About me
A passion for MarTech, SalesTech and AI.

Felix Zeeb
Independent implementation partner for AI in marketing and sales, Zurich
Since 2019 I have been building marketing and sales systems for companies from SMEs to large groups, today mostly AI agents. Through my company Geistzeit I work for teams across Switzerland and the DACH region.
- Focus
- AI agents, CRM and data systems for marketing and sales, in your teams' tools and aligned with your IT.How I build and test AI agents
- Offer
- AI agents from my kit: set up at a fixed price, live in about 30 days, then continuously tested in operation. For larger work also as a project or as part of your team.How it works
- Clients
- NZZ
- CH Media
- Swisscom
- Groupe Mutuel
- Switzerland Travel Centre
- ETH Zurich
- Swisspor
- Allianz
- PwC
- Mobility
- Dyson
- EWZ
- Simpego
- Medi24
- 360° Kinderwunschzentrum
- Data protection and security
- With NativeAI I am co-building a Swiss AI security startup and sell a security product there myself. I know the compliance requirements of regulated industries first-hand.Where data and models run
- Handover and backup
- The agent runs on your side or on a server in Switzerland. Code, runbook and documentation are with you from day one, so your team can take over at any time. Should I drop out, a partner I have worked with for years steps in.How operations work
- Experience
- 16 years in marketing and sales, entrepreneur since 2020: growth, marketing, sales and technology in my own companies.
- Education
- M.Sc. Marketing & Data Science, University of Vienna; B.Sc. Economics and Business, University of Hohenheim
You talk directly to the person who builds your project. Questions, adjustments and decisions are settled directly. Offer, contract and invoice go through my Swiss company Geistzeit.
I work with your IT, not around it: I align architecture, security and interfaces with them, and sign-off stays with them. I run operations for a monthly fee and document everything so your team can take it over at any time.
So that AI does not get stuck in the pilot.
These are projects I have already delivered, for clients and in my own companies. Concept, build and operations from one hand, until the workflow runs in daily work, in the team's tools and approved by people.
Agentic AI for marketing and sales
Sales agent for inbound requests
The agent reads every request, knows the history and writes the reply draft.
- Placed in conversation history and business context
- Reply strategy plus a finished draft
- A human reviews and sends: much faster response
- Scope
- About 2 months
- Collaboration
- With the sales team
Ad production as an agentic flow
One concept turns into every asset of a campaign in a single run.
- Copy, visuals and every format per channel
- Always on brand guidelines
- A human briefs and approves
- Scope
- About 3 months
- Collaboration
- With the marketing team
Business context layer: the company knowledge behind every answer
Every reply to a lead knows the full knowledge of the company, no matter who writes it.
- Product, industry and customer knowledge in one place
- Knowledge stays when people leave
- Language models on own infrastructure, no data leaves
- Scope
- About 4 months, then ongoing operations
- Collaboration
- With sales and IT
Steering agent for marketing and sales
An agent challenges the annual marketing and sales plan with data.
- Analyzes the performance of every channel
- Recommends which measures are worth it
- Monitors in real time and flags when to pivot
- Scope
- About 6 months
- Collaboration
- With marketing and sales
Personalized 1:1 lead nurturing in B2B
Nurturing per contact instead of per segment, through the whole B2B cycle.
- The agent picks the next step from context
- Every message personal
- Lead becomes pilot customer, pilot becomes customer
- Scope
- About 3 months
- Collaboration
- With the sales team
Campaign briefings and marketing documents automated
Briefings and factsheets are produced automatically from systems and company knowledge.
- Always the same structure and quality
- Hours of busywork become one step
- Scope
- About 6 weeks
- Collaboration
- Independently, with the marketing team
AI copy editor: editorial rules at the push of a button
An agent applies a style guide of several hundred pages.
- Rules in a vector database, retrieved via RAG
- Workflow in n8n
- Scope
- About 2 months
- Collaboration
- With a colleague
Agentic AI setup reviewed before it was built
Independent review of a planned agentic AI implementation, before the build.
- Structured feedback on the planned setup
- Recommendations for building and running it
- Scope
- Advisory, about 4 weeks
- Collaboration
- With the AI and MarTech team
One agent for web chat, WhatsApp and email
One agent for every channel that replies in the writing style of each team member.
- Rules apply instantly across all channels
- Reply drafts right inside the email thread
- Every reply sent improves examples and rules
- Scope
- In stages, ongoing operations
- Collaboration
- With the team that improves it
Website elements in hours instead of weeks
A Slack message turns into an on-brand website element.
- Banners, overlays and conversion elements
- Based on the template library and brand guidelines
- After approval, a pull request is created automatically
- Scope
- About 2 months
- Collaboration
- With development and marketing
AI role-plays as a lead magnet
Instead of a whitepaper, a role-play in which prospects practice real everyday scenes.
- Three agents: counterpart, goal check, feedback
- Sign-up with double opt-in
- In B2B, for example as sales or negotiation training
- Scope
- Prototype in one week, ongoing expansion
- Collaboration
- With the management team
Agents that measurably improve with every conversation
An agent is not finished at go-live: every conversation becomes traceable and a test case.
- Every conversation as a session in Langfuse
- The domain team flags mistakes, they become tests
- Regression tests with synthetic users before every change
- Scope
- Setup about 2 weeks, then part of every agent project
- Collaboration
- With the domain team
Agent-led assessments that qualify leads
An agent runs the assessment as a conversation instead of a rigid questionnaire.
- B2C: adaptive placement test
- B2B: AI Project Planner that turns an idea into a project plan
- Result after contact details, then a talk with a human
- Scope
- B2C live, B2B: the AI Project Planner on this website
- Collaboration
- With the domain experts
Website migrated in days, SEO and GEO run agentically
From WordPress to React in a few days, then an agent optimizes the site every month.
- Prerendered, multilingual, old links still work
- The agent improves copy and links from search data
- Approval by a human, plus A/B tests
- Scope
- Migration in days, optimization ongoing
- Collaboration
- Independently
Alongside consulting I build my own companies and own growth, sales and technology there, B2B and B2C alike. What I recommend to clients, I live myself.

Swiss-hosted security layer for the safe enterprise use of LLMs: protection against prompt injection, jailbreaks and data leakage, with data sovereignty in Switzerland. Co-founded, based at the EPFL Innovation Park in Lausanne, 2026 semi-finalist at >>venture>> (top 10 of 451 startups). Internally, NativeAI runs on an AI operating system with its own B2B CRM that I built and operate.
Digital-first language school: a flexible online programme plus private 1:1 lessons up to the telc or Goethe certificate. We built the entire course platform ourselves, and with AI we can finally implement the ideas that have piled up over the years, such as AI role-plays. Our learners love the course: learning German has never been this easy. As founder, I built up the platform, growth and AI features.
Tribe Meguri
I see the loss of community and closeness as one of the big problems of our time. I first looked for a solution for myself; at the request of friends and acquaintances it became a service: an app that helps people actively nurture their relationships, in private life and in business networking. Go-live planned for Q4 2026, the beta test is running.
ML-as-a-service for marketing operations: churn prediction and customer lifetime value as directly usable CRM fields, the entire service chain built myself. The market was not ready yet; the approach lives on in my consulting.
On site in the Greater Zurich area
Greater Zurich area
On site whenever needed.
Rest of German-speaking Switzerland
For example Basel, Bern, Lucerne, Zug, St. Gallen or Schaffhausen: on site by arrangement.
Otherwise remote
In my clients' infrastructure and tools.
Languages
German as my native language, English at business level.
Geistzeit is my consulting company and my lab for experiments. When a business experiment works, I move it into a company of its own.
Three developments strike me as the big problems of our time: we are losing closeness, as families get smaller and communities erode. AI and robotics take over work faster than our societies can rethink taxes and the distribution of value. And the climate is getting more extreme.
Tribe Meguri is my first contribution to this: technology that brings people together.
CONTACT
Let's talk about your project.
Tell me about your idea: on a video call from anywhere in German-speaking Switzerland, or over lunch in the Zurich area. I'll tell you straight away how I would build it.
I reply within one working day with a proposed time.
Prefer email? felix.zeeb@geistzeit.ch

