After two years of experimentation and missteps, AI is crossing into a new phase. Malt asserts in its study “Malt Tech Trends 2026” (1) that we have entered the era of operationalization.
This paradigm shift is first visible in the numbers. AI is now the second most in-demand skill across all categories. It also shows up in 22% of non-technical briefs in the realms of marketing, design, and management consulting.
In 2023, 70% of AI projects were concentrated in data-related roles. In 2026, by contrast, 65% of demand concerns software engineering profiles. The revolution has thus moved from the labs to permeate the entire stack.
From assistants to agents: the most dramatic shift
The report’s most striking takeaway is captured in a single figure. Demand for AI agent skills surged by a factor of 60 in a single year. Related search queries rose by 26. We are moving from an AI that responds to requests to an AI that executes tasks—working autonomously, chaining steps, and interacting with complex environments.
In 2024, the flagship skill was Retrieval Augmented Generation (RAG). It allowed grounding LLMs in an organization’s own data. This component remains solid: demand for it rose by 129%. It has also standardized as a core building block of agent architectures. Yet, in Malt’s quadrant, it was dethroned by AI agents.
On the orchestration frameworks front, LangChain still holds 75% of the market. Yet it slipped by 16 points in a year. At the same time, the ecosystem is fragmenting. LangGraph, CrewAI, and LlamaIndex are gaining ground. No single standard has emerged clearly. This fragmentation signals that the market is still in an architectural exploration phase.
The reshaping of skills
The report notes a clear retreat of once-dominant skills. The share of profiles highlighting JavaScript, Python, or Java fell by 20 to 30% in a year. This does not mean those technologies disappear. It indicates that freelancers are actively repositioning their profiles toward more differentiating expertise.
In 2024, nearly 40% of new tech freelancers listed JavaScript as a skill on registration. That rate dropped to below 20% in 2025. Meanwhile, the share listing AI as a primary skill more than tripled. WordPress remains the most demanded skill in absolute volume. Python, meanwhile, remains the king of languages for GenAI.
Workflow automation tools are growing at an astonishing pace. n8n stands out as the most telling example. Its Malt projects surged by 1,390%. It now rivals Java in overall volume. A telling detail: 71% of its users are senior technical professionals who view it as a way to bypass traditional bottlenecks. Moreover, 82% of demand comes from companies with fewer than 50 employees.
Zapier and Make round out the podium in this category. Use cases span: CRM and marketing automation (35%), AI agent integration (25%), API connectors (15%), and data pipelines (15%). The benefits cited are manifold: time savings (80%), cost reductions (60%), and scalability (50%).
The label “low-code” remains somewhat misleading for tools like n8n. In reality, the value of these platforms does not lie in the absence of technical skills. It lies in reducing the cognitive load on engineers. Less time spent deciphering past decisions, more time for tomorrow’s actions.
Towards a hybrid profile: the orchestrator mindset
This transformation creates a highly valued new profile on the market: the “super-connector.” It blends deep domain expertise with technical mastery of automation. These profiles no longer merely prototype. They deploy autonomous AI agents capable of executing complete operational workflows. Value is no longer found in code-writing alone; it resides in the ability to steer intelligent systems.
According to Malt’s CTO, Claire Lebarz, competitive advantage no longer rests on the size of engineering teams. It isn’t about stack maturity either. It hinges on what she calls “intelligence liquidity”—the ability to mobilize, at the right moment, the optimal combination of AI-driven automation and human expertise.
(1) Malt Tech Trends 2026 — Analysis of the behaviors of more than one million freelances and 90,000 client companies on the Malt platform, covering the EU, the United Kingdom and the Middle East. 2025 data compared to 2024.
