In the technology field, what seems niche today often becomes tomorrow’s industry standard—and that’s definitely true of software development. From test-driven design and AI-powered code reviews to embedded intelligence and proactive security, leading-edge development practices promise to make modern software more resilient, efficient and adaptive than ever before.
Below, members of Forbes Technology Council share the software development approaches they believe will be commonplace within five years. Their insights reveal how today’s experimental techniques are shaping the next era of software development.
1. Runtime Component Visibility
Visibility into what software components are live in runtime will become standard. Static scanners often miss what components are active and exposed during execution. As software environments become more complex, this visibility gives security teams the context they need to assess current risks accurately, prioritize remediation and avoid being overwhelmed by irrelevant vulnerabilities. – Mehran Farimani, RapidFort
2. Test-Driven Development
Test-driven development will go mainstream—write a failing test, make it pass and then refactor. The payoff isn’t “more tests,” but design pressure and fast feedback. We will get clearer APIs, small decoupled units and safer refactors. Paired with continuous integration, ephemeral environments, contracts or property tests, and AI code, TDD will become the guardrail that ensures synthetic code is always production-ready. – Jefferson Ding, AI clade
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
3. Purpose-Built AI
Purpose-built AI in software development feels niche now but will be standard in five years. Instead of generic models, teams will use specialized AI trained on their industry’s rules to flag compliance gaps, optimize code and surface best practices in real time. Embedding these AI checks in development cycles will make compliance proactive, turning today’s edge practice into tomorrow’s baseline. – Scott Reynolds, UpCodes
4. AI-Powered Code Review
AI-powered code review and security validation will be embedded directly in integrated development environments. Today’s manual reviews and separate security scans will be replaced by real-time AI assistants that catch vulnerabilities, suggest optimizations and ensure compliance before code is even committed, with deep understanding of security patterns and business logic—making secure, efficient code the default, not the exception. – Neil Morris, REDAPTIVE INC.
5. Security-Oriented Unit Testing
Security-oriented unit testing will become standard. While functional tests are routine, security tests covering input sanitization or external dependencies remain niche, used mostly in regulated sectors. But with damage costs from cybercrimes projected to reach $10.5 trillion in 2025, proactive integration will soon be a standard part of every developer’s workflow. – Ankit Narayan Singh, ParallelDots, Inc.
6. ‘Continuous Everything’
“Continuous everything” is the next leap after CI/CD. Security, compliance, observability and performance will be embedded from day one and checked at every merge. Tools and standards are catching up, with AI now driving testing, audits and live insights. In five years, releasing code without continuous safeguards will look as outdated as shipping software without automated builds. – Darren Jack, AssetFynd
7. Ongoing Product Discovery
Ongoing product discovery—driven by real-time experimentation and user feedback—will become the norm. Unlike CI/CD, it shifts product development from roadmap-based to insight-based, ensuring teams build what users truly need. As AI and analytics evolve, embedding discovery into daily workflows will become standard, not just a competitive edge. – Kashif SaleemSubscriptionFlow
8. Agentic AI Integration
As a physician interested in healthcare software development, I think the integration of agentic AI systems directly into the software development lifecycle is of paramount importance. Agentic AI will impact autocoding, QA testing, documentation and even architecture design. – How are they?, Yale University
9. Cultural Awareness And Bias Control In AI Systems
We can safely assume that any software built within the next five years will have significant AI components. It is important that these AI components do not have bias built in and that they are culture-, identity- and context-aware. To me, these functionalities will be table stakes in five years, just as privacy, consent and security are today. – Syam Adusumilli, GroundGame Health
10. Everything As Code
A practice that feels niche now but could be standard in five years is Everything as Code—treating not just infrastructure, but also dashboards, alerts, policies, schema and observability as versioned, automated code. EaC improves consistency, auditability and speed while reducing manual errors and configuration drift. – Upendra Kumar, Capital Group
11. Health Price Transparency
For healthcare software, it’s health price transparency. Can you imagine buying products on Amazon without knowing the prices? That is how the U.S. healthcare industry operates today. Some leading health organizations are finally sharing this data, but this is still new. In five years, intuitive and transparent pricing information will become the standard for any patient-facing software. – Virgil Bretz, MacroHealth
12. AI-Assisted Development
AI-assisted development will be the norm. Right now it still feels niche, but in five years, using AI agents to generate code, write tests, refactor legacy modules and even review pull requests will be as standard as using version control today. Teams that learn to pair human judgment with AI copilots now will be miles ahead. – Kartik Subbiah, Healthlink Dimensions
13. ‘Everyone Ships’
Today, “everyone ships” feels edgy but achievable, given the right upskilling and access to low-code tools. In five years, it will be a requirement. With AI agents assisting in coding, testing and debugging, roles like product managers, support reps and even marketing teams will be able to contribute directly to the product. – Paul Derval, NinjaCat
14. The Product Operating Model
The product operating model will be the default way enterprises operate. Strategy, funding and execution will converge into a unified, outcome-driven model. AI will support prioritization and resource allocation, and in the next five years, teams will shift from project tracking to continuous value delivery through connected value streams, OKRs and real-time capacity planning. – Razat Gaurav, Planview
15. AI-Powered Development Task Automation
AI-led testing, coding and platform UI/UX design may feel niche today, but I believe they will all be standard within the next five years. As cloud and SaaS models evolve, most automation and AI tools will replace manual processes, from code testing to basic DevOps scripting. Even platform UI/UX is likely to shift toward conversational interfaces, with chatbots replacing traditional designs. – Naaman Raval
16. Natural Language Programming
Natural language programming will feel as normal as version control does today. Developers will describe features in plain English, and AI will handle the translation into production-ready code. Low-code/no-code tools will sit alongside it, making development more democratic. Just as in finance, where once-niche data signals became standard, these tools will redefine what it means to “write” software. – Artem Lalaiants, RiskSeal, Inc.
17. Data-Centric Development
In five years, data-centric software development will be standard. Apps will natively optimize data movement, locality and parallelism across hybrid and AI-driven environments. Today it’s niche, but embedding data intelligence into code will become essential to performance and innovation. – Sven Oehme, DataDirect Networks
18. Vibe Coding
Vibe coding—using natural language prompts to generate working software—may feel experimental today, but it’s the logical next step in development. In five years, prompting AI to build, test and refine code will be as standard as version control, transforming how teams prototype and deliver software. – Judit Sharon, OnPage Corporation
19. Software Supply Chain Security Management
Software supply chain security practices, like sourcing vulnerability-free open-source packages and images and auto-fixing, feel niche today but will be standard in five years. While many teams still rely on costly, reactive scanning and patching, attackers are weaponizing AI to execute attacks at breakneck speed. To compete, organizations will quickly need to adopt a more proactive, AI-driven approach. – Javed Hasan, Lineage
20. Composable Architecture
Composable architecture in enterprise software, especially with micro frontends and API-first design, feels niche now but will be standard in five years. It enables modularity, faster innovation and AI-native integration, aligning perfectly with the shift toward autonomous systems and adaptive user experiences. – Hemant Soni, CAPGEMINI AMERICA INC.
