We are IGT - a thirty-something-person software and system integration company in Bratislava. Like most SMEs of our size, the company's operations were spread across six or seven different tools. One for attendance. Another for project tracking. A third for invoicing. A spreadsheet for shift planning. Another spreadsheet for benefits. Each tool had its own login, its own data model, its own export format.

To answer a question like "how much did we invoice this client in April?" or "how is this project tracking against budget?" someone had to know which tool held the answer, get into it, find the right view, and reconcile against another tool.

We saw the same pattern in every Slovak SME we worked with.

Challenge

Fragmentation isn't a software problem in the conventional sense. Each individual tool works. Each one solves the thing it was bought to solve.

The problem is structural. When operational data lives in six different tools:

Nobody owns the answer to a cross-functional question. Asking "are we profitable on this project?" requires pulling time from one system, costs from another, invoicing status from a third. Someone has to do that reconciliation by hand. Most of the time, nobody does.

Decisions get made on instinct. Owners and managers can sense how the business is doing, but they can't see it. The data exists. It's just not in a form anyone can act on.

Adding a tool is easy. Removing one is much harder. Every new system layers another login, another export, another reconciliation step. Operational debt compounds the same way technical debt does.

We were watching it happen at IGT. We were watching it happen at every Slovak SME we worked with. The pattern wasn't accidental - it's the natural shape an SME's operations end up in if nobody actively prevents it.

Mid-market ERPs exist to prevent it. But for a thirty-person company, they're priced for organisations five times larger, and configured for processes that don't match how small businesses actually run.

So we started building.

Solution

Core 3 is an operations platform built around one principle: one place, one data model, one set of truths. Every operational function a small or mid-sized business needs runs as a module on the same platform, sharing the same underlying data.

Informal development started in 2024. Formal development kicked off on 1 January 2025. The platform went live for IGT's own operations the same year. Today, the company runs entirely on Core 3.

Seven modules are live:

Attendance: planning, events, meeting-room occupancy, and the reports payroll actually needs at month-end.

Project management: agile and waterfall methodologies, Kanban boards, project reporting. Every hour logged here flows directly into the Attendance module - no second entry, no reconciliation.

Ticketing: built for the developer side. Find a bug, raise a ticket, route it.

Finance: cost centres, automated reading of POs and invoices. The system extracts line items from documents directly.

Benefits: order lunch, rate it, see what colleagues ordered, sign up for a massage. Built for the way IGT actually works, not for what HR theory says they should do.

Shifts: for businesses that run on shift patterns: planning, visibility, overlay against attendance.

Migration: the platform's quiet workhorse. Every new tenant arrives with data scattered across the same six or seven tools we're trying to replace. The Migration module is programmable to pull data from any legacy source and reshape it into Core 3's data structure. Onboarding becomes a configuration exercise, not a data-entry exercise.

The AI layer sits across all of them

Core 3 ships with an integrated LLM that has access to the platform's full data - and only to it. The LLM operates inside the tenant's environment with no outbound internet connection. No data leaves the boundary of the customer's installation. No prompts logged in someone else's cloud. No accidental training on someone else's payroll.

What that enables is the answer to the question we started with:

How much did we invoice company XY in 04/25?
How many people are going to be in the office today?
Which project is at risk of being delayed?

Ask Core 3. That's the layer that makes the data useful at the speed business actually moves at.

Delivery Approach

Three decisions shaped how Core 3 got built - and each cost something to make.

Dogfood first, sell second. Most product companies build something they'd like to sell and then look for customers willing to try it. Core 3 reversed the order. Before the platform was offered to anyone else, we moved our own operations onto it - payroll-affecting attendance, real customer invoicing, live project tracking. Every bug found by us is a bug a paying customer doesn't find. Every feature that survives a year of our own operations is a feature that's earned the right to ship.

The Migration module is a feature, not a service. Most SaaS products treat data migration as the customer's problem. Core 3 treats it as the product's problem. This is the difference between onboarding a Core 3 customer in weeks and onboarding them in months - and it's a direct response to what kept SMEs trapped in their fragmented stacks in the first place: nobody wanted to do the migration work.

The data doesn't leave the building. A standard SaaS LLM integration sends prompts and context to an external API. That's a non-starter for the SMEs Core 3 is built for: finance data, payroll data, customer commercials, and HR records are not data they'd consent to send to a third party. Core 3 ships with a sandboxed LLM that runs inside the tenant's environment, with full data access but no outbound internet. Slightly less powerful than calling a frontier model. Significantly more deployable in a real business.

Result

Twelve-plus months of internal operations. Core 3 isn't a roadmap with screenshots. It's the system that runs IGT itself - attendance, projects, finance, shifts, benefits, the lot. Every change to the platform goes through us before it reaches anyone else.

Two Slovak SMEs in active production rollout. The first is a custom carpentry business making kitchens and furniture - multi-machine shop floor, shift patterns, warehouse operations. They're migrating department by department, and the first measurable impact came in project management: hours that had been quietly absorbed into the old tooling are now visible, billable, and accounted for.

The second is a wholesale business with a large warehouse. The rollout there is moving slower for a telling reason: they have fewer existing systems to migrate from. We're building their operational data layer from the start rather than restructuring it from legacy sources.

3–5 legacy tools replaced per deployment. Each Core 3 rollout retires the same fragmented stack we replaced internally: separate attendance, separate project tracking, separate invoicing, separate shift planning. The economics follow - one platform subscription beats five point-tool licences on every line of the spreadsheet that matters.

Manual admin time has dropped at IGT. Pilot SMEs are tracking similar trajectories on operating costs. These are conservative figures - they reflect what's measurable today, not what's modelled. The full case will sharpen as the pilots run longer.

And there's a quieter result that matters more than any single metric: managers can ask the platform questions and get answers. Which project is at risk? Who's in the office today? How much did we invoice this client last quarter? For SMEs, this is the thing mid-market ERPs were supposed to deliver and didn't.

Core 3 makes the data that already exists actually accessible to the people who need to make decisions on it.

That's the bet we're testing.