Learning Library

← Back to Library

Generative AI Accelerates Application Modernization

Key Points

  • Modern applications are deeply embedded in daily life, and their heterogeneity and inter‑dependencies across an organization make upgrades risky without comprehensive, enterprise‑wide planning.
  • Application modernization means updating legacy systems with modern capabilities to generate new business value, driven by goals such as leveraging innovation, boosting productivity, or meeting compliance requirements.
  • The modernization journey typically follows two stages: an advisory phase that inventories and assesses each application’s fate (retire, SaaS, containerize, refactor, or rewrite) and a planning phase that sequences migrations while ensuring coexistence with legacy assets.
  • Generative AI can dramatically accelerate both phases by automatically summarizing poorly documented “spaghetti” code, identifying change‑needed components, and generating documentation, cutting advisory cycles from weeks to days.
  • By using AI to uncover embedded business logic early in the planning stage, organizations can create more accurate migration roadmaps, reduce reliance on scarce specialist developers, and achieve faster, lower‑risk modernization.

Full Transcript

# Generative AI Accelerates Application Modernization **Source:** [https://www.youtube.com/watch?v=AINoTKMZSGI](https://www.youtube.com/watch?v=AINoTKMZSGI) **Duration:** 00:07:43 ## Summary - Modern applications are deeply embedded in daily life, and their heterogeneity and inter‑dependencies across an organization make upgrades risky without comprehensive, enterprise‑wide planning. - Application modernization means updating legacy systems with modern capabilities to generate new business value, driven by goals such as leveraging innovation, boosting productivity, or meeting compliance requirements. - The modernization journey typically follows two stages: an advisory phase that inventories and assesses each application’s fate (retire, SaaS, containerize, refactor, or rewrite) and a planning phase that sequences migrations while ensuring coexistence with legacy assets. - Generative AI can dramatically accelerate both phases by automatically summarizing poorly documented “spaghetti” code, identifying change‑needed components, and generating documentation, cutting advisory cycles from weeks to days. - By using AI to uncover embedded business logic early in the planning stage, organizations can create more accurate migration roadmaps, reduce reliance on scarce specialist developers, and achieve faster, lower‑risk modernization. ## Sections - [00:00:00](https://www.youtube.com/watch?v=AINoTKMZSGI&t=0s) **Challenges of AI‑Driven Application Modernization** - The speaker explains how pervasive, heterogeneous software systems require organization‑wide, AI‑guided planning to upgrade legacy applications without disrupting complex interdependencies. - [00:03:03](https://www.youtube.com/watch?v=AINoTKMZSGI&t=183s) **AI-Driven Legacy Code Transformation** - The passage describes how AI can speed up planning, reverse‑engineer, generate, and automatically convert legacy code to modern platforms, citing IBM’s real‑world success. - [00:06:14](https://www.youtube.com/watch?v=AINoTKMZSGI&t=374s) **Embracing Continuous AI Transformation** - The speaker highlights the rapid rise of generative AI models for application modernization, urging organizations to begin AI‑driven transformation now to gain a competitive edge. ## Full Transcript
0:00There is software all around us. 0:02From the gadgets we use, to how we communicate, 0:05it's embedded everywhere. And because applications 0:08are so integrated in our everyday lives, 0:11we need to make sure that they are maintained 0:13and upgraded constantly. 0:14The challenge when it comes to modernizing applications 0:17lies in heterogeneity and complexity. 0:20An organization, as you can imagine, 0:22is built on multiple systems 0:24and each system is built on many applications. 0:27These applications, in turn, can rely on external services, 0:31like APIs, to perform specific functions. 0:35Each piece plays its part. 0:36Upgrading any one component could potentially 0:39disrupt this delicate balance. 0:41So, to modernize successfully, the planning must span 0:45the entire organization. 0:47That's where I come in. 1:05Welcome to AI Academy. 1:07My name is Maja Vukovic. 1:08I'm an IBM Fellow and Researcher specializing 1:10in strategies for AI driven application modernization. 1:14So, what is application modernization? 1:16Simply put, application modernization involves 1:19taking an outdated computing system 1:21and updating it with modern, well-aligned capabilities 1:24and features to create new business value. 1:26Businesses are constantly reinventing their ways of working 1:29for multiple reasons. 1:30To take advantage of the latest innovation, 1:33to enhance productivity, or sometimes just to avoid 1:36non-compliance risks. And every time business operations 1:40are refined, the applications powering them 1:43need to be modernized as well. 1:45Imagine a system with a lot of unstructured code-- 1:48tangled, messy and difficult to read. 1:51To make sense of this spaghetti code, 1:53you either need the subject matter expertise 1:56or you need to apply complex techniques 1:58that can help you extract it. 1:59But because of the shortage of developers 2:02with the right skills, it's easier said than done. 2:05This is where generative AI can make a critical difference 2:08to accelerate your application modernization journey. 2:15Once you make the decision to modernize, 2:17the process will kick off in two phases. 2:19The first is the advisory phase. 2:21Here you take the portfolio of applications 2:23written in different languages on different platforms 2:26and understand which among them need to be retired, 2:28transformed to software-as-a-service (SaaS), 2:30containerized, refactored, or rewritten. 2:32This is to understand the ROI 2:34and figure out if a transformation is needed at all. 2:37Next is the planning phase, where you lay out 2:40the right order to migrate the applications, 2:42understanding how they will coexist 2:44with legacy applications and test them. 2:46The advantages of generative AI start to take effect 2:49as soon as you begin the modernization process. 2:51In the advisory phase, for instance, 2:53you can use generative AI to summarize the legacy code, 2:56which very often is poorly documented, 2:59identify pieces of it that need to be changed 3:01and generate documentation for it. 3:03As a result, in some scenarios, AI can help shorten 3:06the advisory cycle from weeks to days. 3:09As you begin the planning phase, 3:11you may first want to understand the business logic 3:13that's embedded in the existing code. 3:15With reverse engineering and AI, you can quickly summarize 3:19the code and analyze legacy applications to decide 3:22what to retain and remove. 3:24Once you have figured out what to do with the code, 3:27the code generation capabilities of AI let you create 3:30millions of lines of code very quickly, 3:32saving valuable time and labor. 3:35With just the natural language prompt, 3:37developers get a head start, or in some cases, 3:40even fully generated code where developers simply validate. 3:43It's also likely that you will have some legacy applications 3:46that are not compatible with the new platform, 3:49a cloud platform, for instance, 3:51and you need to retain these applications. 3:53In these cases, you can use code conversion to rewrite 3:56or rather translate the existing languages to newer ones. 3:59With AI, there is the opportunity to automate 4:02the conversion and enable it for many languages. 4:06During the conversion process, developers can now use AI 4:09to identify and eliminate similar code snippets 4:12and replace them with common services. 4:14And beyond theory and strategy, we are seeing success 4:18in real world scenarios. 4:20IBM has been working with a large IT organization 4:22on their transformation. 4:24They were looking for their legacy code 4:26to be automatically converted to the new target language. 4:29And with AI existed code conversion, we were able 4:32to automatically translate about 85% of this custom code. 4:36And not only that, this translated code also aligned 4:40with their existing DevOps practices, 4:42so massive time and cost savings. 4:44In another instance, a large banking client was looking 4:47to modernize and refactor their applications 4:49built over the past decades. 4:51They wanted to make applications more agile in the cloud 4:54while avoiding significant investments in redevelopment. 4:57We used AI to recommend the components 5:00and generate the stubs for these new components 5:02and identify any debt code, 5:04any unused code in their applications. 5:06With AI, we were able to accelerate the process 5:09of refactoring and reduce the application transformation 5:11time from months to weeks. 5:14Think about how this was done in the past. 5:16Transformation planning, to give an example, 5:19was done through extensive rules and code in spreadsheets. 5:22When you have hundreds of servers 5:23and applications to migrate, imagine the complexity 5:26of managing it all through these sheets. 5:28Every time you hit a bump, you had to revisit 5:30the spreadsheets to make changes. 5:32And to further complicate matters, 5:34each customer has their own unique requirements, 5:36which means you had to manually craft these journeys. 5:39AI helps to radically streamline the workload migration 5:42to the new architecture. 5:44It also enables teams to observe workload performance 5:47in the new environment 5:48so they can optimize and refine the process. 5:51With AI, you gain the capability to adapt on the fly. 5:59One of the biggest advantages of AI assisted transformation 6:02is the ability to learn and adapt faster. 6:05There's a saying that has always held true in tech, 6:08"What's new today is legacy tomorrow." 6:11The point is that modernization isn't one and done effort. 6:14It's an ongoing journey. 6:16So, having the best processes and practices in place 6:19is often more important than one specific 6:21modernization outcome. 6:23I love this stuff. 6:24The challenge of getting better and faster every day! 6:26We are in what feels like a golden moment 6:29for transformation. 6:30There's so much innovation happening 6:32that my Slack channels are always lighting up. 6:35A new model comes up practically every minute 6:37across various domains. 6:39And the large language model trained for coding 6:41that we rely on for application modernization 6:44is also seeing a lot of action. 6:46And with the pace of innovation in AI models, 6:48this is just the beginning. 6:50So, what does the future hold? 6:53With the developments in AI coming in fast and furious, 6:56the future will be full of endless opportunities. 6:59Imagine generative models teaming up 7:01and collaborating on tasks, 7:03or systems starting to self-heal and self-evolve. 7:06Remember the periodic upgrades to applications 7:08I talked about earlier? 7:10What if the applications could seamlessly transform 7:12and get better on their own? 7:13Compared to traditional transformation, 7:16this stuff sounds like science fiction, 7:18but it's going to get real very quickly. 7:20The possibilities are limitless, and my advice to everyone 7:24is to start planning now. 7:25Whether that's on your own or with a partner, 7:28kick off your AI transformation practice today 7:31because it's going to be a huge competitive 7:33differentiator going forward. 7:35Thank you for watching. 7:36Please join us again for future episodes of AI Academy 7:39as we explore some of the most important topics 7:42in AI for business.