Leaders at Amdocs, which provides software and IT services to communications and media companies, have leaned into using generative AI (GenAI) in both technical and nontechnical teams. To succeed in that effort, the company tasked Victoria Myers, its global head of talent attraction, with transforming how the company prepares its employees and teams for the era of AI.
Collaborating with Amdocs’ Learning and Development team, Myers ensures that people are recruited with AI skills in mind and that employees are equipped to thrive in an AI-powered world. She helped launch an extensive training program for all 30,000 of its employees, specialized to each person’s function. About 76% of the workforce has participated in GenAI training.
Myers has experienced the value of this proactive approach firsthand. “I've taken a bunch of training myself,” she says. “It’s helped me stay at the forefront. I can sit in conversations with technical people and have a basic understanding of what we’re doing.” Equipped with knowledge of GenAI and given a practical introduction to the available tools, she’s now an example of how it can positively impact her work: “I’ve got a copilot open pretty much all the time,” she says. “What used to take me hours can now be done in minutes.”
But when it comes to embracing AI, Amdocs is more of an exception than the rule. A Society for Human Resource Management (SHRM) survey in August 2024 found that 80% of U.S. workers classify their AI knowledge as beginner or intermediate at best.
The reality is that AI is new, complex, and constantly evolving. It is challenging even for experts to keep pace. But it’s also necessary, especially for HR leaders, to understand what critical skills are required for the AI era, how to recruit for them, and how to nurture them.
AI is a vast landscape made up of countless tools and techniques. Nobody needs to be an expert in all of them, and few companies need to hire the full stack of AI personnel, from the researchers building the foundational AI models to the knowledge workers applying them to their businesses. But to understand how to succeed in a fast-changing world, it’s crucial to grasp the foundational concepts and keep current with emerging tools and trends. This approach ensures HR leaders are adept at what’s essential now and prepared for the innovations on the horizon.
Bhawna Bist, a managing director in Deloitte Consulting LLP’s Human Capital Consulting practice, says the firm found that many organizations are reporting a growing need for technical AI staffers: the people who create AI software, algorithms, and systems. Even companies that use AI systems created by AI providers such as OpenAI often require engineers to fully exploit these tools or extend them into their unique business domains.
As these new technological tools are often used by nontechnical people, they also need to hire people with skills in using AI tools. “There is a growing need for AI translators,” Bist says. “They bridge the divide between the business and technical staff at the front and back ends of AI solutions.”
Looking at AI technologies first, these are areas of expertise in high demand:
Machine learning is the backbone of artificial intelligence. It allows AI to identify patterns, learn from data, and improve predictions without requiring explicit programming. Deep learning is a subset of machine learning: “It is the technology behind many of today’s AI innovations,” says Emily B. DeJeu, a professor at Carnegie Mellon University’s Tepper School of Business and author of multiple papers on AI in communication. “It’s like a big sifter, which looks just for anomalies. It's making possible something that previously was impossible. As an example, one of my colleagues used deep learning to build a system that mines huge amounts of data to detect financial fraud.“ DeJeu adds, “But there is still a need for human expertise to make sense of the results.”
For numerous business functions, from analyzing financial data to programming industrial controls, companies may need experts in these areas. They look for people with theoretical backgrounds in these sciences as well as mastery of programming languages such as Python and R, alongside expertise in popular machine learning frameworks such as TensorFlow and PyTorch.
As AI adoption continues to surge, the demand for machine learning professionals is rising. The salary range for machine learning and deep learning engineers is $133,000 to $212,000 across all experience levels. (All salary data in this report is sourced from Glassdoor.)
ChatGPT, Gemini, and Claude can create seemingly humanlike responses to plain-language questions, summarize information, translate content, and even generate programming code. They are all using large language models (LLMs), a technology that was only recently rolled out to power these new tools. LLM science is a subset of natural language processing, one of the original areas of AI theory, with its foundation dating back to Alan Turing in the 1950s.
Working knowledge of LLM theory and coding are the hottest and most in-demand skills in AI, and that’s reflected in the compensation the top LLM coders can expect. Engineers working on these tools at leading AI companies, such as OpenAI, can expect extraordinary compensation, according to Levels.fyi, with base salaries over $250,000 and total compensation (stock included) over $400,000 at the entry level and well over $1 million for the most experienced engineers.
Hiring AI engineers and scientists outside of companies such as OpenAI and Google is not as expensive, with a salary range of $124,000 to $264,000 across related job titles.
It may seem like few companies need workers with these skills, but ServiceNow’s Enterprise AI Maturity Index 2024 shows that 31% of organizations say they’re developing LLMs in-house, and a further 36% are taking a hybrid approach, where prebuilt models are also in the mix. Clearly, specialists with the expertise required to do this work will remain in high demand.
Computer vision allows machines to interpret their environment: to recognize patterns, identify objects, and track movement. Whether for enhancing security systems, enabling autonomous vehicles, or streamlining manufacturing processes, computer vision plays a critical role in automating tasks that were once solely reliant on human judgment. With capabilities to understand the visual world in real time, analyze live footage, detect anomalies, and make real-time decisions, these systems are revolutionizing industries.
Amazon exemplifies the power of visual object detection in its fulfillment centers, where robots autonomously pick, pack, and transport products with incredible precision, optimizing the entire supply chain.
Computer vision engineer salaries range from $126,000 to $228,000.
Beyond purely visual systems, Bist says that “skills for computer vision are becoming skills for multimodal AI, which is about integrating text, video, and audio data for more comprehensive analysis.”
AI systems need high-quality data—for both training and day-to-day use. Without good data, AI products generate inaccurate or biased results. This critical data need has fueled demand for specialized data engineers. They help with tasks such as designing, building, and maintaining data pipelines. They also clean messy data sets, manage databases, and support large-scale systems.
“A lot of data nowadays doesn't fit in a single spreadsheet on a single computer,” explains Noah Giansiracusa, a professor of mathematics at Bentley University and author of How Algorithms Create and Prevent Fake News. “Data engineers will help with the big data coding of the system across vast, distributed data centers.”
Data scientists are seeing significant growth in earning potential. The job’s salary ranges from $128,000 to $207,000. The U.S. Bureau of Labor Statistics projects a 36% growth in job opportunities by 2033.
Giansiracusa notes that the AI-driven surge in data needs has led to growth in other critical roles, including those for cloud computing infrastructure and AI model deployment.
A 2024 Goldman Sachs report projects that cloud computing sales will reach $2 trillion by 2030, with AI driving much of that growth. This expansion highlights a growing need for cloud engineers and architects who design, build, and manage the massive and complex infrastructures that support scalable AI solutions. Cloud computing experts often pursue certifications to validate their expertise, such as AWS Certified Solutions Architect, Google Cloud Professional Cloud Architect, and Microsoft Certified: Azure Solutions Architect Expert. These certifications, some of which require years of study to obtain, cover crucial skills, including designing scalable architecture, managing resources, and implementing security and ethics best practices.
Glassdoor reports that cloud engineers earn between $121,000 and $192,000.
A recent Deloitte survey points out that 68% of generative AI projects are still stuck in pilot or proof-of-concept phases. It turns out that IT teams can’t just drop machine learning and AI models into existing systems that were designed for different types of computing loads. And it can be a puzzle to ensure reliability, reproducibility, and cost-effectiveness of AI systems while wrestling with regulatory concerns that dog many AI projects.
Meet the LLMOps (large language model operations) engineer. This is an emerging role for people who ensure that AI models can integrate smoothly into existing systems, scale across an organization, and be monitored by a business’s IT team.
Companies, especially those in regulated industries such as healthcare, may also look to hire specialists in AI ethics. Candidates for these roles typically have a background in a technical specialty such as computer science, as well as education in law or a liberal arts field such as philosophy.
The EY 2024 Work Reimagined Survey shows that GenAI adoption has gone from 49% of employees in 2023 to 75% in 2024.
For those looking to work in this evolving space, the required skills are well beyond casual chatbot interaction. AI users do not have to understand how the technology works or how to build an AI model from scratch, but some skills can help get the most out of the technology and leverage a company’s investment in AI tools. “Prompt engineering is most important,” says Myers. “I’ve taken four courses on it, and each time, I’ve gained valuable insights.”
Prompt engineers earn from $143,000 to $237,000.
There are many paths to developing AI skills, just as there are many different roles in the field. Higher education is a comprehensive route to skill building and an easy way to find team members, especially now that many universities are adapting their curricula to include practical experience with AI.
But if time and flexibility are priorities, online courses, certifications, and boot camps can be powerful tools for learning core AI skills. Platforms such as Coursera, Udacity, and edX provide focused, practical training.
As Giansiracusa points out, “You can get a lot of education in six weeks.”
There are also many open-source tools that make it easy to build AI models. Tools such as TensorFlow, PyTorch, and Keras break down complex AI tasks into manageable steps; there are online tutorials for all of them. And when people are looking for help, communities such as Hugging Face and Stack Overflow are packed with resources and helpful people ready to lend a hand.
Thanks to the multiple paths available to workers, those who are recruiting and hiring AI specialists need to be flexible as well. While many AI roles require technical knowledge, many do not. ServiceNow's Jayney Howson, SVP of global learning and development, says that knowing how to leverage the capabilities of AI tools does not require a theoretical foundation in how they work. The field is advancing quickly, and many products successfully make the technology approachable. She says, “It is already true that we don’t all need to fully understand [AI]. It depends on the person’s role.”
AI education should be ongoing for anyone whose work touches the technology, as the field is constantly changing. But according to Amdocs’ Myers, “It’s not always about taking a class. It can be watching bite-sized videos, say, on YouTube.”
Maintaining AI skills isn’t about getting drilled on concepts and theory. “It’s about practice, practice, practice,” says Terrence J. Sejnowski, a computational neurobiologist at the Salk Institute and author of ChatGPT and the Future of AI.
Books and classes can give people the scaffolding of theory, but true mastery comes when people dive in and observe firsthand the quirks and limitations of the tools they’re using.
AI's depth and rapid evolution mean that the learning approach needs to be equally dynamic. Staying skilled in AI is about cultivating a mindset as much as skill acquisition. AI product users and engineers can benefit from working on mini-projects or challenges that mimic real-world scenarios. Think of each project as a stepping stone, teaching you something unique and practical. Each experiment and error brings people a step closer to building not just an understanding, but a confident intuition about AI.
And as Giansiracusa aptly observes, students are discovering ways to bridge whatever theory or skill they are learning by using tools like ChatGPT as their very own professor. With this approach, they’re asking questions—about, say, how a particular AI algorithm functions—and getting instant feedback that helps cement complex ideas in a practical, memorable way.
Looking ahead, Amdocs’ Myers points to the increasing role of AI-powered robotics as it moves beyond industrial applications. This capability will lead to fresh demands in the workforce for expertise in using and managing these technologies and the delicate balance of human/machine interaction.
AI is also poised to transform how we reskill the workforce. For example, when using an AI tool such as a copilot or chatbot, we can ask it questions about how to use it effectively. But the system will also learn from users’ actions and be proactive with recommendations and advice. It will then evaluate progress and may even send a message saying a user has earned a badge or certification without taking a traditional course.
“Learning should be embedded into workflow,” says ServiceNow’s Howson. “This is our vision: predictive, active, and personalized learning at the point of need.”
AI is a new field full of opportunities for people with new skills. From mastering core areas such as machine learning and natural language processing to staying informed on emerging trends, today’s in-demand competencies provide a roadmap for future-proofing careers in tech and beyond. The pathway to an AI-adept workforce is not a company of theorists, but fostering in people an adaptable mindset and helping them build practical experience.
Generative AI (GenAI) Machine learning (ML) Deep learning Natural language processing (NLP) Prompt engineering
A type of AI that creates new content such as text, images, and code based on learned patterns from existing data
A branch of AI where systems learn and improve from experience without explicit programming, using data to make predictions
A machine learning subset using multilayered neural networks to analyze complex data patterns, often for tasks such as image and speech recognition
An AI field enabling machines to understand, interpret, and generate human language for applications such as chatbots and translation
Designing specific prompts for AI models to generate relevant and optimized responses
