How do you introduce AI into a company?

Start with processes, not tools. At omnius we first go through where your people lose the most time, for example on phone calls, email requests, receipts or searching for information in Slack. Then we pick the automations with the biggest payoff, connect them to the tools you already use, and deploy them. That is exactly how we did it in our own company.

How much can a company save with AI automation?

At omnius, automations save an hour a day for four people in management and project leadership. At average Slovak salaries in 2026, that is roughly €3,700 a month and €41,000 a year. The exact savings depend on the number of people, their salaries and the processes being automated.

How does the AI phone system with call recording and transcription work?

The phone system is virtual, routes calls to the right people or to regular phone numbers, and has its own app. At the start of a call it informs the caller about recording, and the call continues only if they confirm consent by pressing 1. If they do not, the system tells them how else to reach us and ends the call. The recording is transcribed and AI works with the transcript.

Do we have to replace the tools we use, such as Slack, to adopt AI?

No. We recommend keeping the tools that work for your people and connecting them. We tried to replace Slack with our own solution and it was a dead end. That is why we build AI automations on top of existing tools such as Slack, email, calendars or accounting software.

Where is our data stored?

omniHub can run in the cloud with data in the EU, or on-premise on your own server. If data must not leave the European Union, we deploy language models through EU-hosted SovereignAI.

What kind of companies benefit from this automation?

Companies with roughly 10 to 200 people, where a lot of information flows through phone calls, emails and chat and management cannot keep track of it all. We are deploying the same approach in several Slovak companies today.

How omnius uses AI: a phone system, Slack digests and omniHub working as one

Challenge

Not long ago we had the same problems as most growing companies. The only difference was that we solve them for clients every day and had overlooked them at home for a long time.

Phone calls. People answered calls the traditional way and then retyped what was said by hand. When nobody picked up, a request was sometimes lost entirely.
CRM. We used Odoo. It was complicated and disconnected from our other tools.
Slack. For 20 people and dozens of channels, Slack is a great tool, but a lot gets lost in it too. Decisions got buried in threads, and promises like "I'll send it on Friday" were forgotten.

The people who felt it most were those whose time is most expensive: project managers and management.

Our approach

We put our own problem on the table like any other client project. The goal was not to deploy as much AI as possible, but to connect our tools so that nothing gets lost and there is a single source of truth.

omniHub instead of Odoo

Odoo was replaced by omniHub, our own CRM, with AI agents built in. Why your own software beats an off-the-shelf box is covered in a separate article. More processes are moving into omniHub step by step:

The receipt and invoice tamer is integrated directly in omniHub. Receipts from the field are processed without retyping.
Meeting transcripts from Plaud flow into omniHub through an MCP server, so the record of a meeting sits with the client and project it belongs to.
New modules are added wherever manual work costs us the most time.

Email requests as tickets

Customer requests also arrive by email. Each one is automatically turned into a ticket in omniHub and assigned straight away. The whole company knows about a new request, not just the person whose inbox it landed in.

AI phone system

The phone system is virtual and needs no hardware. It routes a call to the right person or to a regular phone number, and it has its own app.

Every call starts with a recording notice. The caller has to confirm consent by pressing 1. If they do not, the system tells them where to find us on the web and by email, and ends the call. Consent to recording is not hidden in the terms; it is explicit.

The recording is transcribed and AI works with the transcript. It does not talk to the caller; a person still handles the call. The phone system makes sure no request is lost and that there is a record to come back to at any time. The next step is creating requests in omniHub automatically, straight from the transcript.

Slack digests and reminders

Mentat, our company memory with AI agents, runs on top of Slack. It prepares digests and sends them by email:

daily on tasks and what was discussed,
weekly on commitments and items with a weekly deadline,
monthly as an overview for management.

The digest also flags any ticket nobody picked up within its deadline. It also sends automatic reminders about promises. When someone in the company commits to something, the system reminds them so it does not slip.

What did not work

At first we tried to replace Slack with our own solution. It was a dead end. Slack is a good tool, people like using it, and it made no sense for us to spend effort replacing it. Instead, we connect it into the whole. We give clients the same advice: don't replace tools that work for your people, connect them.

Results

Management and project managers have a complete overview without reading dozens of channels
Promises and tasks no longer get forgotten; reminders arrive on their own
Every customer request that comes in by phone has a transcript and a single source of truth
A missed call no longer means a lost request
Every email request has an assigned owner, and if nobody picks it up in time, the digest flags it
Receipts, meetings and client communication all come together in one system
Client satisfaction went up, because nothing falls through the cracks

The automations save an hour a day for four people: the project manager, CTO, COO and CEO, the people whose time is most expensive in the company.

What an hour a day is worth

We used average Slovak market salaries for 2026 and the full cost of employment including employer contributions (gross salary × 1.362). With 21 working days a month, it comes out as:

Role	Average gross salary	Cost per hour	Saved per month
CEO	€6,794	€55	€1,157
CTO	€6,292	€51	€1,071
COO	€6,000	€49	€1,022
Project manager	€2,906	€24	€495
Total			≈ €3,700

Over a year, 231 working days after holidays, that is 924 hours and roughly €41,000. That time now goes to clients, sales and growing the company instead of reading channels and chasing tasks.

The calculation uses public averages (Platy.sk, Jooble), not our actual salaries. There is no public average for a COO, so we used a conservative €6,000. In a company with higher salaries or a larger management team, the savings are proportionally higher. And there is even more room: employees typically spend 2 to 3 hours a day on email and searching for information (McKinsey, IDC), so automation pays off beyond management too.

What's next

We extend automation step by step, always where manual work costs us the most time. This is on our roadmap:

AI will talk to callers. The phone system will do more than transcribe. It will answer common questions, find out what the caller needs, and route more complex matters to the right person.
Ticket triage. AI will assess every new ticket, assign it a priority level and send it to the person who should handle it, so urgent matters don't get lost among routine ones.
Tickets straight from calls. A phone call transcript will become a ready ticket in omniHub without anyone typing it up.
Draft replies to requests. AI will prepare a draft reply to an email request, and a person just reviews and sends it.
Tenders and price quotes in omniHub. Tracking public tenders and preparing price quotes, which we currently do outside omniHub, will move into the same system.

We will write about each step on our blog, including what doesn't work the first time.

We take care of it

We don't just hand over a tool. We take care of it for you.

We don't deliver software and walk away. We take the whole automation on ourselves, from the first analysis to day-to-day operation.

We map how you work

We go through your processes, people and tools and find where time is lost and errors happen.

We adapt it to your processes

We design the solution around how you work, not your company around a tool.

We automate

AI takes over routine work: phone calls, requests, receipts, digests and reminders.

We look after it long term

We service it, handle changes and add further automations as the company grows.

Automations take over work you would otherwise have to hire more people for, and remove the errors that come from manual retyping and forgetting. Part of the value can be measured in hours saved, but the peace of mind that nothing gets lost cannot.

What we do for companies that want to adopt AI

Everything described in this case study, we also deploy for clients. We don't sell licences or an off-the-shelf box, and we are not a training company either. We can run a training, but we guide the company through the whole journey: from a pilot, through deployment into operation, to long-term care. We want to be a long-term partner, not a one-off delivery.

Consulting and process audit. Together we find where your people lose the most time and where AI pays off most. More about our AI solutions.
omniHub with AI agents. A CRM you own, with no per-user fees. AI agents in it process requests, receipts and meeting transcripts.
AI phone system as a service. A virtual phone system with call routing, explicit consent to recording, and AI call transcription.
Automatic digests and reminders. Daily, weekly and monthly overviews of company communication, plus reminders of commitments, so nothing gets lost.
Integrations with your existing tools. We connect email, calendars, Slack, invoicing and the accounting systems you already use.
AI with data in the EU. If data must not leave the European Union, we deploy language models through SovereignAI.

We are now deploying the same approach at other companies, and case studies from those projects are on the way. If you want to know how much automation would save in your company, get in touch via contact and we will prepare the calculation for your situation.