01
01 / On-Premise
In your data center.
Container, Kubernetes or VM — we deliver, you operate. Air-gapped possible. No data path to the internet.
for: Banks, insurers, manufacturing, critical infrastructure, government.
// products / time-tracking-ai-agent
Time tracking that finally talks back.
The TimeTracking AI Agent connects to your existing time-tracking software — including legacy systems with no API love — and gives your team a language they already speak: a chat.
Made by createIF Labs · Hosted in Germany
— problem
The problem
Your team hates time tracking. The software is old, the form is ugly, the reporting is late. Nobody logs in time, everyone groans at month-end. You know how it goes.
— solution
The solution
We connect to what you already have — Atoss, Personio, ZEP, Tempo, BMD, Datev or your own build — and put an AI agent in front of it. People log time in a sentence: "Today 8h project Falcon, 30min lunch break." Done.
features
What makes the agent tick.
-
Chat-first
Log hours in natural language — German, English, Swiss German. On Slack, Teams, or your own web widget.
-
Widgets in chat
When clicking is faster than typing, the agent surfaces quick actions: tables, slot pickers, calendars. Form follows function.
-
Legacy-ready
SOAP, REST, CSV imports, RPA as a last bastion — we get in where others give up.
-
Audit-grade
Every entry has a full trail. Compliance officers sleep well. So do auditors.
-
GDPR by default
No data in US clouds. No data sent to OpenAI. Optional fully air-gapped on-premise.
-
Multilingual
Your team in Stuttgart logs in German, your team in Lisbon in Portuguese. One agent, many languages.
// deployment-modes
Three ways to host it.
You decide where your data lives — and which model sees it.
02
02 / LLM hosted in Germany
We host — in Germany.
Managed service on European infrastructure. Frankfurt, Nuremberg or Hetzner Helsinki. SLA, backups, updates included.
for: Mid-market that doesn't want to operate it but stays GDPR-compliant.
03
03 / BYOLLM
Bring your own LLM.
You already have an internal Llama-3 or an Azure OpenAI tenant? We plug in. Model-agnostic, no vendor lock-in.
for: Enterprises with their own AI infrastructure.
// interaction / chat
Language instead of forms.
The agent understands context: which projects are active, what hourly rates apply, which cost centers are allowed. It asks back when unsure. It corrects itself when corrected. And it stays quiet when everything's clear.
Also via Slack, Teams or your own web widget.
user · 01
Today 8h on project Falcon, 30min lunch break.
tta
Logged: 7.5h Falcon (8:00–16:30, lunch 12:00–12:30). Confirm?
user · 02
Thursday was home office, full day on customer 4711.
tta
Thursday, 8h, customer 4711, location home office. Right?
user · 03
How much did I log this week on Falcon?
tta
22.5h on Falcon this week so far. Plan target: 32h. Want me to suggest Thu/Fri?
// interaction / widgets
When clicking is faster.
Some tasks are quicker with a click than a sentence. The agent recognizes that and sends structured widgets straight into the chat: slot pickers, weekly overviews, bulk corrections, approvals.
Weekly view
A compact table right in the chat. Gaps are immediately visible. Click a cell = new entry.
Slot picker
Pick a time window instead of typing it. Snap to 15-minute grid. Mobile-friendly.
Bulk operations
"Move all Falcon hours from Wed to Thu." The agent shows a diff, you confirm with one click.
// integrations
Works with your software.
We integrate with anything that has an interface — and a few things that don't.
- 01 Atoss
- 02 Personio
- 03 ZEP
- 04 Tempo (Atlassian)
- 05 BMD
- 06 DATEV
- 07 SAP S/4HANA
- 08 Microsoft Dynamics
- 09 Custom Legacy
Want to see this in your environment?
We schedule 60 minutes of demo with a test instance, connected to a read-only sample of your software. No slides.
Book a demo → 30 min · no commitment