Discover
Audit
€1,900
2 weeks
One end-to-end process mapped with your team, a scored list of automation candidates, a GDPR and data-residency assessment, and fixed-price quotes for the three strongest candidates.
Start with an AuditMosaik Automations starts with the work your team does today. We connect the systems you already use, add AI only where it helps, and keep a person in charge of every decision that matters.
Works with the systems you already use
Most businesses don’t need one giant AI project. They need a handful of recurring jobs done reliably, like getting an enquiry into the CRM, checking an invoice or building the monthly report. We build each piece, connect it to the systems you keep, and make sure someone owns it.
Individual pieces. One shared system.Structured handoffs, approvals and a log of every run.
The Mosaik method
Six stages, from MAP to OPERATE. It’s a sequence of decisions, not six projects to buy: you can start later in the sequence when the earlier answers are already clear.
01 · Map
We sit with the people who do the work and map one workflow end to end: where requests arrive, what gets copied, which decision waits and where exceptions cause rework. You get a shared picture of the workflow, systems, owners and open questions.
| Task | Hours / month | Fit |
|---|---|---|
| Re-entering quote requests in the CRM | 22 | High |
| Typing invoices into DATEV | 16 | High |
| Chasing approvals by email | 14 | Medium |
| Assembling the monthly report | 9 | Medium |
| Searching for internal answers | 4 | Low |
02 · Improve
Before anything is automated, we remove steps that shouldn’t exist, settle which system holds the truth, and decide who approves what. You sign off the target workflow before we build.
03 · Connect
We connect the systems you keep, with checked fields, limited retries, alerts and clear access rules. Information moves once, and you can trace where it went.
04 · Build
Where your existing tools aren’t enough, we build the missing piece: an internal tool or an AI-assisted step. It ships with tests, documentation and defined behaviour for when something goes wrong.
05 · Measure
After launch we compare against the starting point: handling time, errors and quality instead of gut feeling. Then we decide together whether to adjust, roll out or stop.
06 · Operate
Every system gets a named owner, monitoring, a clear path for incidents and an agreed way to handle changes. A regular report shows runs, hand-offs to your team and errors.
Guardrails
An automation that embarrasses you in front of a customer costs more than it saves. Every system we deliver has limits your team can see, and a person who owns it.
How we handle dataAnything that touches money or customers waits for a named person. The approval is part of the design, not a footnote.
A working system of record remains the source of truth. We connect viable software instead of replacing it by default.
Error handling, limited retries and a log of each run, so you can see what happened and why.
Documentation, a runbook and training at handover. Your accounts, knowledge and exit path stay with you.
How we help
Five situations we see again and again. Each one starts with a question about your workflow, not with a tool.
Nobody reliably sees who takes over and which detail is still missing.
Where we’d start: bring intake, ownership and missing details into one clear flow, then check a structured handoff into your existing system.
Document, spreadsheet and business system hold different versions. The mismatch shows up too late.
Where we’d start: agree one source of truth and shared field rules. A controlled transfer can replace the double entry if access and data quality allow.
An approval sits in a message. The deputy, the criteria and the status are unclear.
Where we’d start: settle who approves what, up to which limit, and who stands in. A shared review space can record each decision and its reason.
Several exports are merged by hand. Later, nobody can explain where a figure came from.
Where we’d start: define the figures and their sources first. A traceable data path can make the report repeatable, with gaps kept visible.
Internal knowledge is scattered, and what’s current and who may see it varies.
Where we’d start: sort sources, validity and access. Then test a search or assistant that cites its sources. Sometimes better filing is enough.
Services
Start with a bounded problem. We sell a decision before a large build, and agree scope and acceptance criteria in writing before work starts. You never need all four.
Every project includes
Discover
€1,900
2 weeks
One end-to-end process mapped with your team, a scored list of automation candidates, a GDPR and data-residency assessment, and fixed-price quotes for the three strongest candidates.
Start with an AuditConnect
€4,900
3 weeks
One process live in your environment, not a prototype. One API connection, human approval where money or customers are involved, a log of every run, a runbook and a training session.
Scope a SprintBuild
€8,000–15,000
scoped from the Audit
An app, dashboard or integration across up to three systems, with test coverage, documentation and two training sessions. Quoted from the Audit, never blind.
Plan a buildOperate
€800–1,500
a month · 3-month minimum
Monitoring with defined response times, a set number of changes each month, and a quarterly report on runs, hours saved and errors.
Ask about careNo VAT is charged: Mosaik Automations is a small business under § 19 of the German VAT Act (UStG).
Each service can stand on its own. An Audit may end with the advice to keep a workflow as it is. Map yours
Who you’ll work with
I’m Kareem, a developer trained at 42 Wolfsburg with a background in sales at several start-ups. I started Mosaik because I kept seeing teams and entrepreneurs lose hours every week to work a machine could do in seconds: copying data, chasing leads, answering the same questions. Mosaik builds the automations that give that time back.
Short answers to what business owners ask us first. Anything else, just ask us by email.
Not necessarily. First we check whether the work can be simplified or an existing feature is enough. Clear rules stay rule-based; AI is considered where interpreting text or documents brings a benefit you can check.
A working system of record should stay. Whether a connection is possible and sensible depends on data quality, access and the interfaces your software supports, among other things.
No. We need one person who knows how the process runs today and can make decisions about it. Every stage names an owner on your side and one at Mosaik.
We document which data moves where, who can access it and what gets logged, before anything goes live. The Audit includes a GDPR and data-residency assessment for each automation candidate, so sensitive data is considered before a build is chosen.
From a bounded problem, clarified prerequisites and written acceptance criteria. Changes to the scope and ongoing operation get their own agreements.
No. The services describe different outcomes. An analysis can end with the decision to keep a workflow as it is, or to improve it without any new technology.
Every workflow has error handling, limited retries and a log of each run. When a run fails or the data is unclear, the case goes to a named person instead of passing through unchecked. Ongoing monitoring is part of Ongoing care.
Quick Map · about 3 minutes
Eight short questions about one recurring workflow. You see your result straight away, then decide whether to send it to us.