The dominant story of 2025 – one that infused new challenges and opportunities for software development, delivery, security, testing, observability and more – was the rapid, widespread adoption of artificial intelligence.
Organizations have been kicking the tire on AI for the past several years, but 2025 saw an explosion of AI-powered offerings all along the software development life cycle.
We’ll take a look back on how AI impacted and disrupted every corner of technology, and in a separate series of 2026 prediction articles, we’ll look at what industry leaders see for the coming year.
AI in Software Development
They can be called copilots; they can be called coding assistants, but no fewer than 15 companies this year introduced AI tools that can generate code much more quickly than humans can write. The downside? These assistants often hallucinated when they couldn’t come up with an answer, and stressed developers by overwhelming them during code reviews to ensure the code met business goals, was secure, debugged and conformed to company policies. Companies also began to bring out AI agents that can spot vulnerabilities in code and offer remediation, as well as spotting anomalies when changes to the code base are made and integrating with other parts of the system.
There were some variations on the theme this year. In August, Codeium (rebranded as Windsurf, then split into pieces) introduced its Cortex assistant that, as opposed to autocomplete or code completion, supports large scale reasoning, code generation, reviews, and knowledge transfer, with greater accuracy, lower latency, and reduced costs.
Later that month, Google introduced an agent mode for VS Code and IntelliJ to expand the capabilities of Code Assist beyond prompts and responses to support actions like multiple file edits, full project context, and built-in tools and integration with ecosystem tools. It also had the ability to edit code changes using Gemini’s Inline diff, user-friendly quota updates, real-time shell command output, and state preservation between IDE restarts.
GitHub also made its Copilot coding agent available to users from anywhere on its UI as a lightweight overlay on GitHub.com. And Microsoft added Copilot-powered debugging features for .NET in Visual Studio.
Meanwhile, AI is impacting the developer experience by reducing the amount of code they are creating and seeing the role change. GitKraken has created a suite of tools that help developers work more efficiently with AI and development leaders gain the insights they need to make sure projects are on track. At the GitKon Conference the company produced earlier this month, it defined the Builder’s Era to elevate the craft and to integrate AI into the software life cycle.
AI in Testing
In 2025, test companies continued to add AI into their products, offering wider test coverage, automated script generation and the ability to predict when and where things might break.
Parasoft released more AI functionality this year, targeting C/C++ test automation, creating an AI agent for service virtualization, and last month launched AI-driven autonomous testing workflows for CI/CD pipelines.
Another industry leader, Appvance, in March launched its GENI generative AI offering that eliminates the need for manual testing, scripting and recorders by automatically converting English test cases to test scripts in bulk at a rate of 100 scripts per hour. In October, the company rolled out AI ASSERT, which allows testers to validate anything from animations to medical visualizations to 3D models, just by speaking in plain English what should be checked.
AI in Data
AI has changed how organizations work with and derive insights from their data. AI is great at data prep and generating Python or SQL code, and LLMs enable spoken querying to transform raw data into business intelligence. It’s that data intelligence that organizations rely on to stay ahead of their competitors.
While IBM, Microsoft and Oracle are among the leaders in space, other players have stepped up their AI game. For instance, Informatica, which was acquired by Salesforce in November, launched CLAIRE GPT, a conversational AI assistant to discover, analyze and execute complex data tasks.
And in April, Observe Inc. introduced an AI-powered observability data lake. The data lake stores structured and unstructured data, while the AI can identify patterns with a knowledge that identifies patterns across logs, metrics and traces.
Meanwhile, ValueOps by Broadcom, at its Spark VSM Summit in June, discussed how AI automates tasks, assesses risk and progress, and augments work like user stories, while VSM offers the alignment, visibility and metrics AI needs.
AI in Security
Software security companies have incorporated AI into their products to do things like find and fix vulnerabilities or analyze massive amounts of network traffic logs to detect anomalies.
For instance, OpenAI announced a private beta for its security researcher that can find and fix vulnerabilities in code, while GitHub added Copilot Autofix to GitHub Advanced Security, which can analyze vulnerabilities, explain their importance, and offer suggested fixes.
Companies this year also learned the hard way just how prevalent security risks in AI-generated code are. For example, a Veracode survey published in July found that AI coding assistants introduced vulnerabilities in 45% of 80 curated coding tasks across over 100 LLMs.
To combat these issues, OX Security launched VibeSec, which embeds dynamic security context into coding models to cut down on the number of insecure suggestions that are made in the first place.
