Keysoft’s journey: from legacy systems to AI-powered platform with Ratio1

In an era where rapid innovation and resilient infrastructure are paramount, Keysoft - a decades-old software provider specializing in enterprise business applications - embarked on a partnership with Ratio1.ai to modernize its legacy systems and introduce AI-driven capabilities. This two-phase technological transformation showcases how a traditional enterprise can leverage decentralized, edge-optimized cloud and specialized custom deep learning and AI agents to enhance scalability, agility, and developer productivity.

Keysoft is a Romanian software company (founded in 2001) known for its high-performance information management platform and bespoke enterprise applications. They serve a broad range of industries - from customs and finance to healthcare, media, and auditing.  Keysoft’s platform could take a project from concept to production in 1 month versus the typical 6+ months and this was for many years the key value proposition. Nevertheless, in a landscape of continuous innovation and technological disruption, the company management decided and urged the need to modernize its legacy architecture to keep pace with evolving technology standards. To achieve this, Keysoft partnered with Ratio1 - a pioneering decentralized cloud and AI platform - to execute a dramatic two-phase transformation. 

In Phase 1, Keysoft overhauled its deployment strategy by embracing containerization, continuous integration/deployment (CI/CD), and Ratio1’s Deeploy orchestration system. In Phase 2, Keysoft augmented its applications with advanced AI capabilities, leveraging Ratio1’s AI ecosystem to introduce intelligent code generation and domain-specific machine learning features. This case study details how Keysoft successfully navigated both phases, blending cutting-edge infrastructure with AI innovation to revitalize its product suite and better serve industries like finance, healthcare, and audit in 2025 and beyond.

Phase 1 - Modernizing Legacy Infrastructure with Ratio1 Deeploy

In the first phase, Keysoft focused on migrating from a traditional 3-tier legacy system to a modern cloud-native environment. Originally, Keysoft’s software was deployed in a classic on-premises model: a Windows-based application server (with a proprietary logic engine and “K-script” language) connected to a relational database (SQL Server, MySQL, Oracle) and a thin client GUI. This setup, while stable, posed typical legacy challenges - manual deployments, limited scalability, single points of failure, and difficulties in continuous updating. Keysoft’s development team often had to roll out updates by installing new versions on each client’s server, and scaling up for higher loads meant provisioning additional dedicated servers (with data synchronization hurdles). 

To eliminate these pain points, Keysoft adopted Ratio1 containerization and DevOps best practices, converting its application into Docker container images and establishing a CI/CD pipeline. Each new software version is now automatically built and packaged as a container, ensuring that it runs the same everywhere (solving the “works on my machine” problem). Containers encapsulate all necessary code, libraries, and settings, so deploying Keysoft’s app became as simple as launching the container on a target environment. Containers also start faster and use resources more efficiently than the old VM-based deployments which improved Keysoft’s resource utilization.

Enter Ratio1 Deeploy

Keysoft didn’t just move to any cloud - they moved to Ratio1’s decentralized orchestration platform, Deeploy, to deploy and manage these containers. Ratio1 Deeploy essentially brings “the power and flexibility of cloud-native orchestration (like Kubernetes) to the edge, but without relying on any centralized data center. Instead of hosting Keysoft’s app on a single server or a traditional cloud provider, Ratio1 uses a distributed network of nodes (potentially spread across multiple locations and owners) to run the containers. This unique approach, coordinated by blockchain and smart contracts, provides several tangible benefits:

• High Availability & Horizontal Scaling: Deeploy automatically handles replication and scaling of containers across multiple nodes. If demand for Keysoft’s application increases, Deeploy can spin up additional container instances on different nodes; if a node fails or goes offline, another node seamlessly takes over the workload . This means Keysoft’s services now have no single point of failure - uptime is greatly improved compared to the old single-server setup. Even during maintenance or outages in one location, the application remains available on other nodes, which is crucial for mission-critical uses (e.g. a healthcare client’s system can keep running locally even if cloud connectivity is lost)

• Decentralized Edge Deployment: Ratio1’s edge-first design allows Keysoft to deploy applications closer to end-users or data sources when needed. For example, a Keysoft installation for a hospital can run on hospital-owned Ratio1 nodes on-premises to meet data sovereignty requirements, while still being part of the global orchestration. Deeploy was built to handle the dynamic nature of edge computing with “extreme high-availability horizontal scaling” on heterogeneous nodes. Nodes can be regular servers, cloud VMs, or even PCs - any authorized device and KYB-ed node operator can join the Ratio1 network to contribute compute. The platform’s trust-minimized blockchain coordination ensures that tasks are assigned and verified transparently across these participants. In short, Keysoft can now run its software on a coalition of distributed mini-clouds rather than a single cluster, improving resilience and performance for clients in different regions.

• Seamless Container Orchestration & DevOps Integration: Developers at Keysoft found Ratio1 familiar in workflow. They package the software in standard OCI/Docker containers and declare resource needs (CPU, memory, etc.) as usual. The Deeploy platform then securely distributes and runs these containers on suitable edge nodes in the network. It takes care of crucial orchestration tasks automatically - starting/stopping containers, restarting them on another node if one fails, scaling out/in based on load, and monitoring health - all of this without any central Kubernetes master or manual intervention. This “DevOps-free” deployment model let Keysoft’s small team achieve reliability on par with big cloud setups, but without maintaining their own Kubernetes cluster or hiring dedicated site reliability engineers. Ratio1’s open-standards approach also meant Keysoft did not have to rewrite their software - the existing app could be migrated as-is into containers and deployed on Ratio1 with minimal changes. 

• Integrated Decentralized Services (R1FS & ChainStore): By moving onto Ratio1, Keysoft’s applications can leverage built-in decentralized services that enhance a cloud-native app. Two notable ones are ChainStore and R1FS. ChainStore is Ratio1’s Redis-like key-value store that operates across multiple nodes, allowing Keysoft’s container instances to share state and communicate in real-time even when running on different machines. For instance, if Keysoft’s app modules need a fast caching layer or a distributed session store in a scaled-out scenario, ChainStore provides that without relying on a single in-memory database server. R1FS is Ratio1’s distributed file system (built on IPFS) which gives Keysoft a private, encrypted storage layer for files, data exports, or backups, with full data ownership and no censorship. Instead of uploading documents or attachments to a third-party cloud storage, Keysoft apps can now store them on R1FS - the data is automatically synced peer-to-peer among the network and remains accessible even if one node drops, with content addressing ensuring integrity. These tools allowed Keysoft to match (or exceed) the capabilities of modern cloud platforms in their product: e.g., an audit firm using Keysoft can securely share large reports via R1FS (no central file server needed), and a CRM built on Keysoft can use ChainStore to keep multiple running instances in sync. All data is encrypted and stays within the customer’s trust boundary, addressing privacy concerns out-of-the-box.

• Cost Efficiency & Flexibility: By using Ratio1’s network, Keysoft can also potentially reduce hosting costs and gain flexibility in provisioning. The Ratio1 model taps into a pool of distributed compute resources - some provided by smaller data centers or even spare capacity - often at lower cost than traditional cloud VMs. Keysoft now pays only for the resources its containers actually consume on this shared network, in a usage-based model, avoiding the expense of idle dedicated servers. Deeploy’s economic design (with its token-based escrow and pay-for-uptime contracts) ensures that Keysoft gets reliable service from nodes while cloud service providers on the network compete at fair prices. In addition, Ratio1’s data sovereignty features allow Keysoft or its clients to specify where data should (or should not) run. Need to keep a database within the EU for GDPR compliance? Deeploy can restrict container placement to European nodes. Want an extra backup node on-premise? A local Ratio1 node can be added to the deployment. This level of control, combined with the inherent auditability of blockchain (every deployment action can be logged immutably), gave Keysoft and its enterprise customers greater confidence in moving to a cloud-like platform.

By the end of Phase 1, Keysoft had effectively transformed its infrastructure: moving from “old-school” manual deployments on single servers to an automated, containerized deployment on a decentralized cloud. The immediate outcomes were faster release cycles, as CI/CD pipelines deliver updates seamlessly to containers, and improved reliability due to Ratio1’s self-healing, multi-node design. Keysoft’s small team could manage many client installations through one interface (the Deeploy dashboard/API) instead of juggling individual servers. This lays a robust foundation where scalability, uptime, and global accessibility are no longer limitations for Keysoft’s software. It also set the stage for the next phase: capitalizing on Ratio1’s built-in AI capabilities to enhance the software’s functionality.

Phase 2 - augmenting Keysoft with Ratio1 native AI

Traditionally, Keysoft’s platform featured a DDL/DML SQL editor and a proprietary K-script for workflow automation. In partnership with Ratio1.ai, these editors evolved into high-level specification interfaces. At the same time, building on Ratio1’s neuro-symbolic framework and tokenized economy, Keysoft integrated domain-specific AI deep learning model pipelines. With the move to Ratio1’s ecosystem, Keysoft gained access to a suite of AI tools and “agents” that could be embedded into their applications. Phase 2 of the transformation leveraged these capabilities to introduce game-changing features powered by artificial intelligence, aligning perfectly with Keysoft’s mission of “designing business logic, not code”. In essence, Keysoft shifted from a traditional low-code platform into an AI-assisted development platform - drastically simplifying how new features are created and how end-users interact with the system. Last but not least the new proposed paradigm of Ratio1: “Your AI, you Data” where the final consumers are truly in possession of their data.

Natural Language to SQL (NL2SQL) Agent

One of the pain points in enterprise software is writing complex database queries or definitions by hand. Keysoft’s system historically provided a DDL/DML SQL editor for advanced users to define data schema or craft custom queries. Now, that module has been reinvented as a Specifications Editor, where users can simply describe in plain English what they need from the database, and an AI agent automatically generates the SQL code. For example, instead of manually coding a query to pull “the top 5 products sold in the last month”, a user can just ask for that in natural language - the AI will understand the request and produce the correct SQL behind the scenes. This is made possible by Ratio1’s specialized SQL Agent, an LLM (Large Language Model) fine-tuned for SQL generation. Such an agent “translates natural language queries into SQL commands to retrieve data from relational databases”, essentially acting as a query copilot. In Keysoft’s context, the agent is further customized with knowledge of the Keysoft database schema and business terms, so it can generate Keysoft-specific SQL that is syntactically correct and semantically appropriate for the application’s data model. The benefit is enormous: non-technical analysts can get the data they need without writing a single line of SQL, and experts save time on writing and debugging queries. This follows a broader industry trend of using LLM-powered “SQL assistants” to make data access more intuitive and self-service. Keysoft’s integration of an Ratio1 native NL2SQL agent has increased productivity and opened up data analysis to a wider audience within client organizations.

AI-Assisted K-Script Development

Keysoft’s platform includes a proprietary scripting language (“K-script”) used to implement custom business rules and workflows within the applications. Writing K-script traditionally required knowledge of Keysoft’s API and some programming effort - typically done by Keysoft’s team or trained power users. To drastically simplify this, Keysoft worked with Ratio1 to deploy an AI code generation assistant that can produce K-script code from a high-level description. Now, a business analyst can literally tell the system, “I need a feature that calculates late fees for overdue invoices and sends a reminder email”, and the AI agent will generate the corresponding K-script logic to implement that feature. This capability is akin to having a smart co-developer who understands Keysoft’s entire codebase and framework. In fact, Ratio1’s own flagship AI assistant, Ratio1 J33VES, is described as an AI “team simulator” that “understands your entire codebase, enabling rapid, full-context coding and problem-solving… from pinpointing bugs to generating production-ready code”. Keysoft leveraged this concept by training an AI model on its platform’s existing modules (the “100 individual business objects and modules” that Keysoft’s engine consists of) so that the AI has a private knowledge base of how features are implemented. This means the AI doesn’t create code in a vacuum - it writes new script in line with Keysoft’s established patterns and data structures, ensuring the generated code fits seamlessly with the existing system (no syntactic errors or incompatible logic). The AI essentially functions as a Keysoft expert developer on demand, available 24/7. It can take on multiple roles in the development process: writing code, suggesting improvements, checking for edge cases, and even explaining the code to the human user. This continuous dialog and refinement loop is similar to how advanced systems like J33VES work with developers until a problem is solved. The result for Keysoft is a radical reduction in development time and effort for customizations - many tasks that used to require a programmer can now be accomplished by a business analyst working with the AI assistant. This empowerment of non-developers through AI is a key trend in modern software: studies have shown “AI generation of boilerplate code and unit tests significantly reduces manual coding effort”, freeing up humans to focus on higher-level design. 

Keysoft’s new “describe it, we build it” approach exemplifies this, making enterprise software development more conversational and intuitive.

Specialized Predictive Models and Domain Assistants

Beyond code and queries, Keysoft also infused machine learning intelligence into the daily operations of its applications. Using Ratio1’s AI platform, they trained and deployed domain-specific predictive models to add value in different verticals. For instance, in the finance/audit sector, a Keysoft application can now include an AI model that analyzes transaction data to predict risk factors or flag anomalies (acting as an “AI auditor” assistant). In healthcare, a Keysoft module could leverage a predictive model to forecast patient appointment no-shows or optimize staff scheduling based on historical data. These models are built using deep learning techniques and hosted on the Ratio1 network, taking advantage of its support for distributed AI workloads (the platform can even perform federated learning across multiple nodes holding sensitive data, which is useful if Keysoft’s healthcare clients want to collaboratively train a model without sharing raw data). Ratio1 was designed to accommodate advanced AI/ML tasks - from training to serving - as part of its “AI OS” vision, so Keysoft could plug in custom models with relative ease. Moreover, Ratio1’s privacy-focused design (data remains on user-controlled nodes, and AI models can be served locally) means these features don’t compromise confidentiality. For example, an AI model providing financial forecasts to a Keysoft CRM system can run on the client’s own infrastructure via Ratio1’s edge node, ensuring no sensitive financial data ever leaves their premises - yet they still benefit from powerful AI analytics.

Keysoft personal assistants

To make these AI features accessible and even more, to offer users customized and specialised in-app assistants, Keysoft introduced decentralized specialized assistants within their software. Think of these as chatbot-like helpers embedded in the UI, tailored for specific roles or domains. 

Domain-specific advisors, like a Healthcare Assistant that can answer questions about patient data trends (using the predictive models) or a Financial Analyst Assistant that can generate custom reports and insights from accounting data on the fly. These are feasible because Ratio1 allows fine-tuning or grounding general LLMs on proprietary data and knowledge bases. In fact, Ratio1’s J33VES is described as extending to other domains, where if it’s loaded with legal contracts or financial data, it can assist a lawyer or accountant just as well as it assists a developer. Keysoft applied this principle by creating internal knowledge bases (using R1FS to store company-specific documents, policy texts, etc., and turning them into embeddings) so that the assistants can give informed answers.

Each of these assistants runs securely and locally via Ratio1’s decentralized AI serving infrastructure where basically the Keysoft containers can access local endpoints that provide the whole range of Ratio1 features - from J33VES to R1FS and ChainStore. Keysoft’s clients loved this because it meant they could have “ChatGPT-like” functionality within their business software without sending any data to external cloud AI services. All the AI processing happens either on the customer’s own Ratio1 nodes or across the encrypted Ratio1 network, ensuring compliance with data privacy regulations. (As Ratio1 emphasizes: “you run the AI on your own hardware (or someone’s hardware you trust)… no middlemen, no hidden data mining”. For many industries like healthcare and finance, this is a critical advantage - they get the benefit of AI-driven automation while keeping sensitive data fully under control.

In summary, Phase 2 turned Keysoft’s product suite into a much smarter, more user-friendly platform. Tasks that used to require technical skills (writing queries, coding business logic) have been democratized through AI natural-language interfaces. End-users can interact with the system in their own language, and the heavy lifting is done by specialized AI agents behind the scenes. Keysoft’s development team, on the other hand, is supercharged by AI - their productivity in delivering new features has multiplied, as routine coding is handled by the AI assistant (indeed, developers using tools like J33VES can build features “2-4× faster than before,” according to Ratio1’s observations). Quality is improved too: the AI, having full context of the codebase, generates consistent code and can catch integration issues that a human might overlook. 

All these enhancements translate into a better experience for Keysoft’s clients: faster customizations, richer analytical capabilities, and intelligent assistance available 24/7 within the application.

Phase 2 of Keysoft’s transformation - the AI-powered functionality - is currently in alpha stage and entering its initial testing round. While the architecture and core capabilities have been defined and partially integrated, broader availability will follow after further validation, refinement, and user feedback. This staged approach ensures the new AI features meet enterprise-grade quality and domain-specific expectations before public rollout.

Conclusion

From an industry perspective, Keysoft’s journey exemplifies the major tech trends of 2024-2025. Many organizations with legacy systems are looking to achieve exactly what Keysoft did: containerize and automate their legacy applications, move to cloud or hybrid-cloud infrastructure, and inject AI capabilities to stay competitive. Gartner and other analysts highlight containerization, CI/CD, cloud migration, and AI integration as key pillars of legacy modernization strategies in 2025. Keysoft hit all these marks in one cohesive project. Perhaps most importantly, they did so by leveraging a decentralized platform (Ratio1), which is an emerging approach that addresses concerns of privacy and control in the era of cloud computing. As more companies worry about data ownership and cloud costs, decentralized cloud frameworks like Ratio1’s could present an attractive alternative - “AWS-like experience for deploying apps, but on a distributed network”

Keysoft’s partnership with Ratio1.ai exemplifies how legacy enterprise applications can undergo a transformative two-phase modernization: first, by migrating to a decentralized container orchestration platform that ensures high-availability and scalability; and second, by embedding specialized AI agents to automate code generation, predictive analytics, and domain-specific workflows. 

This case study demonstrates a clear blueprint for organizations seeking to combine low-code platforms with cutting-edge AI and decentralized infrastructure to achieve unprecedented agility and innovation.

By leveraging Ratio1’s ecosystem for both infrastructure and AI, Keysoft achieved a comprehensive modernization that touches every aspect of its business. Key outcomes of this transformation summarized include:

• Greater agility and faster delivery

• Improved reliability and scalability

• Cost efficiency & resource optimization

• Enhanced functionality and user empowerment

• Alignment with industry standards 

Keysoft emerges as a disruptive tech innovator

Ultimately, Keysoft emerged from this journey not just as a software vendor, but as a disruptive tech innovator in its own right - one that combined its deep industry know-how with Ratio1’s state-of-the-art technology to deliver a solution that is faster, smarter, and ready for the future.

Next steps?

With the platform improvements, Keysoft can explore offering its software in new ways. The decentralized cloud approach means they could provide tokenized Keysoft-as-a-Service to clients globally, deploying on Ratio1 nodes near each client for performance. The efficient DevOps means even small custom projects are economically viable, possibly opening up new markets or smaller clients that previously couldn’t afford a custom solution. The AI features give Keysoft an upsell opportunity - e.g., tiers of service where advanced AI analytics or assistants are premium features. All told, modernization not only solved technical issues but also equipped Keysoft with a more scalable business model for growth.