AI Enters the Codebase
What if you could talk to your code—and it talked back with meaningful answers? That question isn’t theoretical anymore. With the integration of ChatGPT into SMART TS XL, development teams can now extract insights, document logic, and make high-impact changes faster—simply by asking.
SMART TS XL already gives teams visibility across legacy and modern systems, spanning COBOL to cloud, JCL to APIs. It indexes, maps, and analyzes millions of lines of code across distributed platforms. But now, with conversational AI layered on top, the platform doesn’t just show you what’s happening. It helps you understand it, convert it, and act on it.
This isn’t about replacing developers. It’s about accelerating their thinking—through explainability, automation, and system-wide intelligence that responds in real time.
What Is SMART TS XL and Why It’s Built for Integration
SMART TS XL is more than a code search engine. It’s an enterprise-grade system analysis suite that builds a live, navigable inventory of everything inside your application landscape. From COBOL copybooks to SAP function modules, it gives users fast access to what exists, what it connects to, and what might break when you change it.
A Platform-Agnostic Powerhouse
One of SMART TS XL’s greatest strengths is its platform independence. It works across:
- Mainframe systems (COBOL, JCL, PL/I, etc.)
- Midrange (AS/400, RPG, CL)
- Distributed applications (Java, Python, C#, etc.)
- Databases (SQL Server, Oracle, DB2)
- Web services, shell scripts, SAP, and beyond
Because it doesn’t depend on any one environment, SMART TS XL becomes the unified source of insight across silos—making it the perfect foundation for intelligent, AI-driven interaction.
Full-System Awareness Across Legacy and Modern Code
In a single SMART TS XL instance, users can index hundreds of thousands of programs, millions of lines of code, and every related data element. It understands not only the content of your codebase but also:
- Which programs call each other
- Where datasets are defined and used
- How job streams interconnect
- What business logic lives inside programs
This system-wide awareness is what makes ChatGPT integration powerful. Instead of throwing isolated code snippets at an LLM, SMART TS XL contextualizes them—automatically.
The Scale Behind the Intelligence
Even in a modest demo system, SMART TS XL may process over 50 million lines of code across 450,000+ artifacts. At enterprise scale, that grows exponentially. And yet, search results return in seconds, linking data across every layer of your environment.
Now, add to that an intelligent assistant that understands natural language, can explain logic in plain English, estimate development effort, or convert COBOL to modern formats—and you have a toolset that shifts how teams interact with legacy systems altogether.
How ChatGPT Is Integrated into SMART TS XL
Integrating ChatGPT into SMART TS XL didn’t mean adding a chatbot into a developer’s workflow. It meant enhancing a precision engine with conversational understanding—layering a powerful natural language assistant directly into a deeply technical environment. And it was designed to be as flexible as it is secure.
Safe by Design: Client-Controlled AI Access
Not every organization is ready to use AI tools out of the box—especially in highly regulated environments. That’s why SMART TS XL puts full control in the hands of each client. By default, the system does not connect to any external AI services. Instead, administrators explicitly configure and enable GPT access, ensuring complete control over when, how, and where AI features are used.
For clients who do allow AI, integration is streamlined and fully modular. Once enabled, users gain access to intelligent automation without compromising security or compliance boundaries.
Two Paths to Activation: Select Code or Full Program
SMART TS XL offers two intuitive ways to activate ChatGPT on your code:
- Select Any Lines of Code: Users can highlight a specific code segment from any language—COBOL, Java, Python, PL/I, RPG, and more—and trigger GPT. A pop-up appears with options like:
- Explain this logic
- Annotate or document
- Convert to another format
- Ask a custom question
- Analyze the Entire Program: Users can also choose to analyze an entire program at once. A single click loads the entire member into the ChatGPT pane, making it ideal for documenting purpose, converting structures, or extracting high-level business logic.
This two-pronged approach allows teams to work top-down or bottom-up, based on what they’re trying to understand or improve.
Three Modes of Asking: Free Text, Static Commands, Interactive Suggestions
Not every developer wants to type prompts. And not every prompt needs to be written from scratch. That’s why SMART TS XL includes three ways to interact with GPT:
- Free Text Entry – Just type what you want.
Example: “Explain what these lines of code do in plain English.”
GPT responds immediately, translating raw logic into understandable insight. - Static Suggestions – Prebuilt, one-click commands that perform specific tasks instantly.
Examples:- “Convert this field to a SQL Server table”
- “Summarize this program in pseudo code”
- Interactive Suggestions – More advanced options that ask for user input before executing.
Examples:- “Translate this code to another language” (asks: which language?)
- “Estimate the effort to change a field from length X to Y” (asks for field name and lengths)
This triad of interaction styles ensures that both power users and casual users can benefit from the AI—no learning curve required.
Real Use Cases: What You Can Do With AI in SMART TS XL
Integrating ChatGPT into SMART TS XL is more than a novelty—it unlocks meaningful use cases that improve developer productivity, modernize legacy systems, and accelerate understanding across platforms. Whether you’re documenting business rules or estimating effort for a field change, GPT provides fast, contextualized answers that used to take hours of manual review.
This section breaks down how real-world teams are using ChatGPT inside SMART TS XL to work faster and smarter.
Explaining Code in Plain English
One of the most powerful and immediate applications is having GPT explain code in human terms. A developer can highlight a few lines of COBOL, RPG, Python, or any other language and ask:
“Explain in detail what this code does.”
Within seconds, GPT provides a line-by-line description in plain language, followed by a concise summary. This is especially useful for onboarding, reverse engineering, or understanding legacy logic someone else wrote 20 years ago.
The results can be saved, printed, or exported for documentation.
Auto-Documenting Programs with COBOL Flower Boxes
COBOL programs often lack structured documentation. With a single click, SMART TS XL allows users to send an entire program to GPT and generate a standard flower box comment—those classic COBOL-style header blocks that describe the purpose, input, and output of the program.
The AI formats the documentation and gives teams a copy-paste-ready comment block that improves readability and institutional knowledge.
Code Conversion, From Fields to Full Language Translations
ChatGPT isn’t just explaining—it’s converting.
Users can highlight a single COBOL field definition and invoke a static suggestion like:
“Convert to SQL Server view.”
GPT returns a CREATE TABLE statement with the correct structure and data types. At the program level, developers can select entire routines and use interactive suggestions to convert them into another language:
“Convert this COBOL program into NATURAL.”
GPT prompts the user for the target language, processes the logic, and returns a translated version. This dramatically accelerates language transition work and supports modernization.
Business Rule Extraction and Pseudo-Code Generation
Understanding business rules buried deep in procedural logic is a huge challenge for enterprises. GPT can now help solve that.
With a single click, SMART TS XL can extract business rule descriptions from code—summarizing them in a table that includes:
- Rule name
- Description
- Purpose
- Line numbers where logic appears
Similarly, developers can ask GPT to summarize an entire program in pseudo code, creating a high-level flow that’s easier to review with non-technical stakeholders or during architectural planning.
Estimating Level of Effort for Changes
Change estimation can be subjective. SMART TS XL now allows users to run GPT-based effort estimation tasks directly inside the interface.
Example: a developer wants to change a COBOL field from length 1 to length 5. They simply:
- Enter the field name
- Provide the current and desired length
- Click “Estimate change effort”
GPT returns a breakdown by task: understanding the code, applying changes, testing, QA, deployment—each with an estimated number of hours. The entire estimate is saved and exportable as a report.
Building Interoperability: From COBOL to MongoDB and C
In one of the more advanced demos, SMART TS XL was used to:
- Select a COBOL record layout
- Ask GPT to convert it into a MongoDB schema
- Generate both a COBOL program and a C# program that could read/write that schema
This effectively created a communication bridge between a mainframe and a distributed system—based on AI-generated code, using a shared document model.
The entire workflow was accomplished in moments and saved for export in the Action Center.
Empowering Teams With Developer-Inspired Suggestions
The most impressive part of SMART TS XL’s ChatGPT integration isn’t just what the AI can do—it’s how easily the system evolves based on what real developers need. Suggestions aren’t static. Teams using the platform actively shape the GPT assistant by contributing ideas, workflows, and specialized use cases that are deployed in near real time.
This section explains how suggestions work, how they’re built, and how organizations can tailor AI-powered actions to their environment.
How Smart Suggestions Are Built and Shared in Minutes
Every suggestion in SMART TS XL is backed by a predefined GPT prompt, carefully crafted to extract a specific result. But unlike other AI assistants, these are not locked behind rigid templates. When users think of something new—a task, a transformation, a documentation format—they can click “Ask a New Suggestion” and submit it directly to the IN-COM team.
Within 10 minutes, that request can be:
- Added to the backend suggestion library
- Pushed live across the SMART TS XL user base
- Available to anyone with the GPT integration enabled
This keeps the system dynamic, developer-driven, and continuously improving. If a team identifies a repetitive task—like summarizing a job stream or analyzing copybook reuse—they can automate it quickly and make it available org-wide or enterprise-wide.
Letting Users Request New GPT Actions in Real Time
The GPT assistant isn’t limited to a fixed menu. Every developer has the ability to contribute and request:
- New code translation targets
- Domain-specific formatting (e.g. compliance headers, audit blocks)
- Documentation or testing templates
- Legacy system insights unique to their architecture
Once submitted, these requests become shareable assets—allowing teams across departments, regions, or business units to standardize and reuse AI-powered workflows that match their environment.
It’s GPT as a collaborative toolkit, not just a passive assistant.
Enabling Custom Queries and Search Intelligence with GPT
SMART TS XL also allows teams to integrate GPT with its powerful query assistant—a structured syntax engine that lets users build intelligent searches across their entire codebase. This means you can:
- Ask GPT to build a proximity or block query
- Use GPT to convert sample logic into a Smart TS-compatible search string
- Analyze code structure and detect duplication based on GPT-enhanced search criteria
Example:
“Find all code segments that look like this logic block.”
GPT analyzes the selection, generates a smart search string using Smart TS syntax, and executes the search—finding duplicates, clones, or pattern matches across the entire enterprise.
This combination of structured querying and conversational intelligence turns SMART TS XL into a hybrid environment: one that understands both code structure and developer intent.
Why This Integration Matters
The combination of SMART TS XL and ChatGPT is more than a clever overlay—it’s a shift in how teams engage with their systems. AI doesn’t just speed up analysis. It transforms the relationship between humans and legacy code. Where once understanding required a specialist’s memory or hours of reading, now it takes a question and a click.
This section explores why this matters for enterprise teams, and how it supports faster, safer, and more confident system evolution.
From Static Metadata to Conversational Insight
Traditional metadata repositories provide structure, but they don’t explain logic. You can see relationships, field types, and call graphs—but you can’t ask them why something was built a certain way or what it actually does.
ChatGPT turns that static structure into a living conversation. You can now:
- Ask what a routine does and get an answer in plain English
- Request documentation and receive it in minutes
- Query the business intent behind logic, not just the technical syntax
This bridges the gap between technical and non-technical stakeholders, making legacy systems more accessible to architects, analysts, and modernization teams.
Cross-Platform Understanding for Modernization Teams
Most modernization efforts struggle not because of tooling—but because of blind spots. Teams don’t know:
- Where logic lives
- How programs connect
- What risk is introduced by changing one field or job
By combining full-codebase visibility with ChatGPT’s explainability and conversion tools, SMART TS XL eliminates those blind spots. It becomes the single pane of understanding for:
- Estimating modernization effort
- Translating old logic into new formats
- Designing new services without duplicating legacy rules
With this integration, modernization becomes less about deciphering code—and more about designing the future.
From Legacy to Cloud with Less Guesswork
Whether migrating to cloud-native services, adopting data lakes, or integrating with APIs, one constant remains: you can’t move forward confidently without understanding what’s behind you.
The GPT integration provides:
- Clear summaries of legacy logic
- Migration-ready documentation
- Language and platform translation support
- Testing and QA guidance with effort estimates
This reduces rework, shortens planning cycles, and ensures that modernization teams build with precision—not guesswork.
Code Talks Back, and Teams Move Forward
For decades, understanding legacy systems meant navigating complexity—line by line, job by job, function by function. Even the best development teams relied on documentation that was outdated, experts who were overloaded, and manual processes that slowed down every step of change.
With SMART TS XL and ChatGPT integrated, that era is coming to a close.
Now, developers can ask questions in plain language—and get answers in return. They can extract business rules, translate logic, document programs, and estimate effort without leaving the tools they already trust. What once took hours now takes minutes. What once required niche expertise is now accessible to the entire team.
But more importantly, this integration doesn’t remove the developer from the equation—it empowers them. It amplifies knowledge. It accelerates learning. It supports faster decisions and safer changes in environments where one missed detail can create cascading issues.
This isn’t just a new feature. It’s a new way of thinking about code: as something that can speak, respond, and collaborate.
And in a world where software complexity is only growing, that conversational clarity might just be the most important upgrade of all.