FutureNet World was delighted to spend some time recently with Ibrahim Gedeon, CTO at Telus discussing the drivers, priorities and challenges around AI and Automation.
Tell us a little but about you and your company?
I am the Chief Technology Officer for TELUS, a leading national telecommunications company in Canada, where I’m responsible for all technology development and strategy, security, service and network architecture, service delivery and operational support systems, as well as service and network convergence, and network infrastructure strategies and evolution.
Are you deploying AI and automation on your network? If so, where and how?
TELUS has already deployed AI applications through several chatbots that are customer-facing. We have chatbots aimed at answering customer support questions, a TV chatbot, and a virtual assistant that has been integrated with IVR. We also have chatbots that are internal-facing for our team members.
Beyond chatbot applications, we are leveraging AI to detect security threats (e.g. cyber-attacks) on our IPTV network, as well as to predict network failures and outages, as opposed to reacting and isolating afterwards. For example, when the light at the top of a cell tower is about to burn out, we could proactively replace it. This is especially important when the cell tower is located near an airport and maintenance must be scheduled with the airport authority.
Looking to the future, the anonymised datasets generated through telecommunications data can be integrated into advancements within our world-leading networks to proactively identify customer issues. For instance, AI can be developed to enable a flatter network architecture to enhance user and traffic “awareness” while helping drive low energy consumption in our upcoming 5G networks. TELUS’ work on 5G will also be valuable to support the actualization of the vision of smart cities with autonomous vehicles. While we’re not the actual AV provider, our 5G network can provide the foundational connectivity to allow ecosystems of vehicles not from the same manufacturer to communicate with each other.
What are the main drivers for deploying AI and automation in your organisation?
In cybersecurity, we are applying machine learning to automate threat identification and remediation. We’re also harnessing AI and the idea of the “augmented consumer experience” to evolve our future friendly home (e.g. ambient intelligence and elevating the NEST experience). Smart Home intelligent systems can be enhanced through ambient intelligence and real, or near real-time threat detection.
Other applications of AI that TELUS is exploring include helping us optimise energy consumption within our data centres, manoeuver drones to monitor cell towers in support of driving high network reliability, and enable predictive network maintenance. For instance, we are looking at neural networks/machine learning to predict failures in the network – we’re already good at analyzing root causes and there’s a plethora of algorithms monitoring our alarms currently that are partly AI, so now we are looking at ML/NN to predict failures on a huge data set. This will help us reduce cell tower downtime and increase service availability.
Healthcare will be a crucial application for AI. In our lifetime, each of us will contribute to generating approximately one million gigabytes of health-related data. Currently, this staggering amount of medical information is housed across disparate systems, making it challenging for doctors to fully understand a patient’s medical history. As the largest healthcare IT provider in Canada, there is a tremendous opportunity to create cognitive computing solutions that ingest and analyse these enormous volumes of data and present it in meaningful ways to healthcare professionals. AI services and APIs will enable us to deliver better health outcomes using multiple sources of data and advanced analytics. One way is by augmenting existing developments in medical images and diagnostics including improving disease and cancer detection from images and biomarkers; support faster and more accurate clinical decision-making; and enhance smart monitoring and alerts in ER, surgery and patient care scenarios.
Overall, we are lowering the operational cost of our services, while simultaneously improving the customer experience, by maximizing service availability, system responsiveness, and perceived customer quality through the use of AI.
What is the biggest barrier for introducing automation and intelligence into network operations?
The biggest barriers for introducing automation and intelligence into network operations are that large anonymized datasets are needed to train a neural network in detecting anomalies and recognizing patterns. Additionally, the lack of determinism of AI systems prevents us from deploying it in mission critical environments without maintaining a ‘human in the loop’. All of this also requires rebuilding existing processes and retraining people as well for the next stage of digital transformation.
Where do you see AI and automation having the biggest impact?
The biggest impact for us is in improving customer service and enhancing the efficiency of our network operations for all the reasons mentioned above.
Are you developing your own solutions or working with partners? Which do you think it the best approach?
In addition to the work our teams are doing to advance AI applications to enhance customer service, cybersecurity and healthcare, TELUS is at the vanguard of exciting developments as Canada works to advance our country’s artificial intelligence research agenda. For instance, we have committed to support the recently-established Vector Institute in Toronto as a gold sponsor, and over the next 10 years, will invest $2.5 million to support the Vector Institute’s efforts to establish Canada as a centre of excellence for artificial intelligence.
We also have close connections with amii, the Alberta Machine Intelligence Institute at the University of Alberta. Amii performs advanced R&D in machine learning and AI.
Aside from academic and research partnerships, as with all domains at TELUS, we will continue to investigate both R&D and commercial partnerships to support the delivery of our AI vision.
Ibrahim joins us at FutureNet World 2019 on the Opening CXO Keynote PANEL: Will the promise of automation and intelligence really deliver the network requirements of tomorrow?
Interested in attending? Register today.