“They are going to want cognitive systems that learn, reason and interact across cloud-based applications,” she says.

IBM has partnered with Cisco to inject real-time analytics into everyday tasks such as email and conferencing. For instance, using IBM’s cognitive learning system Watson, its email application could read the body of a message, analyze it to understand that a colleague wants to have a meeting with a certain group and then offer to set up a video conference in Cisco Spark with the members of that group — without the user having to switch between windows or applications.

“The technology offers the ability to read, summarize and prioritize content, and create an action for the user,” Suh says.

She sees infinite real-world potential for the intelligence cultivated between UCaaS and business-critical SaaS. For instance, a car service could use data about preferred customers’ locations, typical passenger load requests and time-of-day order trends to automatically position appropriate vehicles nearby for faster service.

Integrating analytics SaaS to SaaS does have its challenges though.

“SaaS has to be abstracted and written in a resilient way,” Suh says. At some point, a network, a firewall or a router will go down, “so the application has to be resilient across multiple active data centers and the quality to the user, seamless.”

She recommends businesses consider the advantages of applications built with open APIs so they can add and subtract to their pools of SaaS integrations as needed. “We know the client might choose one provider right now, but might want to choose another later. Therefore, the interface has to be designed in a thoughtful manner,” she says.

Greg Collins, founder and principal analyst at Exact Ventures in Burlingame, Calif., agrees that open APIs will be critical so that businesses can choose their own combinations of services. “A lot of times, UC applications can be walled gardens. But as vendors realize that customers want them opened up, they will [open them].”

Businesses, he says, should walk through the feature sets important to workflow to make sure they map properly and that integration doesn’t cause issues. “Have beta testers in the organization try the combination in different environments. You need to make sure that the features retain your desired quality of service,” he says.

UCaaS-to-SaaS integration … as a service

Because SaaS integrations could prove more complicated in larger organizations, Lazar says he expects midsize-to-large companies to support hybrid models in the near term. Complex voice services would remain on-premises or on a virtual server, with ancillary services such as messaging and conferencing overlaid in the cloud.

Before such organizations make the full leap to the cloud, they will need to know who is responsible if something fails.

“If I kick off a call from Salesforce and it doesn’t go through, who is responsible?” Lazar says. He believes third parties will prove critical to wrapping managed services around these UCaaS-to-SaaS integrations.

“Organizations won’t buy integrated UCaaS/CRM directly from vendors like Cisco and Salesforce. They’ll instead get the integration as a fully managed service,” he says.

That option might also interest early adopters like Cartelligent. Carstens says she has run into some issues with having two major applications so dependent on each other and the cloud.

“When a service goes down, like during a power outage or routine maintenance, it can be challenging. If it’s Salesforce, [sales agents] can connect with clients over Gmail, and if it’s 8×8, they can use their cellphones,” she says. “That’s just part of working with cloud-based services.”