KNect365 is part of the Knowledge and Networking Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 3099067.


IoT in 2017: Who's using data analytics and AI correctly?

Industry experts believe that, if harnessed correctly, IoT and AI can drive businesses to new levels of profitability.  Here are some of the best uses of IoT in 2017.

The internet of things  and artificial intelligence have sparked innovation in the consumer realm, but the enterprise has far more to gain by investing in the same technologies.

Industry experts believe that AI and IoT in 2017 can pave the way for fresh revenue streams, increased productivity, and even the evolution of new business models in the enterprise.

IoT, encompassing everything from smartphones to connected cars, has been embraced by consumers seeking more efficient and smart devices in their daily lives. To further increase the sophistication of these devices, in the past few years, technology vendors have begun applying AI technologies to products and services.

You only need to look at Facebook's facial recognition algorithms, Amazon's Alexa voice assistant smart speaker and Apple's Siri to see how far AI has travelled in the last decade to affect our daily lives.

While the initial wave of IoT and AI technology focused on consumers, the enterprise is yet to take the plunge and integrate it fully into their own systems.

The cloud, software as a service (SaaS) and bulk data collection are firmly entrenched in many enterprise firms. While countless companies collect reams of data which could be analyzed to enhance business practices, detect purchase patterns and streamline operations, Deborah Sherry, General Manager at GE Digital Europe, believes that the enterprise is yet to "create value" through Big Data to its full potential.

The problem is simple: there is too much information at hand. Without the help of AI and machine learning (ML) to take the strain off analyzing these vast sets of unstructured data, according to the executive, the enterprise is missing out on valuable clues and patterns to future business success.

Sherry says these technologies are "critical to overcoming the data deluge and unlocking massive potential business value".

"The ever-increasing amount of digital information presents significant opportunities for enterprises to improve safety, health, efficiency, and profitability," Sherry says. "Most importantly, AI and machine learning can enable organizations to drive innovation, introduce new services to market faster and achieve a long-term competitive advantage."

There are many ways AI, in particular, can be applied to enterprise systems to improve business practices — as well as care and safety.

Artificial intelligence can be utilized in intelligent tools for health monitoring, ML can be used in smart cities to tailor applications and systems depending on citizen usage and needs, and manufacturers can couple AI and sensor scans to find defects in parts and order replacements immediately in factories without human intervention or disruption.

In every scenario, learning from business data analyzed by AI systems is at the heart of benefiting from these technologies.

"The fundamental opportunity that AI presents is the ability to learn from data, to use past results to predict outcomes in the future," Sherry says. "Done robustly, and with feedback loops, AI insights can be built into AI products and get better and better over time."

If you define AI as a means to emulate human decision-making when analyzing information to make quicker — and better — decisions in real-time than human analysts, then this technology can be applied to reduce the time it takes for a business to make decisions over products or distribution, supply chains, human resources, marketing and approaches to sales.

One benefit of using AI and ML-based algorithms is that they improve over time as more information is fed into systems.

As noted by Paul Whitelam, Group VP, Product Marketing at ClickSoftware, "they are entirely data-driven systems that can develop increasingly sophisticated analysis through powerful feedback loops, and are broadly applicable in any of the areas where expert systems used to be deployed".

Not only can AI work for enterprise vendors to process big data effectively, but reliable AI systems can streamline a number of business processes and automate both manual and repetitive tasks, thereby cutting costs, increasing efficiency, and potentially reducing human labour requirements.

However, there are potential pitfalls to dashing headlong into the use of AI. The time and cost to implement AI into existing systems or create new platforms altogether may not deliver the results vendors seek, the technology may not have been "trained" to deal with exceptional circumstances and situations, systems may be manipulated for nefarious reasons with human operators none-the-wiser, and there is a lack of capacity for auditing — and accountability — if you place all data and decisions in the hands of machines.

Despite these potential problems, AI does offer an opportunity for the enterprise to capitalize on data in the present to tailor business models and decisions for the future.

Prediction-as-a-service (PaaS) is a new offering stemming from these advantages which highlight how companies are diving into data collection, analysis, machine learning and AI algorithms to create an entirely new business which caters to enterprise players.

Google's Cloud Prediction API is a cloud-based machine learning toolset that administrators can bolt-on to existing systems in order to analyze data and ascertain directions which would benefit businesses. The API also includes features such as customer feedback analysis and spam detection.

Peak Indicators is a company which specializes in PaaS as a subscription service for business users. The firm offers an Oracle cloud-based platform which outsources predictive analysis — removing the need for enterprises to hire their own data scientists — to make sense of bulk business information, highlight behavioural patterns, and make predictions about future business performance and opportunities.

Ken McHugh, General Manager at Tigerspike, Southeast Asia told IoT World News that PaaS is likely to become a "sought-after" service potentially in the very near future.

While enterprise players may wish to take advantage of AI, they may not have the in-house resources available — and in the same manner that SaaS has become popular worldwide, PaaS may also follow.

"That being said, the tools we use will need to mimic the intuitive leaps that the human brain can make as this is something that learning algorithms have trouble getting to grips with," McHugh noted. "In particular, correlation and causation require a creative thinking mindset to fully determine certain patterns, initially these patterns may seem random to a machine but this might only be based on the information that the machine has available at any given time."

"There needs to be a human element involved here, in order to link additional date and extract specific patterns and correlations," the executive added.

Whether or not PaaS captures the imagination of the enterprise as a whole, according to Forrester research, businesses will invest 300% more in AI this year than in 2016.

This might seem like an unrealistic increase in investment over the course of over 12 months, but as ClickSoftware’s Whitelam notes, the statistics may not represent simply a "conscious" investment in new, AI technologies; but rather, the improvement of existing business processes and software.

"Organizations who use workforce management software don't necessarily consider this an investment in AI, but when one considers that the software is using machine learning and other algorithmic approaches to optimize job schedules, then it becomes clear that more and more dollars, pounds and euros are being spent on this influential technology," Whitelam says.

The implementation of IoT, AI and ML processes not only requires the investment of time and money but also demands a cultural change to take place. Bringing human insight, empathy and the ability to connect the dots together with AI algorithms can result in a change in how vendors view business challenges and make decisions concerning potential future routes to success and profitability.

The enterprise cannot rely completely on machines and automation to run its businesses, as taking the human element out of decision-making would likely backfire eventually. No matter how sophisticated an algorithm, AI is not human, but rather, can only emulate decision-making based on its data, programming, and scenario training.

As a result, the successful introduction of  AI,  ML and IoT in 2017 into the enterprise requires companies to create and maintain a smart ecosystem which allows these technologies and human operators to form symbiotic relationships able to communicate with each other to provide the best return on investment.

As GE Digital's Sherry says, "AI can play a key role of helping organisations uncover critical insights that can help them improve business decision making, but this can only go hand in hand with human knowledge and as part of a wider set of analytics tools".


Get articles like this by email