Combining Low-Code and GenAI for App Delivery
Artificial intelligence (AI) is extremely versatile, touching almost all domains, and it’s changing the world of software development. Generative AI (GenAI) in particular can deliver enormous value that spans the entire software development lifecycle (SDLC). However, it must be adopted with a critical eye.
In an OutSystems-sponsored CIO Dive webinar, “The Future of AI in Software Development: What You Need to Know,” OutSystems Co-Founder and AI Product Manager Rodrigo Coutinho and Michael Harper, Managing Director at KPMG US, discussed AI adoption and what is preventing broader implementation. This blog covers the highlights.
Shifting mindsets: AI adoption soars, but choice can be overwhelming
With so many opportunities on the horizon, it’s no wonder that 84% of software executives have already integrated AI into their SDLC processes, with 60% doing so in the last two years.
The effect of GenAI on software development
GenAI has been a major driver of this surge in adoption, largely in two key areas:
- Augmenting existing use cases for AI: Most software executives use AI for testing and quality assurance and security vulnerability. GenAI clearly augments these processes by supporting the creative side of testing while preventing you from always testing for the same things.
- Expanding adoption into the entire SDLC: Extending adoption to other domains, like app deployment and maintenance, provides untapped opportunities. For example, AI can do a better job of translating what users want—through natural language explanations—as part of building a working app. As a result, iteration times will be much faster. Yet, not many software executives are currently using GenAI for the SDLC. One of the reasons is that there is a confusing array of disparate AI tools (and few are GenAI) that only cover certain aspects of the SDLC.
How industries are using GenAI in development
While the adoption of GenAI in software development is growing around the world, there’s a lot of variation in its use in different industries:
- Banking and financial services: This sector has been one of the quickest to broaden its adoption of GenAI for development, especially for use cases such as fraud detection and financial forecasting.
- Healthcare: AI has proven instrumental in developing telemedicine capabilities that integrate clinical data and focus on patient outcomes.
- Manufacturing: AI has been making its way onto shop floors in building applications for use cases like predictive maintenance and supply chain management.
Sourcing AI tools
Currently there isn't one GenAI or AI software development tool to rule them all. Most businesses are working with hyperscalers or independent software vendors (ISVs), encouraging experimentation to identify exactly what works for them. This is likely to continue as companies work to reconcile different capabilities with different vendors, while trying to mitigate AI sprawl.
Seizing an opportunity: Expanding AI adoption to new domains
What are some of the new and emerging use cases for AI in software development and SDLC management? For one, the role of the developer is starting to change. It’s likely that soon, developers will spend less time on things like manually writing or reviewing code and more on supervising AI-supported workflows and outputs.
In the webinar, several key parts of development and the SDLC were identified as ripe for AI-driven innovation.
DevOps and code generation cycles
Instead of doing the first code review manually, GenAI can make sure it follows the right rules and patterns, thus bridging the gap between development and quality assurance. GenAI is especially promising in cases where the rules are very complex. For example, people can define rules and ask questions in plain English to prompt AI to check the code, the deployment plan, or even the underlying environment.
Documentation and knowledge transfer
GenAI can greatly enhance developer knowledge transfer by generating documentation not just for end users, but also when trying to change code created by someone else. GenAI can explain the changes, making it faster and easier to onboard new developers without derailing the project.
User experience (UX) design
GenAI can enhance user experience. For instance, if you’re new to a certain environment or use case, rather than creating something from scratch, GenAI can show the most common ways to do it. It can also provide suggestions for improving on your work.
Breaking down barriers: Bringing better governance to AI adoption
Some of the biggest barriers to the further adoption of AI and GenAI are concerns over privacy, security, and regulations. Also, many organizations feel they lack the skills needed to incorporate AI into existing workflows. This is particularly the case with GenAI. But, it is possible to overcome these concerns by:
- Addressing reputational risk: Issues of biased outputs and misinformation largely come down to the data used to train the model. But it’s only when GenAI is treated like a black box that provides an output but no explainability. With a cross-functional team keeping an eye on GenAI, it’s possible to address the specific risks and the potential issues they could lead to.
- Keeping ahead of regulatory mandates: Regulations concerning the use of AI are evolving quickly. Organizations should keep up with the developments and think about how to proactively adjust to new regulations by making sure that the right teams focus on the changes.
- Overcoming skills gaps: It’s possible to enable developers without AI experience to become comfortable with GenAI tools by using them for proof of concept and proof of value. They should be encouraged to innovate and experiment, albeit with guardrails in place, to get hands-on training and expertise.
- Integrating AI into existing workflows: By evaluating your existing tech stack to identify gaps and technologies that need to be modernized, you can then imagine how your workloads could work with AI to provide value.
AI and GenAI in software development: What success looks like
AI and GenAI-driven application development and SDLC management are inevitable. The key to success is strategic implementation with transparency, security, and privacy top of mind. The goal should be to generate value through enhanced productivity and agility–and to encourage developers to take on the role of AI orchestrator and acceptance tester.
To dig deeper into where Coutinho and Harper see development headed, watch the The Future of AI in Software Development: What You Need to Know webinar.
