Navigating the AI Storm: Three Paths for CSPs in the Age of Generative AI


Navigating the AI Storm: Three Paths for CSPs in the Age of Generative AI

In an AI-dominated world, CSPs face a crucial choice: evolve or become obsolete. JLA has mapped out three implementation paths – Big Bang, Metamorphosis, and Microevolution – each offering a distinct route to harness the potential of Generative AI. Time is of the essence, and the decision is stark: adapt or fade away.

The artificial is (getting) real.

CSPs can no longer sit on the sidelines while disruptive tech companies eat their lunch. Generative AI’s capabilities are rapidly evolving, with use cases ranging from basic customer service to end-to-end network optimization. CSPs must evaluate and select an AI execution approach that aligns best with their internal processes, capabilities, and strategic goals.

And embracing AI needed to happen yesterday.

Industries including banking, high tech, retail, and telecommunications are witnessing a rapidly accelerating embrace of AI, and in particular, Generative AI. Generative AI refers to algorithms like ChatGPT, Bard, GitHub Copilot, and Dall-E that can create new content including audio, code, images, text, simulations, and videos that rival the complexity and quality of content produced by humans. McKinsey estimates that one third of organizations have integrated Generative AI into at least one of their business functions today, with new use cases and applications emerging faster than we can write this essay. The potential for Generative AI to alter critical industries is compelling executives to ask the question, “What should I do about it?” and more critically, “Who can help me?”

The disruptive qualities of Artificial Intelligence are a hot-button issue today, but the stage was set decades ago, at a time when most organizations could not forecast the impact that advanced algorithms would have today. Machine learning within AI started getting traction in the 2000s, focusing on teaching machines to learn from real world data and drive accelerated outcomes. The 2010s witnessed a shift to deep learning’s multi-layered neural networks that simulate the decision-making power of the human, using computer vision to classify and detect objects and voice recognition to power speech assistants.

2020 and beyond is characterized by the rise of Generative AI, marked by the advent of Generative Adversarial Networks (GANs) and the enablement of language-based AI applications with far reaching consequences for knowledge, interactions, and processes. Today, Communications Service Providers (CSPs) must increasingly leverage Generative AI to enhance business and network operations, and in doing so, lower the risk of being disrupted themselves.

Organizations unwilling to adapt or evolve their operations will get left in the dust.

When CSPs have a challenge, AI has a solution. These range from operating use cases like customer engagement to true neural-based network and resource optimization. The potential of AI includes optimizing network performance, implementing predictive maintenance, enhancing customer service, personalizing marketing, and augmenting network security, all while greatly reducing costs. The integration of AI is no longer a competitive edge – it’s a necessity for survival and prosperity in an increasingly digital world.

The current state of the telecom industry isn’t exactly pretty.

Declining ARPU, reduced ROIC, and heightened vulnerability due to margin pressures are creating an urgent imperative for CSPs to innovate and transform. For each of these issues, AI offers a mitigation strategy. By implementing AI in sales, CSPs can combat ARPU decline through targeted marketing and tailored plans that better meet customer needs. By leveraging proactive AI services and self-optimizing networks, CSPs can automate routine tasks, improve resource allocation, and boost ROIC. By introducing AI-powered internal operations and front-line productivity enhancements, CSPs can become more cost-efficient, better handling margin pressures and CAPEX challenges.

But how should a CSP navigate these uncharted waters?

JLA has outlined three potential implementation approaches, each of which is best suited for a CSP based on their underlying infrastructures, capability, adaptability, and objectives. The Big Bang Approach, Metamorphosis Approach, and Microevolution Approach are monikers offering differing frameworks with distinct benefits and drawbacks, ranging from incremental and cautious changes to wholesale transformation. CSPs must evaluate their history, technologies, and overall strategic goals to determine the optimal approach to effectively integrate Generative AI.

The Big Bang Approach

The Big Bang Approach is a radical transformation for organizations looking to reincarnate as an AI Native CSP. CSPs are placing a bigger bet on the inevitable transformative impact of AI and will need to embrace an operating model that leverages the strength of both humans and machines, with cloud native virtualization emerging as the first step in the journey. Once a cloud foundation is in place, leaders can integrate AI into various operational aspects that align with overall digital transformation goals.

The overall driver for a Big Bang approach is a foundational shift in mindset and culture, requiring investment in talent, technology, processes, and governance. Significant shifts in operational strategy, extensive focus on training and upskilling employees, and enhanced awareness of AI capabilities are all critical initiatives. CSPs will need to invest in data quality and quantity build their own foundational models and employ sophisticated data labeling and tagging strategies. Selective acquisitions will help acquire fresh talent and IP, building capabilities in data architecture, vector databases, and data processing pipelines. From a technology perspective, employing this approach will require high-speed storage and processing, greater flexibility for cloud and on-premises deployment, and complex integration with existing systems.

The Big Bang is characterized by higher risk and reward. An overhaul strategy that is not complemented with the appropriate tools, talent, and strategy may expose CSPs to new risks including inaccuracies, cyber threats, and IP infringement. This approach is best suited for companies that are newer, have the DNA to undertake massive transformation, and have already built cloud-native infrastructures, resulting in more effective integration with AI-enhanced functions and processes.

For the Big Bang Approach, organizations must first consider whether the shoe fits. Although we haven’t seen a market-proven example yet, younger, more agile players may find themselves gravitating toward this foundational shift. For example, Rakuten Mobile’s emergence as a cloud-native player in Japan’s telecommunications scene positions it as a potential future case study. Central to Rakuten’s strategy was the initial and crucial decision to build a fully virtualized, cloud-native network. Organizations like Rakuten, who already position themselves as disruptors, may begin accelerating AI-enabled shifts and consider the Big Bang Approach.

The Metamorphosis Approach 

The next approach, Metamorphosis, is more measured and best suited for established operators that require a structured and stepwise migration to the cloud and AI. In this approach, each phase builds upon the success of the previous one, allowing an organization to evolve its capabilities and gradually transform business processes through risk-managed progression. Executives can exercise greater control over their AI integration process, validating the need for greater capital investment as they see the benefits. This staged implementation offers controlled scaling, and manageable risks, but the financial uplift may be more conservative than the Big Bang Approach.

The Metamorphosis Approach requires a larger team of multidisciplinary specialists with greater emphasis on compliance, code validation, risk management, and performance and scalability. Organizations willing to invest in substantial strategy coordination will benefit from learning from front-runners’ experiences and evaluating market trends and competition on the path to transformation. Reduced market risks due to delayed entry and a more strategic go-to-market approach may be more practical for legacy CSPs.

Organizations undergoing this managed path to AI integration must first perform a root cause analysis, determining where AI can have the greatest impact, performing detailed gap analysis, and identifying strategic projects to pursue. Next, CSPs should create an AI roadmap that prioritizes high-impact projects and creates a sandbox environment to train, test, and deploy AI-enabled applications. CSPs can leverage industry models as templates, but they must be fine-tuned to meet each organization’s strategic goals. These pilot programs will deliver lessons learned and codify best practices to drive an incremental shift in mindset and culture.

Established CSPs are shifting towards a Metamorphosis Approach today. However, each CSP is carving a different path to achieve the same end goal. Comcast, for example, is instituting an AI Accelerator Program, investing in technologies that create text, video, audio, and code-based content. AT&T is looking inwards, creating a Generative AI tool for employees, a multi-lingual conversational platform, and supporting initiatives in digital avatars and vehicle routing optimizations. SK Telecom (SKT) is focusing externally and has developed an AI chatbot to compete against ChatGPT that leverages in-house large language models and integrated services including music streaming, e-commerce, and payment apps. By leveraging the power of the cloud and advanced analytics, established CSPs can overcome traditional barriers to entry, drive competition, and position themselves as early adopters – ultimately propelling the entire telecom industry forward.

The Microevolution Approach

Finally, the Microevolution Approach is best suited for smaller players or cautious adopters who choose to focus on gradual AI adoption with manageable investments. The lower-risk, learn-as-you-go approach is more suitable for organizations looking to find an AI footing without the time and capital intensity of the other two approaches. Albeit lower risk and investment, the long-term economic benefits may not match those of bolder strategies. This approach dips a toe into the AI waters, placing a greater focus on aligning smaller-scale AI projects with specific business needs. Benefits lead investments through measured and targeted implementation around specific use-cases.

CSPs pursuing this approach will likely leverage partner AI capabilities to drive improvements in select functions, ingesting pre-existing services through APIs and other interfaces. The initial focus will likely be internal – validating outputs, integrating insights into workflows, and incorporating tools that clearly drive productivity, reduce cost, and improve customer experience. The key to the Microevolution Approach is establishing a process of continuous evaluation with high emphasis on partner selection.

CSPs opting for this approach must carefully choose the right solution and vendor for implementation. With numerous providers touting Generative AI capabilities, distinguishing hype from reality is crucial. Moreover, attracting AI-specific talent poses a challenge, highlighting the need for a strategic partner who guides them through every step of the process – from vendor selection to piloting, implementation, and integration.

Examples of incremental changes that CSPs adopting the Microevolution Approach can implement today include integrating conversational AI in call centers or deploying chatbots for customer service. Customer care functions are low-hanging fruit, providing an immediate opportunity to recognize the benefits of AI on a smaller scale. Organizations looking to institute an easier customer service approach will not need to build capabilities or skills in-house but can instead evaluate and pick a market-tested partner. Orange, for example, is partnering with Google Cloud to test AI in call centers. These next generation contact centers uses Generative AI to transcribe calls, summarize the exchange, and suggest follow up actions to the agent based on the discussion. The result? Dramatic improvements in both the efficiency and quality of customer interactions.

Regardless of the approach a CSP selects, it starts from the top. AI is fundamentally a C-suite conversation. Today’s market is dominated by proof of concepts and showcases, but real-world engagements will define the true impact and financial return that Generative AI can deliver. Reprioritization requires investment and effort from stakeholders who see the writing on the wall and are ready to act. CSPs that are slow to move risk being ousted by competition that provides faster, cheaper, and better services.

A CSPs approach to AI should mimic AI itself— constantly enhancing, upgrading, and evolving.

AI, as expected, is iterating beyond even Generative AI. Causal AI, albeit nascent today, provides an even greater opportunity, allowing AI models to make accurate predictions based on causes rather than correlations. AI will soon be able to make choices the way humans do. For networks, this translates to intelligent routing for apps and services based on compute and processing needs. Retrieval Augmented Generation (RAG) is yet another advancement, which can optimize the output of a large language model by referencing an authoritative knowledge base outside of its training data sources.

Generative AI is only the first chapter, with tools and processes evolving at breakneck speeds. CSPs embarking on this journey must determine the optimal of three approaches and begin instituting critical use cases. Each approach offers varying levels of effort and upside, providing a unique way to balance executive alignment, talent needs, infrastructure requirements, and potential outcomes. Carriers without an AI strategy risk becoming obsolete in an increasingly competitive and technologically innovative environment. Telecommunications is ripe for disruption, and executives must ensure the disruption comes from within the industry as opposed to outside of it. Generative AI unlocks a toolkit of enhanced processes and outcomes that will ensure CSPs stay relevant, and more importantly, exist.


John Trobough

Founding Partner, JLA Advisors

Sangit Rawlley

Sr. Advisor, JLA Advisors

Zain Sharif

Sr. Consultant, JLA Advisors

JLA Advisors is a boutique firm offering our clients a comprehensive end-to-end set of services, from strategy development, technology architecture design and execution, to software operational excellence, with an emphasis on innovation.

Recent Posts