Salesforce’s recent announcement about plans to introduce a major new artificial intelligence product called "Einstein," which will integrate AI across the company’s software suite, serves as a reminder that AI is coming soon and fast.

In June, a number of tech industry leaders predicted that AI and machine learning will soon "change life as we know it."

Meanwhile, Deloitte anticipates use of AI in the enterprise will see at least 50% adoption in the next five to 10 years.

Part of the reason AI is opening up so quickly has to do with wider adoption of cloud technology. One of the biggest resource needs to have a good AI platform is compute power, and cloud enables that. According to Rajeev Ronanki, principal, Deloitte Consulting and leader in Deloitte’s Cognitive Analytics practice, the rise of cloud is making it easier for enterprises to scale and adopt AI.

"The cloud solves two key issues for AI – on-demand availability of low cost computing resources and nominal cost of storage, which makes it cheaper to process massive volumes of data," said Ronanki. "As a result, most AI applications are resident on the cloud. In turn, because most AI applications are on the cloud, it's driving up adoption of cloud computing for non-AI applications."

"There are millions if not billions of activity and data feeds that can be captured in an enterprise depending on its size," said Darshan Appayanna, CIO of Happiest Minds Technologies. "To be able to correlate and make meaningful sense of all this data requires heavy compute power."

A typical enterprise won’t commonly make such a large in-house investment for two reasons. First, the amount of usable data from these millions/billions of feeds is relatively small. Secondly, the refresh cycle for such infrastructure is a big burden on the enterprise. Hence, cloud is a perfect solution.

"Cloud makes it cost effective and efficient as this compute power can be provisioned on an as-needed basis and you pay for just what you use," said Appayanna.

Where is AI today?

While some companies have embraced AI and machine learning at scale, most enterprises are still in the early stages of adoption.

Ronanki said Deloitte is currently employing AI in two ways. First, to develop and implement AI solutions for their clients. In this case, Deloitte’s uses its Cognitive Cloud platform to automate processes, engage consumers through personal digital assistants and by using machine learning algorithms to find insights that can work to drive both efficiency and growth.

Internally, Ronanki said Deloitte uses AI to automate the services they provide their clients. For example, automation of labor-intensive audit processes, identification of tax optimization/compliance opportunities, automation of data collection processes, optimization of travel schedules and the identification of trends.

Goran Garevski, vice president of Engineering, System Software and Tools at Comtrade Software, said he currently sees AI finding its footing mainly in backend support systems from vendors.

"This is where AI is trying to predict potential customer problems by using various pattern matching, machine learning and analytics technologies," said Garevski. 

Garevski said he also sees large entities (for example, telcos) using AI concepts to proactively manage huge network infrastructures.

"But in order for these new AI-based approaches to be massively used, a few major obstacles must be overcome, such as productization — in order to make these solutions accessible to the users — and the usability of these highly complex concept, to eliminate the need for an AI-expert team to extract the value of the solution," Garevski said.

Happiest Minds is leveraging AI internally through its use of Microsoft Office Delve.

"This is a platform that creates 'serendipity' situations for users," said Appayanna. "The Delve platform is able to look at various users’ activities in our Office 365 ecosystem like mails, chats, files being worked on, calendar schedules, etc. and helps users to 'unclutter' their work environment by giving information on what they need to focus on."

For example, said Appayanna, Delve will show him what his team is working on related to a topic or email he sends to his teams.

"So it’s reading what’s top of mind and trying to get related/relevant information for the same," he said.

For now, smart bots!

Some of the most broadly used AI technologies so far are smart bots that can react to human instruction and take the next steps. The most common example is meeting scheduling.

"There are many startups that are using a combination of human and artificial intelligence to parse emails and understand simple and complex rules for putting a meeting on a calendar," said Nikhil Hasija, CEO of Azuqua.

To do so, the bot has to understand a person's meeting preferences (e.g. morning vs. afternoon, are weekends ok?), the priority of incoming meetings compared to existing ones (i.e. what can be scheduled over and under what circumstances), meeting mechanisms (i.e. is it in person? If so how much travel time to book? If not, who calls whom? Does the other person have skype?), among a variety of other variables.

Hasija said Azuqua currently uses smart bots to route leads to sales representatives as they come in from marketing. Based on a set of rules around geography, company name, company size, prospect status, engagement, job title, etc., an incoming lead can be set to route to the appropriate sales rep. The sales rep then has to confirm acceptance of that lead (indicating that he/she is indeed available for immediate follow-up) before the bot then provides the lead information to the sales rep. If the sales rep isn't available, the bot then will try the next rep in the list of appropriate reps until it finds one who is available.

"You can have an intern do this work, but a machine will be much faster and exact," said Hasija.

While leveraging AI for customer service bots, messenger bots, social feed bots, etc. is interesting, such applications only scratch the surface of what AI can do.

"These are all still very nascent steps towards using the full power of AI in enterprises," said Appayanna.