Inconvo: Open-Source Chat-with-Data Platform

In a software landscape increasingly defined by artificial intelligence, the ability for applications to “talk” to their own data has shifted from novelty to necessity. Inconvo, a startup founded in 2023 and based in Limerick, Ireland, has emerged as a compelling answer to this demand. Built as an open-source platform for creating chat-with-data agents, Inconvo enables companies to integrate intelligent, conversational access to production databases directly into customer-facing applications. Rather than treating AI as a superficial chatbot layer, the company positions its technology as infrastructure — a reliable bridge between natural language and real business data.

Founded by Eoghan Mulcahy and Liam Mulcahy, Inconvo is a compact but ambitious team operating with the focus typical of early-stage innovators. Having participated in the Summer 2023 startup batch, the company has maintained active development while refining a product aimed squarely at engineering teams under pressure to implement AI features quickly and safely. Its core promise is deceptively simple: make it easy to deploy trustworthy data agents that can answer real user questions without exposing companies to the risks typically associated with large language models interacting with live systems.

As organizations race to incorporate AI-driven experiences into their products, Inconvo’s approach highlights a broader shift. Instead of building flashy demos, companies now need production-grade solutions that can scale, protect sensitive information, and remain accurate over time. Inconvo’s growing attention stems from addressing precisely these needs.

Why Is Building Chat-with-Data Functionality So Difficult?

Although conversational interfaces powered by large language models have captured the public imagination, translating those capabilities into enterprise-grade tools is far from straightforward. Product teams attempting to implement chat-with-data features often begin with a seemingly logical strategy: allow an AI model to generate SQL queries that retrieve information from a database based on user questions. In theory, this approach promises flexibility and speed. In practice, it introduces a cascade of risks.

One of the most significant challenges is data safety. Production databases contain sensitive and often regulated information. A model generating queries autonomously may retrieve the wrong rows, expose confidential fields, or accidentally cross boundaries between different customers or tenants. Even a single mistake can create legal exposure and erode user trust.

Business correctness presents another layer of complexity. Metrics definitions, permitted filters, join relationships, and formatting rules are rarely simple. They evolve over time and often live across multiple documentation sources. Encoding these rules into prompts alone can lead to inconsistencies, causing AI-generated answers to drift away from the company’s actual logic.

Finally, there is the issue of boundaries. Because models can produce arbitrary SQL, enforcing restrictions after the fact becomes a fragile exercise. Teams frequently end up building additional validation layers, monitoring tools, and manual safeguards — effectively constructing a second system just to keep the first one under control.

The result is a paradox: while demos of chat-with-data functionality can be assembled quickly, deploying them responsibly to real customers becomes an engineering challenge that many teams underestimate. Inconvo’s founders recognized this gap early and designed their platform specifically to address it.

How Does Inconvo Rethink the Problem?

Instead of trying to restrain AI systems after they generate risky queries, Inconvo flips the paradigm by constraining them from the beginning. The platform introduces a structured framework that defines exactly what data agents are allowed to access and how they can interact with it. This approach transforms AI from a free-form generator into a guided system operating within clearly defined boundaries.

At the heart of this methodology lies a semantic model that serves as a contract between the application and the data agent. Companies can specify approved tables, columns, metrics, filters, and relationships in a single authoritative layer. By centralizing these definitions, Inconvo eliminates the fragmentation that typically occurs when rules are scattered across prompts and codebases.

Rather than producing raw SQL queries, the AI generates constrained query plans expressed in structured formats. These plans are validated against the semantic model before execution, ensuring compliance with business rules and data access policies. Only after passing validation does the system translate the plan into an executable query.

This architecture shifts control back to deterministic code rather than probabilistic prompts. It also reduces the need for reactive guardrails because safety is embedded into the workflow itself. Inconvo’s philosophy suggests that reliable AI systems should be designed more like secure APIs than experimental chatbots.

What Features Make Inconvo Production-Ready?

For a tool aimed at customer-facing applications, reliability is paramount. Inconvo’s platform includes several capabilities designed to move beyond experimentation into real-world deployment.

Safe query execution ensures that data agents operate within approved parameters, preventing unauthorized access or accidental exposure of sensitive information. Automatic permission enforcement applies tenant scoping and field restrictions consistently, eliminating the risk of cross-customer data leaks. Stateful conversations allow agents to maintain context across interactions, enabling more natural dialogues while still adhering to safety rules.

Observability is another key component. Engineering teams can monitor how agents interact with data, track usage patterns, and identify potential issues before they escalate. This transparency is essential for organizations deploying AI features at scale, where silent failures or inaccuracies can quickly impact user experience.

Inconvo also offers a simple API endpoint that developers can integrate directly into their applications. This design choice reflects the founders’ emphasis on practicality. Instead of requiring teams to rebuild their infrastructure, the platform acts as a modular layer that can be plugged into existing systems.

Who Benefits Most from Inconvo’s Approach?

While any organization managing large datasets could theoretically use Inconvo, the platform is particularly relevant for companies building customer-facing products. Software-as-a-service providers, analytics platforms, fintech applications, and enterprise dashboards all share a common need: enabling users to access complex data through intuitive interfaces.

For these companies, chat-with-data functionality can transform user experience. Instead of navigating menus or writing queries, customers can ask questions in plain language and receive accurate answers instantly. However, because these interactions occur directly with production systems, reliability and security cannot be compromised.

Inconvo’s open-source model also broadens its appeal. Developers can inspect the underlying mechanisms, customize behavior, and contribute improvements. This transparency fosters trust and encourages adoption among engineering teams wary of black-box solutions.

Startups racing to differentiate their products through AI features may find particular value in the platform’s speed of deployment. By providing out-of-the-box infrastructure for permissions, validation, and monitoring, Inconvo allows teams to focus on user experience rather than reinventing safety mechanisms.

How Does Open Source Shape Inconvo’s Strategy?

Choosing an open-source foundation reflects both philosophical and strategic considerations. In the AI ecosystem, transparency is increasingly important as organizations seek assurance that systems interacting with sensitive data behave predictably. By making its platform open, Inconvo invites scrutiny and collaboration, positioning itself as a trustworthy alternative to proprietary solutions.

Open source also accelerates innovation. Developers experimenting with data agents can build on Inconvo’s framework, extending it to new use cases and industries. This community-driven development model can help the startup scale its influence far beyond what a two-person team could achieve alone.

From a business perspective, the approach aligns with a common pattern in infrastructure startups: offering a free core platform while potentially monetizing advanced features, hosted services, or enterprise support. Although the company’s long-term commercialization strategy remains evolving, the open-source base establishes a strong foundation for growth.

What Role Do the Founders Play in the Vision?

Eoghan Mulcahy and Liam Mulcahy bring a focused vision shaped by hands-on experience with the challenges of deploying AI in production environments. Their decision to target a specific pain point — safe conversational access to data — reflects an understanding that the next phase of AI adoption will hinge on reliability rather than novelty.

Operating from Limerick, Ireland, the founders represent a growing trend of globally distributed innovation. Advances in remote collaboration and open-source ecosystems allow startups outside traditional tech hubs to compete on equal footing. Inconvo’s participation in a prominent startup batch further underscores its ambition to operate on an international stage.

The founders’ technical orientation is evident in the product’s architecture. Instead of prioritizing marketing hype, they emphasize deterministic safeguards, semantic modeling, and developer-friendly integration. This pragmatic mindset resonates with engineering teams seeking tools they can trust.

What Does Inconvo Reveal About the Future of AI Applications?

Inconvo’s emergence highlights a broader shift in how organizations think about artificial intelligence. Early excitement focused on conversational interfaces and generative capabilities. The next phase is about operationalizing those capabilities safely within mission-critical systems.

As AI becomes embedded in everyday software, the distinction between experimental features and core infrastructure will blur. Tools like Inconvo suggest that successful platforms will be those that treat AI as part of the system architecture rather than an add-on. Safety, governance, and observability will become standard requirements rather than optional enhancements.

The startup also illustrates the growing importance of domain-specific solutions. Instead of building general-purpose chatbots, companies are creating specialized agents tailored to particular workflows — in this case, querying production data. This specialization enables deeper integration and higher reliability.

Can Inconvo Become a Foundational Layer for Data-Driven Products?

Whether Inconvo evolves into a widely adopted standard remains to be seen, but its approach addresses a problem that is unlikely to disappear. As organizations accumulate ever-larger datasets, the ability to interact with that information conversationally will become a competitive advantage. Platforms that make this interaction safe and scalable could become essential components of modern software stacks.

If the company continues refining its technology while cultivating an active developer community, it may position itself as a foundational layer for data-driven applications. In doing so, Inconvo would not only simplify how products implement AI but also reshape expectations for what conversational interfaces can accomplish.

In a world where software increasingly speaks the language of its users, the ability to ensure those conversations are accurate, secure, and meaningful may define the next generation of successful technology companies. Inconvo’s early work suggests that the future of chat-with-data will belong not to the flashiest demos, but to the systems that make AI dependable enough to trust with real decisions.