Artificial intelligence (AI) is seen as core to telcos' own digital transformations, especially for its ability to automate processes, according to a new survey of telecom service providers, all of whom are at different states of transformation.

IT services and consulting firm Infosys Technologies Ltd. (Nasdaq: INFY) queried 932 enterprises and 138 telecom service providers on their use of AI in digital transformation. The telcos indicated they are using AI for machine learning (86%), building AI-based applications to amplify and improve their products and services (80%), automating their decision making (71%), building chatbots (66%), completing cognitive AI-led processes for tasks (65%) and automating predictive analytics (62%).

The goal for 36% of them was to use AI for process automation, followed closely by data analysis at 34% in an effort to increase productivity, minimize costs and increase consistency. (See Yvette Kanouff Shares Cisco's Automation Equation.)

Interestingly, 44% said they want to use AI to "refocus people's efforts on other non-repetitive tasks that benefit from human intervention" suggesting that some -- albeit less than half -- are looking to retain jobs even with AI and automation making many redundant. Perhaps the jobs that may be hit the hardest are in customer support as 71% said they want AI to provide human-like recommendations for automated customer support. (See Will AI Create More Jobs Than It Destroys?)

For more on artificial intelligence and machine learning, visit the dedicated automation content page here on Light Reading.

Infosys, of course, has skin in this game as it offers AI and automation technology and consulting to both its enterprise and telco customers in the midst of digital transformation. However, its findings are backed up by Light Reading's own discussions, as well as from what we've seen from early transformation leaders like AT&T Inc. (NYSE: T), Level 3 Communications Inc. (NYSE: LVLT) and Telefónica . (See Together, We Can Build the Telecom 'App Store' and Beyond MANO: The Long Walk to Network Automation.)

Infosys said the O2 division of Telefonica, for example, has earned 650% to 800% return on investment over three years from its 160 robots, which processed about 400,000 to 500,000 transactions each month. AT&T, for its part, is pairing AI with SDN to parse its data at the edge of its network in real time to ensure it has a rapid response to network problems, and Level 3 has said it combines human intelligence with AI to understand the vast ocean of data it collects on its network for threat detection. (See SDN + AI: A Powerful Combo for Better Networks and AI Paving the Way for 5G, IoT.)

All of these use cases have process automation at the heart of them, naturally because it's the number one priority for telecom service providers, according to Light Reading's own surveys. It's what's driving their use of NFV, SDN and their digital transformations in general. (See Process Automation Tops Carriers' Goals for NFV.)

Telcos are at different stages of their digital transformations, with 51% telling Infosys they have achieved their transformation goals. For the remainder, the biggest challenges that remain include a lack of data-led insights on demand for 71%, lack of collaborations among teams for 61% and a lack of in-house knowledge and skills around the technology for 47%. (See Digital Network Transformation and CenturyLink CTO Updates on Transformation 'Journey'.)

Some still aren't convinced of the value of AI yet either -- a whopping 75% said they lack clarity on the value of AI, and 71% said they lack financial resources for it.

— Sarah Thomas, , Director, Women in Comms