The OutSystems Advantage in Low-Code for AI
Although it’s early days, some of AI’s benefits and impact for businesses are beginning to emerge. In everyday activities, tasks like summarising meeting transcripts, taking minutes, and undertaking financial analysis can be achieved with some success by the early machine learning tools we’re starting to see. Evidence thus far suggests that there are significant productivity benefits to be gained as the technology develops.
However, in contrast to the marketing rhetoric we’re bombarded with every day, deploying any AI solution is not as simple as signing up for a new SaaS product and setting off toward a brave new world. Existing problems around infrastructure and operational processes need to be addressed before contemplating any AI solution.
In a conversation with Tech Wire Asia, OutSystems‘ Richard Davies discussed the current situation in the organisations he interacts with during his working day and their experiences of modernisation.
Doing the groundwork
There’s certainly a buzz around AI’s potential, but Richard suggested that the claims that every company is well down the road of an AI-driven future are rarely true. “There is some AI adoption, but I would say it’s still very low. If I was to give a finger in the air estimate, [numbers of] anyone who’s done anything [with AI] semi-seriously, it’s probably around about 10%.”
There are clearly any number of tasks in every company that could be at least automated or even enhanced by AI, but the roadblocks that exist are the same ones that impede a cohesive data structure. Cohesive data structure is essential to quickly develop enterprise applications, whether or not they leverage machine learning. If software is easy to build – which low-code platforms like OutSystems make possible – it’s only partially effective with access to partial data.
The required core data, Richard said, is “[…] the important context that you need, the data held in your core systems, like customer data, core product data, things like that. […]. If you want [applications] to know something really specific, you’ve got to give them that particular information and ask very explicitly what you want.”
Bumps in the road
The impediments to getting visibility onto core information include technical issues like disparate data sources, legacy platforms that don’t ‘play well with others,’ incompatible data formats, information in formats difficult to parse, and a half dozen others. Plus, from the business’s side, there is the perception of risk associated with opening up data silos for ingestion (in the case of machine learning) or for use in applications that will automate and bring value. It’s addressing these issues that always form the first stage of implementation of OutSystems’ low-code platform, for example. But doing the necessary work and taking on those challenges paves the way for future AI use, a use that’s beyond a simple, smart customer chatbot, at least. Building applications quickly that leverage AI will be possible; as fast as any other type of application – yet any app is contingent on a coherent base of data.
AI in low-code
While AI applications can be built and used by OutSystems customers for tasks in their business’s domain, Richard pointed out a secondary area where the technology is put to use – in the development process of enterprise applications taking place inside the OutSystems low-code environment.
A company’s internal software developers have low-code in their range of tools, and AI becomes another. “So it’s like a productivity booster, probably the biggest one being the ability to generate full-stack apps with a text prompt and modify them.” Both business apps and development tools that use AI are undoubtedly useful, but the majority of interest, as Richard has seen it, remains the applications companies can build and use to solve particular problems.
“To be honest, most people are actually much more interested in the business software: ‘My business needs this particular [app] to look at this particular use case’, rather than ‘Look how cool OutSystems is because I can build apps with text prompts’. Which is actually pretty cool, but just not as relevant to as many [of our] customers.”
Next steps
Assuming an organization can rationalize its data resources so they become a valuable asset, there is definitely great potential for machine learning technology. Companies are already able to pick and switch between the available large language models and use their inferences.
Adding relevant, local data is the next stage, as is prioritizing the business cases for automation, applications, or solutions that will ease manual processing. Careful planning of any technological initiative in the context of the wider business needs remains imperative, whether the intended outcome is a wholesale, AI-first operational makeover or giving the in-house software development team the low-code tools needed to iterate quickly on essential enterprise apps.
Reach out to an OutSystems representative near you to find out more about low-code development, the steps needed to create a valuable data fabric, or how you can develop the next generation of enterprise-grade software with AI. You can schedule a demo here.
READ MORE
- Unlocking Competitive Advantage: APAC’s Transformative Financial Payments Landscape
- The new age in running an MSP in A/NZ & beyond
- Getting Off The Uphill Payroll Cycle
- Data Strategies That Dictate Legacy Overhaul Methods for Established Banks
- Securing Data: A Guide to Navigating Australian Privacy Regulations