SME stands for Small and medium enterprises, also interchangeably used with MSME, Micro, Small and medium enterprises. Before getting into the definition, let us look at some facts and figures about it :
a) The total number of SMEs in India including both registered & unregistered is estimated to be at 42.50 million, which is an astonishing 95% of the total industrial units in the country.
b) Next to the agricultural sector, this sector employs about 106 million people making it about 40% of India’s workforce.
c) SME sector produces more than 6000 products.
d) 12 million people are expected to join the workforce in the coming 3 years with the sector growing at a rate of 8% per year.
The above numbers clearly indicates the importance of SME sector and justifies it being known as India’s ‘engine of growth’. Banking solutions specially catered to SMEs is known as SME banking. However, there are numerous challenges that these enterprises face, the biggest being financing and funding. In India, hardly 10% of small businesses have access to formal credit.
On a positive note, this challenge opens up new opportunities. In order to curb the financing issues, non traditional or alternative lenders and technologies like AI/ML have entered the SME banking/lending space, disrupting it.
Where can AI/ML play a role in SME banking?
The cost involved in lending to a SME is high, making it one of the biggest challenges. The operational cost that goes into processing small loans to SME is such that it leaves very little margin for loan decisions that are bad. Also, the amount of time and manpower that goes into processing these small loans is nearly the same as that of bigger loans. This is where automation and artificial intelligence (AI) comes to picture.
* Robots can be developed to perform tasks related to customers to back office. These robots can perform routine and repetitive tasks, collating data across multiple accounts, and carrying out credit checks. Thus, making them both more effective and efficient.
* Customer onboarding is a lot of manual work which ultimately leads to a high turnaround time. But AI has been proven beneficial in this case where it automatically fill the fields with the help of external sources of data. It has enhanced these processes from cost, quality and customer experience standpoint. This inventive automation drastically reduces bank’s operational costs that goes in processing of an SME loan.
* Traditional banks are generally reluctant to give loans to these micro and small enterprises because of the anticipated risk of such businesses. Apart from banks, there is an informal lending available through local money lenders but the yearly interest rates charged by them can be as high as 80%-100%.
Challenges like this have triggered the need of alternate lenders, who have brought technology-first approaches, infrastructure and analytical capabilities.
Data has become an inseparable piece for every sector and industry these days. There are numerous lenders who uses data and machine learning to assess the capability and intent of the customers. Different startups use different types of data in order to gain insights and predict.
* ePayLater is a Mumbai based startup that gathers real-time data using microservices, NeoGrowth makes use of digital payments data, while LendingKart extracts around 8500 data points to perform machine learning and deep learning.
* These startups have large varieties of data sources at their disposal such as bank statements, PAN/VAT/TIN address, tax returns, Justdial data, social media data (LinkedIn, Facebook, twitter, etc), defaulters list, etc which can be used to create a credit assessment system.
* ML and DL algorithms also help in reducing the timelines for credit approval. “The customer can start using the credit line within hours of approval. The approval rate is also higher than traditional banks due to leveraging of alternate data,” Akshat Saxena, cofounder, ePayLater.
I believe that most of the startups will agree with me when I say that to widen the access to credit to a larger population(MSMEs), the new-age have to work on an array of use cases for AI/ML which are today majorly used for backend operations, underwriting, and fraud detection.
But with numerous such companies, standing out will be critical. Today, the company having a straight-through digital process along with analytical and predicitve capabilities to create strategies that are specific to segments and capabilities to assess the risk will be the winner.
I can now imagine the day when banks are using robots and AI to a cent percent automated lending decision in case of SMEs under a given amount and the power of data to perform predictive analytics and give the applicants a standard score. Banks such as SBI and Bank of Baroda in India have already started deploying AI to increase efficiency, reduce operational costs, and identify human behaviour.
Soon, digitizing processes will no longer be an option for banks to stay competitive in the future. A bank refusing to accept the use of AI/ML in making faster and better lending decisions may not be able to survive.