803,000
ads citing AI in that family (January 2022 – December 2025)
714,000
ads citing AI in that family (January 2022 – December 2025)
514,000
ads citing AI in that family (January 2022 – December 2025)
416,000
ads citing AI in that family (January 2022 – December 2025)
389,000
ads citing AI in that family (January 2022 – December 2025)
02
AI Skills in Digital Careers
Software
& web development
TASK MIX SHIFT
Code generation, documentation, and unit testing - once the backbone of entry-level developer tasks - are now largely automated. Developers increasingly focus on creative problem-solving, debugging, and system integration. Industry benchmarks suggest that about 12–15% of development time has shifted from manual code production to solution architecture, code review, and AI prompt engineering. The workflow is now heavily collaborative between human coders and AI copilots.
Top AI tools used
Entry-level developers routinely use GitHub Copilot, Gemini Code Assist, and Amazon CodeWhisperer to generate code, test scripts, and inline documentation. LLM-powered debugging in Replit Ghostwriter or VS Code Copilot Chat shortens development cycles. Generative design tools such as Midjourney for UI elements and Gemini for code translation are also becoming mainstream.
Share of job ads citing AI skills
Around 41% of software and web developer postings now reference AI capabilities, from “experience using AI code assistants” to “LLM integration.” This marks a near-tripling in two years, mirroring the broader automation of coding workflows.
Mateo
Amsterdam
Mateo joins his hybrid engineering team’s morning stand-up in Amsterdam, where his fintech startup manages real-time payment APIs across Europe. Overnight, Copilot has generated unit tests and auto-documented the latest commits in the main repository.
Mateo reviews the output, refining logic and ensuring that AI-generated functions comply with local data protection standards and internal style guidelines.
Later, he collaborates with Gemini Code Assist to prototype a new customer onboarding feature, using plain-language prompts rather than writing every function manually. Debugging, prompt iteration, and peer review fill most of his afternoon.
The AI automates syntax, regression testing, and documentation, freeing Mateo to focus on security, usability, and performance. For him, coding is no longer just about writing instructions, it’s about guiding intelligent systems to deliver compliant, reliable software at speed.
Data analysis
& business intelligence
TASK MIX SHIFT
Much of the repetitive data preparation and visualization work once done by entry-level analysts is now handled by AI. Cleaning, deduplication, and normalization are automated through intelligent ETL (extract-transform-load) pipelines. Visualization and narrative generation are often produced automatically, allowing analysts to devote more time to interpreting results and translating insights for business decisions. Estimates suggest that 9-11% of analysts’ time has shifted from “data wrangling” to “storytelling” and validation work.
Top AI tools used
BigQuery ML, Looker Studio, Tableau Pulse, and Power BI Copilot are now widely used to automate analysis and generate natural-language summaries. Gemini and ChatGPT Enterprise assist in building predictive models and generating contextual explanations for non-technical audiences. Automated code assistants in SQL and Python (e.g., Colab Gemini, GitHub Copilot) also help junior analysts write and debug queries faster.
Share of job ads citing AI skills
Approximately 47% of listings for data and BI analysts now mention AI or ML literacy, often specifying experience with LLM-integrated dashboards or model validation. This represents a doubling since 2023.
Leila
Milan
Leila works in the analytics team of a major retail group headquartered in Milan. Each morning, she opens Looker Studio dashboards automatically refreshed overnight by AI-powered ETL scripts.
The system has already cleaned, merged, and normalized data from e-commerce, in-store sales, and logistics platforms, and even generated natural-language insights about changing consumer trends across Italy’s northern regions.
Her day is less about spreadsheet maintenance and more about interpretation. She verifies that predictive models aren’t overfitting seasonal anomalies and uses Gemini to generate executive summaries for the marketing and supply chain departments.
When presenting to leadership, she prompts AI tools to visualize “what if” scenarios in real time. Her combination of business insight and AI literacy turns what used to be routine reporting into a strategic decision engine for the company’s regional operations.
Cybersecurity
& network operations
TASK MIX SHIFT
AI-driven threat detection and behavioral analytics now automate much of the continuous monitoring that junior analysts once handled manually. Routine log parsing, malware identification, and anomaly flagging are increasingly performed by machine learning models, reducing false positives and allowing teams to respond faster. As a result, around 10-12% of analysts’ time is being reallocated toward deeper incident forensics, red-team simulation, and policy hardening. Entry-level cybersecurity staff now focus on interpreting AI alerts, validating outputs, and investigating edge cases that automated systems can’t classify with certainty.
Top AI tools used
Security teams rely on platforms such as Google Chronicle, Microsoft Sentinel, and Palo Alto Cortex XSIAM, all of which integrate large-scale AI threat analytics. Generative AI copilots also support report drafting, breach summaries, and alert explanations. Automated playbooks built with SOAR (Security Orchestration, Automation, and Response) frameworks use AI to prioritize and route incidents dynamically.
Share of job ads citing AI skills
Roughly 42% of cybersecurity and network operations job postings now mention AI familiarity, typically framed as “AI-driven detection,” “ML-based security analytics,” or “automated incident triage.” That’s up from about 14% in 2023.
Jonas
Estonia
Jonas starts his morning in the operations center of a leading Estonian telecom provider in Tallinn. Overnight, AI-driven threat detection systems have processed millions of network logs, clustering anomalies and isolating suspicious patterns through Chronicle’s behavioral models.
His dashboard highlights only high-confidence risks that require human review - phishing attempts, lateral movement indicators, or zero-day exploits that automated tools couldn’t fully resolve.
Instead of manually combing through log files, Jonas focuses on validating the AI’s reasoning, investigating flagged events, and drafting concise threat summaries for internal teams. He uses generative copilots to produce executive-ready incident briefs or simulate possible attack vectors.
His role hinges on understanding both cybersecurity frameworks and AI logic - ensuring automation remains transparent and accountable in a national context where digital infrastructure security is a matter of civic pride.
UX/UI & digital
product design
TASK MIX SHIFT
Generative design tools now automate the early concept and layout phase, producing multiple wireframes and visual assets in seconds. Designers increasingly operate as curators and testers, validating user flows and brand coherence rather than generating every asset from scratch. This shift frees roughly 8-10% of designers’ time for user research, accessibility refinement, and experimentation.
Top AI tools used
Design teams employ Figma AI, Adobe Firefly, and Uizard for automatic layout, copy, and asset generation. Gemini for Workspace assists in drafting design briefs and summarizing usability feedback. Automated A/B testing and sentiment analysis tools (like Optimizely AI and Hotjar Heatmap AI) help prioritize redesign decisions based on predictive insights.
Share of job ads citing AI skills
Roughly 41% of design-related job postings now mention AI proficiency - often framed around prompt design, AI asset generation, or A/B testing automation. This has nearly tripled since 2023, reflecting how generative design tools have become industry-standard.
Amira
Barcelona
Amira works on the product design team of a fast-growing mobility startup in Barcelona. Her mornings begin with reviewing three wireframe variations generated overnight by Figma AI, based on user analytics from the previous sprint.
Rather than sketching from scratch, she curates and refines layouts, adjusting accessibility, hierarchy, and interaction flow. Using Adobe Firefly, she experiments with visual styles adapted for different customer segments, while Gemini drafts multilingual copy in Spanish and Catalan.
Later, Amira runs AI-assisted A/B tests and interprets sentiment data from user heatmaps and in-app feedback. Her role is about direction rather than execution, orchestrating generative design tools to ensure the product feels intuitive, inclusive, and on-brand.
In a city defined by creativity and experimentation, Amira embodies the designer of the AI era: leveraging automation to amplify, not replace, human empathy and design sensibility.
Cloud
& DevOps engineering
TASK MIX SHIFT
Routine monitoring, capacity scaling, and deployment orchestration have become largely automated through AI-powered observability systems. Junior engineers now spend more time on troubleshooting, governance, and optimization rather than on manual script writing or pipeline management. About 10% of operational time has shifted from maintenance to proactive improvement and reliability testing.
Top AI tools used
Cloud environments increasingly integrate Google Cloud AIOps, AWS DevOps Guru, and Azure Copilot for Cloud to predict incidents and optimize resource use. AI copilots in CI/CD platforms (like GitLab Duo or Jenkins AI) flag security risks, suggest rollout schedules, and reduce downtime.
Share of job ads citing AI skills
Approximately 37% of cloud and DevOps listings now require exposure to AI-driven observability or automation frameworks, up from 10% in 2023, underscoring how AI has become central to modern infrastructure management.
Ravi
Helsinki
Ravi begins his day in Helsinki, overseeing infrastructure reliability for a Scandinavian SaaS company serving universities and research institutes. Overnight, Google Cloud AIOps analyzed terabytes of system telemetry, highlighting performance anomalies and suggesting resource reallocations.
His first task is to assess these AI-generated recommendations — deciding which to accept and which to override ahead of the next product release.
Throughout the day, Ravi uses GitLab Duo pipelines integrated with predictive scaling tools to deploy updates and simulate traffic surges. Routine log analysis, patch sequencing, and performance testing are all automated, allowing him to focus on governance, cost optimization, and long-term resilience.
In this increasingly autonomous environment, Ravi’s expertise lies in managing balance — ensuring that the AI enhances reliability and sustainability without removing human accountability from the cloud.