Blog

Article

How We Built Our Own AI-First CRM

5 minutes
How We Built Our Own AI-First CRM
At Infodation, we asked ourselves a simple question: what if a CRM did more than manage customer data? What if it could actively think along with its users? What if AI wasn't an add-on, but an integral part of the platform from day one?

That became the foundation of our own CRM.

In doing so, we stayed true to our core philosophy: building software that adapts to the business, rather than forcing the business to adapt to the software.

Built for the Future

Technology is evolving faster than ever. A CRM shouldn't just meet today's needs.Tt should be designed to evolve effortlessly over the next five years and beyond.

That's why our platform is built on a modern technology stack that combines scalability, security, and development speed.

The user interface is built with React and TypeScript, enabling rapid feature development while using static type checking to catch errors long before they reach production. Vite provides lightning-fast build times during development, while Tailwind CSS ensures a consistent, maintainable user interface.

Behind the scenes, the platform runs on Supabase, which brings together the database, authentication, and file storage in a single environment. This significantly reduces complexity while ensuring security isn't dependent on multiple disconnected services.

The result is a CRM that's easier to maintain and ready to evolve as business needs change.

AI Is Not Just a Feature

Many software vendors treat AI as an additional feature. We chose a different approach.

AI is embedded into the platform itself, supporting users throughout their daily work.

Our AI layer operates independently of the underlying AI model. Organizations can choose between Infodation's own LiteLLM proxy, Anthropic Claude, or Azure OpenAI. Because the AI layer is fully abstracted, providers can be switched without making changes to the application itself.

This eliminates vendor lock-in while ensuring the platform remains future-proof.

A CRM That Thinks Along with You

AI doesn't just answer questions. It helps users get work done.

Sales representatives can ask questions in plain English or Dutch, such as:

"Which leads require my attention this week?"

Or:

"Draft a follow-up email for this customer."

The AI assistant doesn't just retrieve information. It can also take action. Creating activities, adding notes, or suggesting the next best step happens without users having to navigate through multiple screens.

The CRM evolves from a system of record into a true digital assistant.

Smarter Prioritization with AI

Not every opportunity deserves the same level of attention.

That's why the CRM automatically evaluates every lead using nine different signals, including deal value, win probability, recent customer interactions, pipeline progress, and expected close date.

Each opportunity receives a score between 0 and 100, along with a grade from A to F.

Instead of spending valuable time discussing priorities in sales meetings, the platform continuously performs this analysis in the background.

Automation Beyond Traditional Workflows

Traditional automation relies on predefined rules.

Our CRM goes further.

Every five minutes, the platform scans the entire sales pipeline for warning signals, such as:

  • Leads that haven't been contacted for more than seven days.
  • Overdue activities.
  • High-value opportunities without a scheduled follow-up.
  • Deals approaching their expected close date without recent engagement.

Whenever one of these situations occurs, the system immediately presents a recommended action that the user can approve with a single click.

The platform also responds instantly to key events. When a deal moves to the next stage, is won, or is lost, AI automatically recommends the most logical next action.

In this way, AI becomes an active part of the sales process rather than a standalone feature.

Designed for Integration

A CRM never operates in isolation.

That's why the platform follows an API-first architecture.

All core business capabilities are exposed through secure REST APIs, while integrations communicate through event-driven mechanisms that synchronize data in real time without creating tight dependencies between systems.

Today, the CRM already integrates with Microsoft Azure Active Directory and Microsoft Outlook Calendar.

Users sign in through Single Sign-On using their existing Infodation account. When they sign out, they're automatically logged out of the CRM as well.

Activities created in the CRM are automatically synchronized with the salesperson's Outlook calendar, with all sensitive authentication securely handled server-side.

Because of this modular architecture, future integrations with ERP systems, email platforms, or other business applications can be added without redesigning the existing product.

Built the Way We Build Software

Perhaps the most remarkable aspect of this CRM isn't the technology. It's how it came to life.

Just four hours after our initial conversation with the business, the first prototype was up and running. Three days later, the first working version was already being used in practice. Less than a month later, version two was released, incorporating dozens of improvements driven directly by user feedback.

This isn't an exception. It's how we build software at Infodation.

We don't believe in spending months defining requirements before users see anything. We deliver early, gather feedback, and improve continuously.

We don't measure success by the number of features we ship, but by the value they create.

We Build for Ourselves First

Our CRM is used every day by our own sales team. That means we experience the same challenges as our customers.

Every frustration becomes an opportunity to improve the platform. Every new requirement can be validated immediately. Every AI capability is tested in real-world scenarios before becoming available to others.

As a result, the product evolves through everyday use, not assumptions.

What's Next?

Over the coming years, our focus will shift even further toward intelligent assistance.

Think predictive analytics, more advanced lead scoring, churn prediction, intelligent workflow automation, and an AI assistant that increasingly understands what users need before they ask.

By building software that understands what's happening, proactively supports users, and helps them make better decisions, we're creating what we believe is the future of CRM.

That's what an AI-first CRM means to us.

Applied AI, without the theatre

Want to see how this works in your organisation?

We help teams turn AI concepts into working workflows, usable tools, and measurable operational gains.

Start a 14-day AI sprint

Further reading

Continue reading

A few more articles worth reading after this one.