OutSystems Unveils AI and Low-Code Outlook for 2024
I am always keen on visiting Canada (I love the food and the people are incredibly nice). So it was with great pleasure that I joined Forbes journalist Hessie Jones on stage at Collision 2024 in Toronto last week, for a fireside chat where we discussed GenAI and the Future of Software Development. I'd like to leave you with the 5 main takeaways from our conversation.
1. AI is redefining the "impossible"
Lately, I have met many enterprise CIOs who are forced to postpone legacy modernization almost to the point of no return. Companies have invested millions of dollars in solutions that are now end-of-life, the original developers are long gone, and there are no manuals they can refer to. In these situations, legacy has turned into a ball-and-chain attached to the feet of the business—keeping them from evolving or building new systems on top of what they have.
Part of my job is helping businesses determine how to release themselves from these shackles, and we've done it for hundreds of customers throughout the years. For example, global reinsurer Gen Re had a complex IT architecture that needed modernization. They chose OutSystems and ended up building a high-performance development team that delivered 30 applications in its first nine months.
Using OutSystems, teams can compress 4-year legacy modernization projects into 7 to 14-month projects. But the latest AI disruptions have brought us the potential to compress these development times up to two times further. We have infused our platform with AI agents (and made AI agents easy to build for customers too), particularly in steps that used to be done manually—adding these GenAI capabilities is bringing us orders of magnitude of compression.
This was one of the great things that came with the AI disruption–it has turned historically impossible legacy projects (due to no line of path to transformation) into easier, cheaper, and faster projects. They suddenly became possible.
2. Security and governance are top concerns in the GenAI era
According to Gartner, more than 50% of CIOs plan to adopt GenAI in the next 24 months. It's the first time that, in the face of an industry disruption, we don't see a division between visionaries, pragmatics, conservatives, and laggards—everyone is moving fast. But the level of uncertainty is also extremely high.
It's interesting to see how business applications change with the inclusion of AI capabilities. One of the questions CIOs are asking is "Where can I put AI in my software systems and processes in a way that helps me offer more business value?" One of the ways is grounding an AI agent in company information—so that the answers it gives aren't generic, based on the available LLMs alone, but also take into consideration the specific reality of the business because it was fed with company data. But how safe is it to mash company information with a bot?
In my conversation with Hessie Jones, I mentioned the example of a customer who decided to create a ChatGPT that was safe to use for the entire company. Hessie immediately asked me if there was such a thing as a "safe ChatGPT." And, indeed, this is one of the main concerns around GenAI today.
Deploying GenAI securely is making sure you have the necessary guardrails in place so that AI agents don't inadvertently leak sensitive company information. For example, if you are building an HR portal on top of an agent that's grounded on company data, you'll need to define policies to ensure which data is accessible to which employees.
There's a difference between building an AI agent and building actual working software that leverages it. You'll need to surround it with logic, governance, security rules, and policies that control its access to data. Only then can it become really effective. At OutSystems we have focused our efforts on the necessities that surround AI so that our customers can create AI-powered solutions that are both usable and secure. Our recent SOC 2 attestation is a testament to that commitment.
3. Explainability is a fundamental success factor
Another crucial aspect of using GenAI in software development that is very difficult to implement is the explainability factor. One of the ways GenAI is being used by developers is through "copilots," which fundamentally work like companions that help developers generate code.
However, what we're seeing is that the lines of code that are generated tend to be 50% longer than what you'd need if you wrote them by hand—you're creating a large chunk of technical debt just by using code that's being generated by "unsupervised" copilots. In OutSystems, our platform has been generating code from the beginning, so we faced this problem in the early 2000s. This gave us the advantage, right from the beginning, of starting to implement the necessary rules and best practices to ensure there's predictability in what's being generated.
Inversely, today's copilots will produce massive amounts of code that is unpredictable due to their nondeterministic nature. As they generate more code, systems will become more difficult to understand—to a point where, after a while, you'll need bots to generate code and other bots to explain what the first did.
Ultimately, code will become less and less important as a knowledge transfer mechanism. And one of the things about software developers is that they often want to understand the "why" of things—why did the AI generate that particular piece of code? What was its rationale? Explainability and observability will become major fields of research in the next few years and visual development will have a massive role to play in the face of this new reality. This is why I fundamentally believe that every software development platform in the future will look like a low-code platform.
4. The role of developers will change
The software industry is the last guild of artisans in the world. Paradoxically, nothing else is this manual in the twenty-first century. And that's why, when developers see something that automates the toil, the boring parts of their job, they jump on it. GenAI produces the capability to automate more tasks and has the potential to elevate the role of developers to something much more strategic and gratifying. By compressing the time it takes to deliver business value, and increasing pragmatism, developers can become architects, designers of very large systems, who get stuff done quickly and decide things at a more strategic level.
In the future, organizations will need more and more developers, especially those who can add value from a strategic point of view. High-level design skills will be in demand: the ability to come up with the right construction for a system to provide a particular type of outcome, or having a real understanding of how future disruptions can be integrated to get the maximum benefit.
GenAI will influence the role of software developers by elevating the profession, which will become more strategic, but it will also ask for more precision in some areas. On both fronts, low-code continues to tell developers "Your job is not to use the nail on the hammer. Your job is to redirect the machine to do amazing things."
5. Agility will get you ready for the future
I'll end with the first question that Hessie asked me on stage, which was about how my particular vision as a CEO has contributed to the OutSystems journey throughout the years. I believe that my major contribution has been to culture. Creating a culture of innovation, where failure is acceptable and you are not afraid to challenge those who tell you something can't be done.
This mindset started at the end of the 1990's. Back then, there was the perception that delayed software projects were simply a fact of life. CIOs and software developers usually blamed the business because they were the ones who couldn't get the requirements right from the beginning. But we challenged that notion. We decided to assume that requirements would always be wrong and that teams would always have to change and adapt mid-project.
That's how we looked at the problem, and we reverse-engineered it until we got to this notion of a platform that could fundamentally reduce the cost of change. That was how OutSystems was born. This ethos of challenging the status quo has always been part of our culture, and not only through me (because most of the innovations do not come from my head), but fortunately this is something that is now ingrained across our global team.
This notion of agility—having a team and a platform built for change—is what has allowed us to keep the cost of iteration so low and projects moving so fast. Agility has become a fundamental driver for us to build our product roadmap and continue to innovate. We can't predict the future, but we can be prepared for change. Our goal is that customers who choose our technology will never hit a wall regardless of what the future brings—so that adopting OutSystems is the last migration they will ever do.
