M+E Connections

Data Scientist: Companies Should Start Their AI ‘Journey’ Now

The advantages for companies to use artificial intelligence (AI) to analyze all their data are many, and they should get prepared and start developing an AI strategy now, because it’s going to be a “journey, not a trip,” according to Michal Tadeusiak, senior data scientist at deepsense.ai.

“Data is becoming the most valuable resource of our time,” he said Aug. 30 during the webinar “Investing Intelligently in AI Solutions.”

“Machine learning is a way of extracting knowledge from data” and is not only a “proper tool to deal with” all the data that’s available, but also “the only way to process and infer the valuable information from such vast amounts of data,” he said. Machine learning (ML) also works with all sorts of data, including images, signals and text, he noted, adding that similar techniques are viable across many problems and sectors.

There are several AI solutions that are “already working” and are “already ubiquitous” in areas that include spam filters, data security, financial trading, fraud detection, product recommendations, online search, record linkage, marketing personalization, personal security, healthcare, news clustering, handwriting recognition, natural language processing, machine translation, image recognition, speech recognition systems and smart cars, he said.

Advantages of AI-driven anti-fraud solutions include their ability to adapt continuously, their ability to detect complex problems, and personalization, he said. With product recommendation, meanwhile, AI algorithms can automatically infer customer preferences and product latent attributes, he said.

The “hype” level for AI and machine learning has reached its “peak,” he said, citing Gartner data from July 2016. “ML is not a magical tool that is going to solve all the problems,” but the hype around it is “deserved,” Tadeusiak said. It’s important to keep in mind, however, that ML works very well only when it’s “used properly,” he said.

There is, meanwhile, an “awareness chasm” that exists today between industry experts and data scientists, he said, noting that many people who are experts in fields such as banking and medicine “might not keep up with the recent developments in AI and not be aware of the possibilities it offers.” He added that “only when working together” can the “full potential” of AI be realized.

Companies looking to invest in AI can opt to enter the category in one of three ways: building their own custom solutions, buying a packaged application, or outsourcing to a trusted partner, he went on to say. There are advantages and disadvantages to each of them though.

Building a custom AI solution is a good route if it would be a “critical differentiator” for a business and it would provide the “highest level of agility and control” over the AI, Tadeusiak said. But such a solution could take a while to implement and would only work if the company has data scientists on staff up to the task.

Buying a packaged AI solution is wise when cost and time are crucial and also when a company doesn’t mind having a lower level of control over the AI project, he said. Outsourcing, meanwhile, could be a “necessity” if a company’s data science team lacks the ability to develop an AI solution, but the company needs a high level of customization and integration, he said.

The AI software and services market may be the fastest-growing technology market today and it’s predicted to continue to be so for the next few years, according to deepsense.ai. The market is expected to grow anywhere from $2 billion in 2016 to $60 billion globally in 2025 – a conservative estimate, according to the company. By the end of 2016, more than 80% of the world’s largest enterprise software companies already started to implement AI capabilities into their platforms, according to deepsense.ai.