WORK
I do not advise on anything I do not do myself. In my own companies I work in marketing and sales every day, against the same performance targets my clients have. What you see here comes from practice, not from theory.
I am an entrepreneur. Alongside consulting I build my own companies, where I own growth, MarTech, sales and everything technical myself: positioning, funnels, CRM, data models, AI systems, code.
That gives me first-hand experience in both B2B and B2C: NativeAI sells security software to enterprises with long sales cycles, German Academy Zurich grows as a B2C platform with learners across several continents, and Tribe is emerging as a consumer experience. 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 full data sovereignty in Switzerland. Co-founded; the company is scaling fast from the EPFL Innovation Park in Lausanne and is a 2026 semi-finalist at >>venture>>, Switzerland's largest startup competition, ranking in the top 10 out of 451 participating startups. Internally, NativeAI runs on an AI business operating system with an AI brain and its own B2B CRM, which I built and operate live. Honestly, it keeps amazing me: what has become possible in the past year alone for accelerating and improving marketing and sales processes borders on the miraculous.
Digital-first language school: a flexible online program plus private 1:1 lessons up to telc and Goethe certification. As Tech Lead and Head of Growth I own the platform, growth and AI features such as AI role-plays. Voted "Best German Course of the Year 2025" by its learners.
Tribe
A relationship experience for individuals: a system that helps people actively nurture their relationships and build real community, in private life as well as in business networking. The conviction behind it: people with more and better relationships live healthier lives. Go-live planned for Q4 2026. The beta test is running: get in touch if you would like to try the application and share your feedback.
ML-as-a-service for marketing operations: churn prediction, predicted customer lifetime value and behavioral predictions as directly usable CRM fields. I built the entire service chain myself, from data integration to model operations. The market was not yet ready for this technology; the approach lives on in my consulting.
Projects are deliberately anonymized. What you see is what was done and how: scope, duration, and whether I worked alone or together with the teams of the company. The details stay with the companies.
Agentic AI for marketing and sales
Business context layer: the brain behind every reply
A central context layer where everything comes together: business and product knowledge, the specific knowledge of the industry niche, and every interaction with every customer, including the replies of all sales people. Every answer to a lead carries the right context, no matter who writes it. When someone is on holiday or leaves the company, the knowledge stays in the system. The entire setup is built to be data protection compliant, and every language model in use runs on infrastructure the company controls itself. Nothing leaks out: neither personal data nor the business secrets and competitive advantages that live in exactly this knowledge.
- Scope
- Project, approx. 4 months, followed by ongoing operations
- Collaboration
- Together with the sales and IT teams of the company
Sales agent for inbound inquiries
An agent that processes inbound customer inquiries: it places each request in the conversation history and the business context, derives a response strategy from it, and writes the draft reply. A human reviews and sends. Response times drop significantly, and no reply goes out without the full context.
- Scope
- Project, approx. 2 months
- Collaboration
- Together with the sales team of the company
Marketing and sales steering agent
An agent that puts a full year of marketing and sales planning to the test: it knows and analyzes the performance of every channel and how marketing and sales work inside the company. From that it derives recommendations on which measures are worth pursuing. On top, it continuously researches new channels and market trends and feeds them into its recommendation. Marketing and sales plans are not simply adopted, they are questioned on data. Throughout the year the agent monitors in real time how the measures develop and flags when a pivot from the original plan makes sense.
- Scope
- Project, approx. 6 months
- Collaboration
- Together with the marketing and sales teams of the company
Dynamic website elements in hours instead of weeks
An agentic flow that turns a plain-language Slack message into an on-brand, technically valid website element: conversion elements, banners and overlays, built on the existing template library and brand guidelines. Once explicitly approved, the agent automatically opens a pull request for the development team. Marketing tests new variants in hours while code review and deployment stay with engineering.
- Scope
- Project, approx. 2 months, from platform evaluation to handover
- Collaboration
- Together with the development and marketing teams of the company
Agentic AI setup reviewed before the build
Review and challenge of a planned agentic AI implementation during the planning phase: structured feedback on the intended setup and recommendations on how the organization should build and operate agentic AI systems. The independent expert view before the build.
- Scope
- Advisory, approx. 4 weeks
- Collaboration
- Together with the AI and MarTech teams of the company
Personalized 1:1 lead nurturing in B2B
Lead nurturing that thinks per contact instead of per segment: the agent guides every lead strategically through the long B2B cycle, decides the next step based on the full context, and personalizes every message down to the individual person. Lead becomes pilot customer, pilot customer becomes customer.
- Scope
- Project, approx. 3 months
- Collaboration
- Together with the sales team of the company
Ad creative production as an agentic flow
An agentic flow that produces complete ad creatives: copy, visuals and every variant per channel and format, consistently following brand guidelines. A human briefs and approves, the flow handles production. One concept turns into every asset of a campaign in a single run.
- Scope
- Project, approx. 3 months
- Collaboration
- Together with the marketing team of the company
Campaign briefings and marketing documents automated
Document generation as process optimization: campaign briefings, factsheets and recurring marketing documents are produced automatically from the connected systems and company knowledge, in consistent structure and quality. What used to be hours of manual work per campaign is now a single step.
- Scope
- Project, approx. 6 weeks
- Collaboration
- Delivered independently, aligned with the marketing team
AI editor: house style rules on demand
A proofreading agent that applies the house spelling and writing rules from a style guide several hundred pages long. The rule corpus lives in a vector database and is retrieved via RAG; the workflow runs in n8n.
- Scope
- Project, approx. 2 months
- Collaboration
- Delivered together with a colleague
CRM and marketing automation
+96% to +233% conversion in experiments
Customer lifecycle marketing as a continuous experimentation program: define segments, design experiences per segment, orchestrate consistently across all channels, test hypotheses, evaluate statistically, iterate. Individual experiments achieved significant conversion uplifts between +96% and +233%; a winback experiment lifted conversion by +117% and revenue by +103%.
- Scope
- Retainer since 2019, ongoing
- Collaboration
- Together with the CRM team of the company
Platform migration without interrupting lifecycle communication
Migration from Salesforce Marketing Cloud to MoEngage inside the running lifecycle marketing: data fields created, experiences redesigned, personalization logic coded (AMPscript to Jinja), email templates built for the target system, automations and multichannel orchestration set up anew.
- Scope
- Migration phase approx. 6 months, within the retainer
- Collaboration
- Together with the CRM team of the company
One template, 19 brands, 19x efficiency
A single email template that serves all 19 brands through dynamization: build once instead of nineteen times. A structural 19x efficiency gain on every setup, hundreds of working hours saved to date, plus use cases a per-brand setup would never have allowed. CloudPages treated the same way. Ongoing operations since: new modules, automations, surveys feeding back into the system, onboarding of new team members.
- Scope
- Build approx. 3 months, ongoing retainer since
- Collaboration
- Build delivered independently, operations together with the CRM team of the company
Newsletters every recipient sees differently
Email modules that show each recipient individually relevant newsletter recommendations: nothing they already subscribe to, prioritized by interest affinity, with a reach-based fallback. Plus a conditional footer that shows app promotion only to contacts without the app, including device-specific store routing.
- Scope
- Retainer, approx. 3 days per month
- Collaboration
- Delivered independently, aligned with the marketing team of the company
Emails that fill themselves with fresh content
Content feeds that automatically pull the latest articles and videos per section from the editorial CMS into the email at send time. Every email shows fresh, topically relevant content, with no manual work per campaign.
- Scope
- Retainer, approx. 3 days per month
- Collaboration
- Delivered independently, interfaces with the CMS team of the company
Winback as an interactive multichannel experience
Evolving the winback flow for churned subscribers from a linear email series into an interactive experience across email, app, push and onsite. Offers and mechanics are validated systematically through A/B experiments instead of gut feeling.
- Scope
- Retainer, approx. 3 days per month
- Collaboration
- Together with the marketing team of the company
Lifecycle journeys rebuilt on data
Onboarding and retention are redesigned on data, and the conversion journey from trial to full subscription is rebuilt. The basis is an email-level performance analysis from which measurable hypotheses are derived and validated with the team before implementation. Built so that a sister product can adopt the journeys one to one.
- Scope
- Project, approx. 2 months: analysis, concept, implementation
- Collaboration
- Together with the product and marketing teams of the company
Recommendation system migrated from Oracle to Adobe
An inbound recommendation system that shows call center agents ML-based recommendations ranked by business impact and purchase probability. Evaluated and documented the legacy Oracle system, analyzed data needs and integrations, separated legacy from essentials, and supported the migration into the Adobe landscape (Adobe Experience Cloud, Journey Optimizer).
- Scope
- Project, approx. 6 months, approx. 2 days per week
- Collaboration
- Embedded in the MarTech team of the company
Hyper-personalization across inbound and outbound
MarTech process and operating model evaluation around the BSI Customer Suite: how must campaigning work in the future so that every customer gets an individually tailored experience? Top-down analysis with expert interviews, mapping of systems, processes and tool landscape, design of trigger-based and real-time campaigning (Kafka) and of the orchestration between call center and outbound in both directions.
- Scope
- Project, approx. 3 months
- Collaboration
- Together with marketing, CRM and IT of the company, with expert interviews across the division
HubSpot from zero to production
HubSpot implementation from the ground up: account setup, data model, workflow automations and an email template with dynamic personalization, making booking data usable for automated lifecycle marketing.
- Scope
- Project, approx. 3 months
- Collaboration
- Delivered independently, aligned with the marketing team
B2B lead scoring: sales works only the best leads
More inbound leads than the sales team could handle: lead scoring in Microsoft Dynamics fed by website behavior (downloads, interactions, product configurator). Below the threshold nurturing, above it handover to sales. Plus role- and interest-based experiences matched to long B2B sales cycles.
- Scope
- Project, approx. 4 months, approx. 1 day per week
- Collaboration
- Together with the marketing and sales teams of the company
Marketing automation for a university library
Evaluation of marketing automation systems against the library's requirements, vendor management, implementation of the chosen tool and building the experiences on top, aimed at growing subscribers and improving the customer experience around the library.
- Scope
- Evaluation approx. 2 months, implementation approx. 3 months
- Collaboration
- Together with the library team
Lifecycle marketing and the technology behind it
Customer lifecycle marketing and the marketing automation technology that carries it: designing customer communication along the lifecycle and working on the technical setup behind it.
- Scope
- Project over several months
- Collaboration
- Together with the CRM team of the company
End-to-end implementations on four more platforms
Always the same chain: evaluate the tool against the use cases, decide, analyze data sources, build the ETL into the CRM or marketing automation system, set up automations and templates, train the team hands-on. Delivered on Mautic, Brevo, ActiveCampaign and Customer.io.
- Scope
- 1 to 3 months per implementation
- Collaboration
- Delivered independently, with training for the specialist team
Data foundations
Data warehouse, daily ETL, free reporting
Evaluated solutions, then built: Azure Synapse as the warehouse, an ETL pipeline in Azure Functions that syncs, cleans and lands the data daily, connected to Power BI. Every report has carried up-to-date booking data ever since.
- Scope
- Project, approx. 14 months
- Collaboration
- Delivered independently
Booking data into the CRM, usable for marketing
ETL process that syncs booking data into HubSpot in the right format, making it usable for email marketing and lifecycle communication. The link between warehouse and marketing tool.
- Scope
- Project, approx. 6 weeks
- Collaboration
- Delivered independently
Website behavior as the data basis for lead scoring
ETL pipeline that brings website behavior (downloads, interactions, configurator usage) into the CRM in a scoring-ready format: the data foundation for lead scoring and role-based nurturing.
- Scope
- Part of the lead scoring project, approx. 4 months
- Collaboration
- Together with the IT and marketing teams of the company
One tracking for paid and owned media
Reviewed, unified and dynamized tracking across all channels so that one structure enables unified reporting across paid and owned: channel performance comparable, attribution correct. Governance set up so that every new channel automatically gets the same structure; the rules live in the tool settings, with no manual tracking to-dos.
- Scope
- Project, approx. 3 months
- Collaboration
- Together with the marketing and analytics teams of the company, including training
- Swisscom
- NZZ
- Allianz
- Suva
- ZKB
- PwC
- CH Media
- Groupe Mutuel
- Mobility
- ETH Zurich
- Dyson
- EWZ
- Switzerland Travel Centre
- Swisspor
- Simpego
- Medi24
- 360° Kinderwunschzentrum
- NativeAI
- German Academy Zurich
- and many more
I am very grateful for the trust and the collaboration with these companies and their teams. On request I am happy to arrange reference calls with people I worked with on these projects.
Beyond scoped projects, there are systems I build again and again: running in my own companies and for clients, refined daily. These are not abstract ideas, they are my daily work. In marketing and sales, almost everything can be accelerated, made more precise, and enriched with context. A new era has begun: the way of working has changed completely, the pace is a different one, and there are possibilities that simply did not exist before. What counts now is having good ideas, implementing them as fast as the technology allows, and reacting just as fast to market developments.
AI setups that empower your team
AI for marketing and sales teams that empowers instead of replaces: full knowledge and full context on brand guidelines, company tone of voice and customer history, for everyone, with more personalized customer interactions and noticeably more speed. Better campaigns, better customer communication, and thanks to a context layer no knowledge is lost when someone leaves the company.
Generate campaigns in one click
Copy, images and every element that belongs to a campaign are created in one workflow. Best practice here is human-in-the-loop: your team collaborates with the agent, for example via Slack, reviews the quality, and generates high-quality output fast in a multi-step process, within defined guardrails such as your brand guidelines and in a data protection compliant way. From idea to finished campaign, without bottlenecks in copy or design.
Generate email flows and journeys
Complete email campaigns and journeys are generated for your target system: structure, content, personalization logic. What used to take weeks now takes hours.
Automate briefings, documents and factsheets
Campaign briefings, factsheets and recurring documents are produced automatically from your systems and your company knowledge: consistent, current, no manual work.
Reporting that interprets itself
Reports build themselves and come with interpretation: raw data turns into an executive report with context and recommendations, not just another dashboard.
Business brain and AI-based CRM
A central brain bundles product, industry and company knowledge plus every customer interaction. Knowledge no longer walks out the door with people, and the system suggests the best next interaction. Especially powerful in B2B. On request entirely data protection compliant on your own infrastructure with self-hosted language models, so that neither personal data nor business secrets ever leave the company.
Sales that never forgets
Your sales team is always perfectly prepared, forgets nothing and never misses a deadline: key stakeholder research, a complete picture of every customer, personalized outreach down to the individual contact. Find, win and nurture better leads without anything slipping through.
Your use case is not on the list? Chances are I have built something similar already. Let's talk about it.
Request a callCONTACT
Let's talk about your project.
If what you see here resonates and you want to run a project that fits my topics, or you need an expert for them: just tell me about your idea.
In that conversation I'll tell you straight away how I would build it, and share my experience with you. Afterwards, take your time to decide whether we work together or you build it yourself. No strings attached.
It starts with getting to know each other: a video call about your initiative or, ideally, a meal together. I am based in Zurich and happy to travel anywhere in German-speaking Switzerland, whether Zug, Lucerne, Basel or Bern. I want to understand who I am dealing with and what your team is planning, so I can support you in the best possible way.
I enjoy the exchange and always love hearing about interesting projects. Just reach out, I'll reply within one business day with a proposed time to talk.
Prefer email? felix.zeeb@geistzeit.ch