the layout of the statement and key elements on the page, understands the data relationships in embedded tables and extracts everything with its context intact. This means you can instantly use the extracted data or store it in a database without any complicated code in between. What took hours, if not days, to our customers, now we deliver more than their expectation in few seconds.
This automated workflow processes data contained in the statement to a visualised information which is further used for initiating credit checks to approve the loan so that borrowers can get instant results of their application rather than having to wait several days for manual review and validation.
Take a free demo to know what we are upto.
This post was originally published on Medium by Kumar Tanmay.