Satya Nadella recently stirred the tech world with his prediction that SaaS applications, as we know them, may soon become obsolete. He suggested that AI agents could replace the business logic traditionally coded into these applications, reducing the complexity of querying databases and interacting with business systems. While this vision is inspiring, realizing it in the context of complex transactional systems like CRMs and ERPs is far from straightforward. However, it offers a glimpse into a transformative approach to building and evolving these systems—a future powered by AI-driven code generation.
Let’s dissect this vision and explore how it might redefine the development of ERP systems like Lensing ERP.
The Challenge: Business Logic in Transactional Systems
Business logic in systems like CRM and ERP is not limited to simple database queries. It encompasses intricate workflows, validation rules, multi-step transactions, and dynamic decision-making. For example:
In a CRM, sales pipelines rely on multi-level approval processes, automated follow-ups, and quota-based forecasting.
In an ERP, revenue recognition requires adhering to intricate accounting standards, consolidating data across subsidiaries, and managing cross-border transactions.
These processes go far beyond generating SQL queries from natural language prompts. They require extensive transactional logic that integrates various system components, maintains data integrity, and adheres to local regulatory requirements.
Replacing such logic with AI agents necessitates not only the ability to “understand” natural language but also to generate, maintain, and interconnect code modules that handle these complex processes.
The Solution: AI-Driven App Development
Rather than viewing AI agents as replacements for business applications, we should see them as catalysts for a new paradigm in app development. Here’s how:
1. Automating Specification and Design
Modern AI models, like large language models (LLMs), excel at synthesizing information and generating structured outputs. For example:
A user could prompt an AI agent with: “Design a sales planning module for a CRM system that includes quota management, forecasting, and approval workflows.”
The AI could generate a detailed software specification, complete with user stories, data models, and process flows.
2. Generating a system architecture
Based on a detailed design- and specification document, one can enter an iterative "dialogue" with a LLM to come up with a system architecture, laying out the right APIs for data storage, configuration, and UI layer (to name a few).
3. Generating Modular Code
Once the specification and architecture is in place, a new breed of software engineers can use code generation tools (IDEs) like Windsurf to iteratively generate modular code for
each business function.
APIs, database schemas, and user interfaces.
ensuring compliance with architectural patterns like microservices or event-driven systems.
4. Integrating and “Gluing” Components
Building transactional systems requires seamless integration of modular units. This is where AI agents shine:
AI Agent Integration: AI can generate “glue code” that connects independently developed modules.
Validation: AI can verify data consistency and ensure compliance with system-wide constraints.
Testing: AI-driven testing frameworks can automatically generate test cases to validate integration.
5. Iterative Development and Adaptation
AI agents thrive in iterative environments:
By monitoring system performance and user feedback, they can identify bottlenecks or gaps in functionality.
They can regenerate or modify code to adapt to changing business needs, reducing reliance on manual updates.
The Impact: Building Business Apps at Unprecedented Speed
This AI-driven approach to app development has profound implications for ERP systems:
1. Faster Development Cycles
Traditional app development involves lengthy cycles of requirement gathering, design, coding, testing, and deployment. AI-driven code generation could reduce these timelines by an order of magnitude.
2. Democratizing Development
Business stakeholders could interact directly with AI agents to specify requirements, reducing the dependency on technical experts for initial development.
3. Enabling Hyper-Personalization
AI agents can tailor ERP systems to the unique needs of each organization. For example:
A SaaS company could request an AI agent to design a subscription billing module that aligns with its revenue recognition policies.
The same system could be adapted for a manufacturing firm by generating modules for inventory management and production planning.
4. Continuous Evolution
With AI agents, ERP systems can evolve continuously:
New features can be added incrementally without major disruptions.
AI-driven monitoring can proactively suggest enhancements based on usage patterns.
Lensing ERP: Pioneering the Future
At Lensing Systems, we’re not waiting for this future—we’re building it. Our ERP system is designed to harness the power of AI agents and generative technologies:
Sandboxing Concept: Inspired by version control systems, our sandboxing feature allows developers to experiment, iterate, and integrate AI-generated modules seamlessly.
Native AI Integration: Our platform incorporates AI for tasks like report generation, predictive analytics, and now, app development.
Vertical Solutions: By leveraging AI, Lensing ERP enables rapid development of industry-specific solutions, from subscription management for SaaS companies to compliance automation for financial services.
Imagine that instead of interacting with the system, you're also interacting with an assistant that understands who you are, what your role is, and what different business processes you play in your organization.
It can proactively understand the context in terms of what you need, what's important to you, what you need to be alerted about. What actions specifically require your attention. Before you even log in or show your attention to this is proactively already those types of things. Not only is it giving you kind of a sense of the situation, but it's also offering you context. That way you understand why this needs your attention right now. It gives you context, and then it offers you certain alternatives.
The Road Ahead
While Nadella’s prediction may seem ambitious, the evolution of AI and its integration into software development makes it plausible. The future of ERP is not the death of SaaS but the rebirth of app development—faster, smarter, and more adaptable than ever before. By embracing AI-driven code generation, we can build systems that not only meet today’s needs but also anticipate tomorrow’s challenges.
With Lensing, because we've had an opportunity to grow up and build our system with this current AI revolution being able to anticipate what that future vision of the experience and build into our system those AI-Native capabilities that other ERPs have to bolt-on. We've built a lot of those tools into the system to help ourselves build out the ERPs and be able to support the experiences that we want to build out.
Bolting on AI after the fact, you end up with the two different systems instead of one integrated system. It gives you a chance to better not only automate tasks but to allow you to understand the data and the processes that have been implemented in the system, to learn from those things, to be able to then advise you, and guide you moving forward as well.
At Lensing Systems, we’re excited to lead this transformation and invite you to join us on this journey to redefine what’s possible in the ERP landscape.
