“AI” may be a hot buzzword – and a global market expected to grow to nearly $310 billion by 2026 – but what exactly does artificial intelligence mean? 

The  definition of AI can be a willowy one to pin down because its application is so broad and ranging in degree of complexity, scope, algorithmic underpinnings and methodologies used.

For these reasons, there is an increased call for a more advanced, specific definition of AI beyond “the simulation of human intelligence processes by machines.” Some consider the AI next step – and ultimate evolution – to be cognitive computing.

“It’s such a vast amorphic term that AI has come to mean right now,” said Stephen DeAngelis, founder and CEO of Enterra Solutions. The cognitive computing company has come up with its own name for this evolved AI use case, autonomous decision science (ADS), which it says goes “beyond data science.”

AI needs to not just be autonomous, but to have the ability to “sense, think, act and learn itself,” DeAngelis said. It’s not just operating on its own and generating data and insight – but making decisions based on that information, as a human would.

Cognitive computing offers human-like reasoning

AI is, at its most basic definition, the simulated processing of human intelligence by machines. Machine learning (ML) is its self-learning subset.

Cognitive computing, according to practitioners, goes beyond both by leveraging such techniques as pattern recognition, natural language and “human sense” processing, data mining, and other systems that strive to simulate human thought processing.

According to Markets and Markets, these types of processes working in tandem allow cognitive computing systems to analyze emerging patterns, spot business opportunities and handle critical process-centric issues – all in real time. This can ultimately enhance interaction by “providing relevant, contextual and valuable information” that can inform customized recommendations and decision making, reduce business costs and streamline business processes.

The firm predicts that the global cognitive computing market size will grow to $77.5 billion by 2025, representing a CAGR of more than 30% from 2020. Numerous factors are fueling this growth, including continued evolution in the computing environment (cloud, mobile, analytics), increased hybrid deployment models, and increased human/machine interaction. There is also accelerating demand for intelligent business processes, and companies are more and more applying deep learning techniques and utilizing cognitive abilities to reduce excess operational costs.

The growing number of companies offering cognitive computing tools include SparkCognition, Numenta, Deepmind, CognitiveScale and Enterra. Microsoft Cognitive Services, HPE Haven OnDemand and IBM Watson also maintain a strong presence in the space.

Uniting technologies

Enterra describes its ADS platform as involving human-like reasoning AI software that “serves as a data scientist, subject matter expert and trusted counselor.”

The technology combines capabilities such as inference reasoning, semantic reasoning, symbolic logic capabilities, ontology-based rules engines, industry specific knowledge bases and common-sense knowledge bases. These are then paired with glass box ML techniques and non-linear optimization functions. The latter two together provide more of a transparent, X-ray view into the ML process (as opposed to standard “black box” ML models) while allowing the system to solve optimization problems when constraints or objective functions are nonlinear.

Leveraging all this, the system can analyze disparate data sources at “speed of market” without being constrained by variables in analysis, DeAngelis explained. The system can generate business and consumer insight and understand business processes accurately and with limited human intervention. ADS can then automatically make decisions and learn from those decisions. It also explains in simple language what decisions were made and why, and how to further act on them.

All told, this can help companies to optimize drivers of revenue, enhance supply and demand planning, and gain competitive intelligence, among other benefits, he said.

DeAngelis reiterated the fact that the term AI can be overused and misused, and is often attributed to technologies that are more technically ML. But as he contends, there is “nothing artificial or intelligent” about ML.

ADS and cognitive computing, by contrast, “allow us to gain for the first time enterprise scalability into the analysis of data,” he said. “We can analyze larger swaths of the world, larger swaths of businesses.”

Business use cases

Enterra focuses its ADS technology largely on global consumer products, DeAngelis explained.

One example use case is incentivizing products or providing trade promotions. If the system can understand all the constraints of the manufacturer and the retailer, it can generate trade promotions that are feasible for both players by essentially using “control knobs” that find just the right optimization.

For example, in the frozen pizza category, it can look at weekly promotional campaigns and sales, while also factoring in if a certain manufacturer had a shutdown and couldn’t create pepperoni. It can tinker and “re-optimize at the speed of the market,” DeAngelis said. “The AI can optimize every combination.”

The platform is able to reduce processes that used to take weeks to mere minutes, he claimed, which has been helpful in getting clients through the last few years of COVID, COVID-plus inflation, COVID-plus-inflation-plus another wave of inflation brought on by the war in Ukraine.

“You can never forecast every problem, but if you can get good at responding, you can mitigate risk, take advantage of market changes,” DeAngelis said. “It gives companies a discrete set of control knobs that they can modulate to affect the goal. It is decoding the dimensionality of the human experience.”

Enterra claims that its ADS is 90% as accurate as human experts, and that its apps have derived 1,000% annual ROI. The company’s tools are used by top brands including Nestle, McCormick, Mars, P&G, TPG and Unilever. DeAngelis said Enterra is experiencing roughly 95% CAGR and is on a path to IPO in the next 24 months.He reiterated the fact that the future of AI is not ML, but techniques such as cognitive computing and ADS. “This is the next wave of analytic innovation,” he said.

VentureBeat’s mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Learn more about membership.