We are back with another round of AI Innovators, a series in which we speak with AI and Technology leaders to discuss AI, its implications for their business, how they are staying ahead, and what is top of mind as they explore expanding use cases of AI.

AI advancements occur daily. Keeping up with them can be daunting or exciting, depending on who you are and your role. For Joji Philip, Director of Data Science at Ericsson, it’s exciting. He leads an AI incubation team that uses the latest in AI to brainstorm and incorporate it into their business. 

So, staying ahead of the curve is crucial for Joji. Joji is spearheading AI and ML proof of concepts and innovations with his team, demonstrating the application of AI advancements (including autonomous networks – more on that later)  within Ericsson’s core business sectors. 

Join us as we delve into Joji’s insights on AI advancements, personal interests, and the future of AI technology.

Note: All thoughts and opinions shared are Joji’s and do not necessarily represent the views of Ericsson. 

Introducing Joji Philip, Director of AI and ML at Ericsson

Please tell us more about yourself and your background.

Joji Philip: My name is Joji Philip, and I work for Ericsson as the Director of Data Science. My team and I focus on AI and ML proof of concepts and innovations that test ideas that could potentially be implemented within our products. We create prototypes and POCs, demonstrating their potential to product teams across Ericsson’s various sectors, such as the radio access network, core network, transport network, and OSS/BSS business.

The Personal Side of Joji

What do you like to do in your spare time when you’re not implementing AI at work?

Joji Philip: In my spare time, I thoroughly enjoy hiking and spending time outdoors, which is a regular weekend activity for me. I frequently hike in the Diablo Hills area near Danville, particularly enjoying the Las Trampas trails. Reading books and listening to music are my preferred ways to relax and recharge.

The Future of AI in the Telecom Industry

What are some of the most exciting AI advancements that you see coming in the tech industry?

Joji Philip: One of the most compelling advancements is agentic AI, where companies like OpenAI are building LLM-driven automation agents. In the telecom industry, we’ll see these agents used for managing networks, isolating faults, and enhancing customer experiences. AI will aid in understanding network faults and customer churn and evolve into autonomous networks closer to level 4, enabling more autonomous network operations.

Can you explain how autonomous networks are further for those new to the term?

Joji Philip: Autonomous networks, similar to self-driving car technologies, are set to revolutionize the telecom industry by enhancing efficiency and reliability through varying levels of autonomy. Joji outlines five levels of network autonomy, with the industry currently in levels 2 and 3. 

The initial levels 0 and 1  involve manual and assisted operations with limited automation, in which human oversight is paramount. As we progress to level 2, networks begin to operate with more autonomy, although still requiring significant human intervention to ensure reliability and trustworthiness. Level 3 introduces greater independence, where, based on intents, specified agents collaborate to manage networks efficiently but still with human oversight for most activities, paving the way for better network performance and customer experience.

Finally, level 4 envisions a near-complete autonomous network operation, where AI systems independently manage telco networks, ensuring seamless operation and human intervention is needed only for critical activities. Each stage comes with its challenges and demands for human expertise, but the goal remains to achieve highly efficient, autonomous networks in the coming years.

Challenges and Benefits of AI Adoption

Will AI replace physical workers in the telecom industry?

Joji Philip: I don’t foresee a complete replacement of physical workers, especially in telco networks, within the next few years. Human expertise will remain crucial, particularly for supervisory roles, ensuring the reliability and trustworthiness of AI within critical networks. As AI improves, the need for low-level engineering tasks may reduce, but expert oversight will remain essential at higher levels of autonomy.

“Human expertise will remain crucial, especially in supervisory roles, ensuring the reliability and trustworthiness of AI within critical networks.”

— Joji Philip, Director of Data Science @ Ericsson

The Role of Ethical AI

What are your thoughts on ethical AI and the importance of checks and balances?

Joji Philip: Ethical AI is critical to ensure that AI systems are unbiased, secure, and reliable. While customers are eager to leverage AI’s capabilities, there’s a palpable concern about AI systems making biased decisions or generating spurious results. To address these issues, we find it necessary to implement robust guardrails around AI applications to prevent misuse and inaccuracies.

Recently, an AI model hallucinated by providing a nonsensical response to a prompt that mixed telecom network terminology with a food recipe—demonstrating the potential pitfalls without proper constraints. Addressing ethical AI involves avoiding biased results and ensuring that AI systems do not interact with irrelevant or inappropriate data, thereby maintaining focus and reliability in specific domains.

Checks and balances are important in AI systems, particularly in critical industries like telecommunications, where the implications of an AI error could be significant. The deployment of AI in such contexts must be closely monitored to avoid malicious activities, ensure accuracy, and maintain security. This ongoing vigilance is essential to building trustworthiness in AI solutions, enabling them to operate effectively while minimizing risks associated with autonomous decision-making.

“Ethical AI is critical… Implementing guardrails is essential for ensuring AI remains unbiased and secure, preventing actions that could potentially harm networks.”

— Joji Philip, Director of Data Science @ Ericsson

Skills for AI Integration

What skills are essential for engineering teams to stay ahead in AI and technology?

Joji Philip: As AI continues to evolve and influence various industries, engineering teams are tasked with acquiring a comprehensive set of skills to remain competitive and effectively leverage AI technologies. A crucial aspect of this skill set is familiarity with large language models (LLMs) and their inherent features, whether open-source or proprietary. Mastery of these models allows teams to implement AI solutions efficiently, requiring proficiency in function calling, tool integration, MCP, finetuning of small language models, and the development of applications using agentic AI for specific use cases. 

Additionally, selecting the appropriate tools and frameworks, such as Crew AI, Langchain, Langraph, AutoGen, etc., is vital. This demands technical expertise and strategic decision-making to ensure that technological choices align with broader business objectives.

The transition from AI prototypes to enterprise-level deployment necessitates careful consideration of operational factors, including cost efficiency and security. Engineering teams must adeptly conduct ROI analyses to confirm that AI implementations are economically viable and deliver tangible value. Establishing robust guardrails is essential to maintaining trust and security, alongside developing thorough monitoring systems to assess the performance and reliability of AI solutions in enterprise contexts. 

By cultivating these capabilities, engineering teams can effectively implement and integrate AI into business operations, ensuring long-term sustainability, competitive advantage, and a leadership position in innovation.

“Engineering teams must educate themselves on LLMs and agentic frameworks, understanding how to scale AI applications to an enterprise level.”

— Joji Philip, Director of Data Science @ Ericsson

Role of Digital Workplace Solutions

What value do you see in WorkTech, especially with AI Enterprise Search tool?

Joji Philip: I see tremendous value in digital workplace solutions like GoSearch for large organizations. These kinds of tools are indispensable for navigating the vast array of documentation and diverse information sources that companies use daily. Integrating data from various platforms allows employees to quickly locate specific information and receive direct answers to their queries. This not only boosts productivity but also enhances the efficiency of task completion, particularly in complex environments such as Ericsson.

Beyond simply offering search functionality, the true power of tools like GoSearch lies in their capacity to provide meaningful insights and facilitate seamless workflows. By distilling vast amounts of information and delivering context-specific answers, these solutions enable employees to make informed decisions and perform tasks more effectively. 

This functionality is crucial in environments where multiple tools and data sources are used regularly. It highlights the transformative impact that comprehensive enterprise search solutions can have on organizational efficiency and performance.

“AI enterprise search tools like GoSearch are invaluable for delivering quick, precise answers and actionable insights within users’ workflows, significantly enhancing productivity in complex environments like Ericsson.“

— Joji Philip, Director of Data Science @ Ericsson

Conclusion: The Promise of AI Enterprise Search

As we wrapped up our insightful conversation with Joji, the topic of digital workplace solutions took center stage. This is where GoSearch comes into play—our solution offers comprehensive AI enterprise search capabilities that can revolutionize information discovery across multiple platforms, much like the innovations Ericsson seeks. 

By integrating workplace data seamlessly and offering a secure framework, GoSearch stands ready to support companies in enhancing productivity and decision-making through advanced AI solutions.

Stay tuned to AI Innovators for more insightful discussions and breakthroughs in AI technology.