Is AI really useful? An ROI point of view
19 Jul 2020 Yogesh J
AI-based intelligent data extraction

Technology is constantly evolving. Every innovation is the first step for further improvement. The world of business assimilates these changes and adapts itself, evolving itself constantly. The creation of the world’s first programmable computer is attributed to Konrad Zuse, when he built the Turing-complete Z3 in May 1941. The real boost was delivered by ENIAC a machine built-in 1943. The ENIAC a computing system built by John Mauchly and J. Presper Eckert is said to have executed more calculations in the ten years of its operations than all of humanity had until that time. With the unveiling of the first personal computer, Altair in 1974, the birth of Information technology took place. Rest as they say ‘Is history’.

Today, Information technology is again standing at the start of a new stage of evolution. Some of the emerging technologies are:

  1. IoT: Internet of Things which connects billions of devices through the internet, sensors, and other mediums.
  2. Artificial Intelligence: AI as it is popularly called, is intelligence displayed by machines/ computers, that tries to mimic the natural intelligence of humans and animals.
  3. 5G: The fifth generation of the cellular networks in terms of speed, developed with a focus to support IoT and rapidly growing data needs.
  4. Serverless computing: An innovation which will be based on peer to peer computing.
  5. Blockchain: A cryptocurrency exchange, not very popular as of now, but things might change.
  6. BioMetrics: The latest method of verifying an identity.
  7. 3D printing: A technology which will change the face of manufacturing permanently.

In all the above technologies, IoT and AI have gained immense importance lately.  Artificial Intelligence is being adopted in practically all industries. Barclays is using AI to detect and reduce fraud. Mining sector in Australia is implementing AI to increase productivity by an estimated 23%. AI has near term business and financial opportunities.

Is implementing AI that simple?

The answer is ‘No’. There are many challenges to implement artificial intelligence.

Artificial Intelligence is going to replace mundane tasks. In a study conducted by McKinsey, ~50% of current work activity can be automated. As automation will increase productivity and improve growth, a lot of professionals will have to re-skill themselves to remain relevant in a hyper-competitive job market. With the Covid-19 pandemic, the situation just went from bad to worse. 

Access to data is also a constraint. AI feeds on data if data is not accessible AI is a useless piece of technical wizardry.

And last but surely not the least- cost! Implementing AI is not cheap. It requires infrastructure, investment not only in the form of currency but efforts. An AI system also requires some ‘Training’ before it can start delivering actionable results.

To tide over these challenges requires time and investments.

So, how does one get ROI from AI?

An obvious question any business professional would ask.  But there is no straight answer to this. Before investing in AI one should ask:

  • What are the business goals?
    • AI tools need to be trained, re-trained and fine-tuned to serve a specific goal- and that is an investment.
  • Will the problems be resolved?
    • The question to ask here is what problem needs to be solved and will AI help in it.
  • Availability of skilled professionals?
    • Once an AI tool is implemented, it requires professionals with specialised skills to keep it running and for finetuning it.
  • What are the KPI’s?
    • Any tool requires to be measured against performance indicators. These indicators need to created and implemented.

If the answers to all of these are firm and definitive – then maybe it’s a good idea to stand at the starting point of AI implementation. Once implemented, the returns can be:

  1. Growth & Expansion
  2. Improved Productivity
  3. Greater Savings
  4. Higher utilization of human intelligence

There should be a simpler workaround?

Of course, there is. A business might not want to automate all its functions at once. But go about it bit by bit.

Let’s take an example of Accounts payable function of an organization. Accounts payable function consists of executing a set of tasks aimed at processing data almost daily. A perfect candidate for automation. The most error-prone activity of these set of functions is the entry of supplier Invoice details.

This activity can be automated with KlearStack. KlearStack is an AI-based intelligent data extraction/ intelligent document processing platform. It reads machine-readable documents e.g. Invoices and extracts actionable data from it. Invoices are mostly unstructured in nature i.e. there is no set format. Therefore, care needs to be taken while entering data. The implementation cycle is 7-10 weeks. KlearStack assures 90% accuracy in data extraction within 90 days – #90in90.  This has proven to improve productivity by 200%

Data extraction process automated within 3 months is the real ROI for automation- and there is a tool which is already delivering it. So to end it all, ROI in AI is a logical expectation and it can be delivered provided there is well-planned thought and strategy supporting it.

Conclusion :

It’s high time that organizations across all domains and irrespective of the organization size should adopt such AI based tools to reduce the wastage of resources and leakages in the cash-flow. 

Recent Post

How is the Accuracy Rate of an OCR Scanner Measured?
How Can I Improve My OCR Accuracy Rate?
What is Optical Character Recognition and How Does AI Make it Better?

Meta