AI in the service provider space today is primarily focused on machine learning — a branch of AI that focuses on the development of intelligent computer programs that can predict future events more effectively than humans. These types of programs autonomously train themselves using historical data and can self-improve when exposed to new data.
As service providers begin to map out their AI strategies, there are four key requirements every project should meet:
- Have a large volume of historical data with clear success/fail criteria
- Contain a well-defined information model where the data is easily understood, and content can be parsed
- Have a clear area of value to the provider, such as resolving low levels of process automation, high levels of order fallout, poor customer experience, unpredictable performance of network or processes
- Be efficient in developing a proof of concept, so value can be shown quickly
AI starts at the top
AI will impact many parts of the service provider organization, and thus an AI strategy needs to be understood and embraced by company leaders. It’s not enough to just hire machine learning engineers or data scientists and call that a strategy. But it doesn’t stop there — the employees that execute the AI plans must also understand the company’s goals in deploying AI, along with its benefits.
This is where training comes into play as a best practice. Customized AI training across the organization will not only help leaders make better business decisions but also help day-to-day employees understand what AI technology is being used for and what it will help them accomplish. It’s not uncommon for there to be disconnects among the various groups involved, especially around timing; the business may need results in a few weeks, while fine-tuning an AI model may take a few months. Training will help resolve disconnects between what the business needs — and in what timeframe — versus what can be produced by machine learning engineers given the constraints at the time.
As AI is brought into a service provider’s environment, the question inevitably arises whether the organization should centralize their AI efforts into an “AI division” or let each group leverage their own strengths and resources. While one oversight group may look like an easy solution, most large service provider organizations are not ready to have a single AI group to oversee all internal and external organizational needs from the start. Individual needs and rationales for deploying AI may vary between groups, causing potential confusion and delays. The size of the organization itself can be prohibitive when starting with a single common AI group. However, that does not mean individual groups should run in different directions; developing a company-wide strategy is still the critical first step.
One U.S.-based telecom company that is implementing AI throughout the organization has structured its AI organization this way: three divisions, focused on operations and customer care; global supply chain strategy; and big data and AI systems that create new data products. The divisions themselves are separate, while a common group addresses data governance and management, data warehousing, data lakes, and common analytical and AI technologies across the organization. This helps to facilitate cross-functional, cross-organizational projects. For smaller service providers with fewer divisions and smaller data sets, a centralized AI function may be a viable option. Taking the time to invest in strategic AI planning will be of immense benefit in determining how to best structure AI across the organization.
AI in the back office
One area where service providers see real impact from AI is in Business Support Systems (BSSs). These systems assist with order taking and payment collection, and also include such functions as product configuration, product management, order management, resource management, network rollout management and customer management. All can benefit from AI enablement. Most service providers are implementing new omnichannel and self-service capabilities to improve the customer experience. As they undertake digital transformation efforts, the way they manage, sell and support their core services is becoming increasingly automated, and that means additional support is required from their BSS platforms.
Today, CSPs are also facing many challenges like the consolidation of systems due to mergers and acquisitions, prolonged system deployment, lack of visibility to system performance, inefficient business processes and inability to quantify impacts. A big data and AI-driven approach can certainly help to give new insights, automate and recommend intelligent, innovative and disruptive solutions promptly. Service providers should reach out to their partners in the BSS space to explore possibilities to make their BSS systems AI-enabled, if possible.
AI has tremendous potential across the entire service provider organization. However, planning and execution are critical. Rather than focusing on the idea of AI as a whole, service providers need to focus more on individual use cases, the user experience and getting the right data. A sound strategy derived from best practices can mean the difference between success and failure.
Priyank Mohan is the Head of Artificial Intelligence and Advanced Analytics at Excelacom. A senior strategist, thought leader and program/product manager, Mohan has worked in artificial intelligence, big data, machine learning, cognitive technologies, digital transformation, marketing technologies, CRM, data science, search, EDW, and BI. He was also the Head of Product Management and Head of Strategic Solutions for Silicon Valley product companies in the artificial intelligence, cognitive, and machine learning spaces.
Prior to joining Excelacom, Mohan has held senior technology management positions at, and consulted with, large Wall Street firms, telecom, pharma, and hi-tech companies. He has assisted organizations from Fortune 500’s to startups in all areas including strategy, innovation, product management and strategy, program management, digital transformation, developing roadmaps, managing complex programs, growth acceleration, business strategy, ideation sessions, business planning, market research, product marketing, and business development.