It's easy and fast to generate a lot of content and tons of ideas. But in the end, you're still the one trying to find or establish the red thread," a manager told us during our research.
This captures the shift precisely. Rather than acting as direct creators of IT products, IT professionals are moving towards a curatorial position that requires evaluating, selecting, and maintaining a unifying storyline within AI-generated output.
Our research identified three key shifts in how IT professionals will work:
Shift from operational to strategic focus: IT professionals will move from performing both strategic and operational activities to primarily strategizing. In the AI orchestra metaphor, they become the "audience" rather than the performers, focusing on defining conceptual visions and guardrails, instructing and supervising the "conductor" (an AI agent orchestrating other AI agents), and evaluating outputs with minimal operational intervention. Traditional tasks such as writing code will be replaced by AI agents acting autonomously after receiving initial inputs.
New role as evaluators and curators of AI output: Rather than directly creating IT products, IT professionals will evaluate, select, and maintain coherence within AI-generated content. As one manager stated: "It's easy and fast to generate a lot of content and tons of ideas. But in the end, you're still the one trying to find or establish the red thread." This means curating from multiple AI-generated options, maintaining a unifying strategy across outputs, and understanding underlying goals as AI agents handle operational tasks. Human prompting will become obsolete as AI agents interact with each other, making the evaluation role even more critical.
Acquiring new skills for building and handling AI agents: IT professionals must develop competencies to coevolve alongside AI agents. This includes working directly with AI systems, bridging human language and machine code, and solving domain problems using well-defined AI agents. They will need to understand how to apply agents to specific use cases without necessarily knowing their internal workings, developing proficiency with "out of the box" AI agent fleets that require higher-level abstractions rather than deep technical knowledge. As one manager noted, this represents "a new job that has sort of just emerged."