AI automation in Freiburg

AI automation combines classic process automation with language models. Rules handle what is clear-cut, and AI takes on what people previously had to read, sort or write: emails, enquiries, documents, leads. Kamil Kascha plans and builds these workflows from his base in Freiburg, Germany, working across DACH and globally remotely only.

At a glance

For

B2B companies in the Mittelstand, SaaS and industry

Based in

Freiburg im Breisgau, Germany

Remote

DACH and globally, remote only

Focus

Marketing, sales, CRM, data and content

Practice

+730% inbound leads and over 300% more closed deals at Wiferion

In most companies the most expensive manual work sits between systems: data copied from the website into the CRM, enquiries sorted by hand, reports rebuilt every month. Classic automation fails where information is unstructured. That is exactly what language models can now read and classify.

At Wiferion I built the full path from first touch to closed deal: website, campaigns, AI-supported qualification and CRM workflows as one system. At HRworks I integrated marketing automation with Salesforce. About Freiburg runs on my own autopilot with scrapers for more than 15 sources, duplicate checks, a publish gate and automatic posting.

So you do not get someone who only recommends tools, but someone who runs these workflows himself and knows where they break in daily use.

Process analysis

I map workflows, measure time spent and sources of error, and show which steps are worth it. The result is a prioritised automation roadmap.

Lead-to-deal automation

An end-to-end funnel with lead enrichment, scoring, routing, automatic follow-ups and a clean handover to sales.

CRM automation

Workflows in Pipedrive and Salesforce, data maintenance, duplicate checks, attribution and KPIs tied to revenue.

Marketing automation

Newsletters, campaigns, content pipelines and reporting that run without recurring manual work.

AI agents and LLM workflows

Agents for research, writing, classification and analysis, as checked processes with clear limits rather than loose chat windows.

Document processing

Read emails, PDFs and forms automatically, classify and summarise them, and transfer them into the right system.

API integration

Systems talk to each other, so people stop copying data. Website, CRM, shop, newsletter and internal tools are connected.

Operations, monitoring, handover

Error handling, logs, audits and documentation, so your team can run the solution itself.

Workflows that recur often, touch many systems and are easy to check suit best. Typical examples:

Sales

Qualify incoming enquiries, enrich company data, create CRM records, route to the right person, prepare quotes.

Marketing

Research and draft content, run campaigns, create reports, monitor rankings and AI visibility.

Customer service

Triage tickets, suggest replies from documentation, close standard cases.

Back office

Read invoices and receipts, file emails, write minutes, reconcile data between systems.

Content and data

Collect sources automatically, detect duplicates, publish or submit for approval.

Management reporting

Combine KPIs from several systems and deliver them weekly with commentary.

The project runs remotely in five steps. After each step you decide whether to continue.

First conversation

We clarify process, bottleneck and systems involved. No obligation.

Map the process

I document the current workflow and measure time, volume and errors.

Target process and business case

We define data model, checkpoints, KPIs and the return.

Build and test

Implementation with real data, first alongside the old workflow.

Handover

Monitoring, documentation and a briefing. Your team then runs the system.

The three approaches solve different problems and are often combined. I pick the simplest one that works reliably.

I am not tied to a vendor. Depending on the task, workflows are built with tools such as n8n, Make or Zapier, with custom code in Python and PHP, or directly via your systems’ APIs. Language models come from OpenAI, Anthropic, Google or open models, depending on requirements.

My CRM experience comes mainly from Pipedrive and Salesforce, plus WordPress, WooCommerce and Shopware. Anything with an open API can be connected.

An automation is only as good as its checkpoints. For sensitive steps a person still decides (human in the loop), for example before anything goes to customers. Logs show every action, audits find deviations, and an error raises an alert rather than writing wrong data to the CRM.

For personal data, data flows, vendors and storage locations are defined up front in line with GDPR. Where AI works in direct contact with people, this is disclosed under the EU AI Act. This does not replace legal advice.

It pays off when the cost of building and running it is below the working time saved or the additional revenue. Frequency, time per case and cost of errors all count. A workflow that occurs once a quarter often stays better as manual work, and I will tell you so.

Companies look for the same work under different titles. If you are filling one of these roles, we are talking about the same thing, as a project or a permanent position:

Forward Deployed Engineer

Also searched for as Forward Deployment Engineer: working close to the department and building and rolling out automations in the real systems.

AI Solution Architect

Designing target process, data model, interfaces and checkpoints before anything is built.

AI Automation Engineer

Implementing, testing and shipping workflows, AI agents and API integrations.

AI Transformation Officer

Steering automation as a programme: priorities, KPIs, governance and rollout with the team.

Marketing Automation and RevOps

Owning the lead-to-deal process, CRM and attribution as one system.

I am based in Freiburg im Breisgau. I support automation projects across DACH and globally remotely only, by video call and with secured access to your systems. That is no drawback for this work, because it takes place inside your systems anyway.

At Wiferion: +730% inbound leads and over 300% more closed deals between 2022 and 2025. At HRworks: Salesforce-integrated marketing automation and +30% MQLs in twelve months. At About Freiburg: my own autopilot, calendar up from about 50 to 125 events.