Until a few years ago, building an automated workflow required a developer. You needed someone to write the code, test it, deploy it, and maintain it. This created a systemic bottleneck: the people who knew business processes best weren't the same ones who had the tools to automate them. Low-code started closing this gap. AI is widening it further, in a direction that changes not just how workflows are built but also who builds them.
80% of low-code users already come from outside IT
Gartner predicted as early as 2022 that by 2026, developers outside formal IT departments would represent at least 80% of the user base for low-code development tools, up from 60% in 2021. This shift is driven by two parallel dynamics: the rise of business technologists, profiles with process expertise and enough technical skill to use low-code tools, and the growth of hyperautomation and composable business initiatives that require delivery speed incompatible with traditional IT cycles.
Gartner describes low-code adoption as the organizational response to the need for speed in application delivery and automation of highly customized workflows. Organizations that succeed in equipping both IT developers and business profiles with adequate low-code tools reach the level of digital competence and delivery speed that today's competitive environment demands.
By 2028, 90% of engineers will use AI assistants to write code
On the professional developer side, Gartner predicts that by 2028, 90% of software engineers will use AI assistants to write code, up from less than 14% in early 2024. The developer role is shifting from implementation to orchestration: the focus moves from writing code to problem-solving, system design, and ensuring AI tools produce high-quality output.
By 2027, at least 55% of software engineering teams will be actively building LLM-based functionality. This means that building AI agents and workflows is no longer a research or experimentation activity: it's operational engineering, with quality, testing, and governance requirements approaching those of traditional software.
AI-native software engineering: a new discipline
Gartner identifies AI-native software engineering as one of the strategic trends for 2025 and beyond. It's a set of practices and principles optimized for using AI-based tools in software application development and delivery, incorporating AI into every phase of the lifecycle: from design to deployment. The distinction from traditional use of AI tools for development is that here AI isn't an optional assistant: it's a structural part of the process.
In practice, this means future workflows are built with tools that generate code, suggest integration patterns, identify anomalies in test data, and propose optimizations continuously. The developer doesn't write the workflow from scratch: they direct it, correct it, and validate it. Gartner indicates that for success, teams must balance automation with human oversight, taking into account business criticality, risk, and workflow complexity.
Why low-code isn't being replaced by AI
A recurring question in the market is whether GenAI, with its ability to generate code from natural language instructions, will make low-code tools obsolete. Gartner has explicitly analyzed this question and the answer is no. The reasons are structural: low-code tools don't just generate code. They govern the application lifecycle, manage integrations with enterprise systems, ensure compliance with IT security criteria, and enable long-term maintenance by non-technical profiles. They're platforms, not code generators. AI enhances what low-code tools already do; it doesn't replace them.
The most effective combination is one where low-code tools incorporate AI capabilities: automatic workflow logic suggestions, guided integration generation, assisted testing, and intelligent monitoring of processes in production. This convergence is already underway in the market and is redefining user expectations for what an automation platform should do.
The $58 billion disruption in productivity tools
Gartner has included among its strategic predictions for 2026 and beyond a disruptive estimate: by 2027, GenAI and AI agent usage will create the first real challenge to mainstream productivity tools in the last 35 years, triggering a $58 billion market shift. New vendors will emerge, new formats will take hold, and value will shift toward agentic experiences. The future of work won't be typed but delegated to an agent.
For those building enterprise workflows today, this means that architectural choices made now have a shorter time horizon than expected. Workflows built on platforms that don't evolve toward agentic capabilities risk becoming technically obsolete within a few years. Evaluating a low-code or automation platform can no longer stop at current features: it must include the roadmap toward AI agent integration and the ability to orchestrate multi-agent systems.
Who decides and who builds: the new balance
The combined effect of mature low-code and generative AI is an unprecedented democratization of workflow development. Business profiles can build meaningful automations without writing code. Developers can build complex systems at speeds that were previously impossible. IT teams can focus on governance, security, and architecture instead of implementing every single request.
The risk in this scenario is the proliferation of ungoverned workflows: automations built by business teams without IT oversight, AI agents launched without security policies, integrations built without considering the technical debt they generate. Gartner identifies AI platform governance and uncontrolled agent proliferation as two of the main risks of accelerated adoption. Speed of development is an advantage only if accompanied by the ability to maintain control over what gets built.
