Although SMEs account for nearly 30% of India’s GDP, employing about 460 million people, they continue to struggle in making their presence felt.
:- By Nitish Pandey
Let me start with an outlandish prognosis “Ultimately machine learning and AI will change everything around us. From financing to manufacturing to retail, everything will have a hyper-personalized focus.” Obvious, right? So the question is how and when will it touch the Indian Small & Medium Enterprises (SMEs)?
Although SMEs account for nearly 30% of India’s GDP, employing about 460 million people, they continue to struggle in making their presence felt. With technological innovations causing disruption across sectors, there exists a huge opportunity for SMEs to fix this by leveraging the benefits that AI has to offer.
Advent of augmented interfaces backed by dynamic segmentation of users and predictive analysis, we have a potent ally in improved methods of user engagement too – be it the end user of a SME product or the SME workforce. Another significant example is the confluence of Artificial Intelligence (AI) and fintech solutions for the manufacturing SMEs can help them punch way higher than they do currently.
However, bridging this gap between the SMEs and the promise of AI is a challenge that needs an ecosystem of stakeholders. It is an opportunity of mammoth scale. So, what is the opportunity size? Lets extrapolate from digitalisation spends. (Since digitalisation is the backbone of any machine learning solution). According to estimates of a 2017 report published by Google-BCG, digital spend in India is set to reach $100 billion by 2020, tripling from $33 billion in 2016-17.
While large corporate houses have consistently assigned resources to digitisation and digitalisation, the SMEs have fallen shy of deploying enough budgets in this direction. This has led to sub-par improvements be it in manufacturing efficiency & throughput, transaction efficiencies, workflow automations or procurement planning. In short, they have been under invested in such data capturing automations.
Farther, Faster and Frugal
AI enabled strategies are part of the 4th industrial revolution or ‘Industry 4.0’ that have been seeing rapid implementations in several of the developed economies. Who has not read or heard about the robots, driverless cars, drones and IIoT enabled autonomous information systems?
A certain flavours of AI enabled solutions, have created new benchmarks in speedy prototyping, manufacturing optimisation, logistics and distribution of numerous products. Not only do AI-enabled machines drive greater productivity, the robotics and IIoT kind of deployments ensure minimal human intervention, resulting in agility, faster decision forks, and reduced risk at workplace. These implementations are helping many industries worldwide which include warehousing, manufacturing, logistics, budgeting, marketing etc.
The decision making in the manufacturing context has been captured in the following 3Ps by several commentators:
Predictive: What machinery or component is expected to have a down time? What business impact it can have?
Prescriptive: When should one plan the downtime or procurement of inventory? No longer should one rely on the age old concept of a fixed ‘economic ordering quantity’ (EoQ).
Preventive: Make a recommendation to prevent a certain breakdown by say change of some ball bearing or oil change. This can save hours of defect and solution analysis.
AI applications are not limited to be used directly for manufacturing in manufacturing enterprises but also benefit in non-manufacturing facets. For example AI systems can micro-segment customers and therefore recommend optimal utilisation of marketing efforts, discounts etc.
Even systems’ security as small entities are vulnerable to hacking (loss and theft are equally scary propositions) as they lack cyber security protection. In such cases, AI-enabled security applications are better equipped to deal with hackers by detecting ‘abnormal’ online behaviour faster than other systems.
Are they cheap? It is an emphatic yes. The medium and longer term RoI is a big positive. This message needs to be understood by the SMEs rather quickly unless if they want to be caught on the wrong side of frugality. It goes without saying that short-term requires budgetary outlay and therefore not all flavours of AI technologies need to be chased right away.
It is imperative to become aware, analyse and ascertain a road map for AI-enabled technologies that can be useful for their businesses in an incremental manner. Whatever be the items on the road map the first slew of investments will need to target enablement of data.
Collecting, Securing, Decoding Data
It is essential to acknowledge the centrality of data in commissioning AI. The goal of ‘monetisable insights for all’ should motivate the SMEs to start investing in data. Only given enough data, AI could provide the far-reaching solutions in via implementations of robotics, augmented reality, fintech, customer profiling and the autonomous IIoT (Industrial Internet of Things). But they will need a supportive ecosystem with multiple stakeholders pitching in with respective contributions.
This democratisation of insights can begin to show results in less than a lustrum. The evidence lies in the fact that for several decades now, business analytics has enabled businesses to zero in on probabilistic and deterministic solutions, helping them in better decision-making. This was done by visual analyses and at times, by statistical modelling.
This ‘tech enabled’ albeit manual analytical approach made the lack of volume, density and quality of data to somewhat of a forgiving omission. However, with AI we need an order of magnitude, more data to deliver on the investment and deliver outcomes with much greater precision, dynamism and lesser uncertainty.
How ready are the SMEs? The number is in low digits. Those who have not invested in data emitting and data capturing have a clean slate to start with. They have to partner with IIOT implementers or floor automation implementers.
For those who have been ahead of the curve and have collected data around processes and transactions are better off. These SMEs have to begin investing in data engineering with prioritized goals determining which data to clean. The first step any analytically mature organisation is getting their data tamed.
There are significant upfront costs that are incurred for building processes that capture rich data and store it. The storage costs (hardware, network, security, staff, AMC) have fallen sharply and the potential upside is way too high to be vacillating on data generation & capture.
The value for getting our data in order and processing it can be gauged from these figures – Analytics India Industry Study 2017, the big data industry is estimated to be $2.03 billion annually in revenues and will grow at a healthy rate of 23.8% CAGR. Further, the Big Data industry is expected to almost double by 2020. But once companies make these early investments, the dividends will keep accruing.
With data in place (say in 3 years for those who are starting late in the journey) , where reams of data will require days, weeks or months for humans to analyse and decode, AI solutions will do it in a matter of seconds. So whether it is reams of paperwork, mounds of files or voluminous minutes of meetings, archived emails, memos and messages, AI will run through them within seconds, gleaning the requisite data and keywords.
In this way, companies could garner greater insights into regular processes, employee behaviour, government modalities, etc. and secure competitive advantage.
Similarly, companies adopting AI for fintech solutions can, tame cash-flow issues, while reducing financing costs. With AI in financing space, banks and alternative lending entities can use AI to swiftly sift through the data of potential SME borrowers. With robust results, this can facilitate more well-informed decisions without the traditional, time-consuming & expensive due diligence and risk-assessment protocols. Thereby, AI-enabled technologies plug the issues of delay, mis-pricing and mis-allocation of funds to the SMEs.
Needless to say, SMEs patronising such technologies will hold an advantage over those using conventional means. With data, AI technologies can only get better with time; therefore companies should be well advised that one cannot put off the adoption waiting for the technology/algorithms/price points to mature.
What you may save in delaying the ‘capital deployment’ will surely be lost in the data opportunity. The opportune time to start collecting data is ‘Now’.
It has been bandied about that ‘human intervention is reduced to the minimum’. This may be an alarmist statement in that some may conjure an age of depleting employment. Haven’t we been through this scare on a number of occasions earlier? With the advent of PC the same was feared. We believe that the human intervention will be reduced in repetitive tasks for sure and for some decision making processes.
What automation did for some repetitive tasks, AI will do for repetitive strategic decision making. However, unlike pure repetitive tasks, decision making is unlikely to be 100% software driven for there always are environment considerations that are not going to be part of the AI systems input data points.
Hence, what will happen is the humans will move towards building, maintaining and then consuming the output of the AI systems. New jobs will also organically emerge in and around the harnessing of the AI solutions (say IIOT installations & replacements). There will be a need for the workforce to rapidly re-skill. In general there will be an inevitable yet valuable shift towards higher skills by the workforce.
While the workforce is a stakeholder but it is not the active stakeholder. The trigger to make this happen earlier than later (incremental investment with positive IRR) will have to be a joint effort by the government, progressive SMEs, AI service providers and subject matter experts (consultants).
On a positive note, AI’s pervasiveness is poised to transform multiple verticals, including banking, insurance, automotive, education, healthcare, manufacturing and others. Indian SMEs have to start their journey of education (consulting), laying out a roadmap with concrete budgets and executing of an AI adoption strategy.
On an ominous note, analogous to the digital divide that marked the 20th century, the coming decades will be creating a divide of the AI enabled and not AI enabled. Industry associations and their member SMEs have to make a choice which side of the divide they want to be on.
The author is Senior VP – Product and Technology, Power2SME Pvt Ltd. The views are author’s own.