Turn a real problem in your business into a working AI-assisted tool.
A six-week, live-online course for owners and operational decision-makers in solo and micro businesses. No programming background required, just a willingness to type, test, and examine AI-generated work carefully. You bring a genuine operational problem and leave with a tested prototype, the skills to build the next one yourself, and an honest read on whether it's ready to use.
Built for people who run the business, not just use the software
This course is for owners and operational decision-makers at solo and micro businesses (companies with up to 10 employees) who have a genuine operational problem they understand well enough to scope and solve themselves. Each live cohort itself is kept small, 3 to 6 participants, so there's real time for your specific problem.
A good fit if
- You have a real, recurring operational problem, and a fallback one
- You're comfortable using a keyboard and testing results, even without a programming background
- You're willing to work with made-up or non-sensitive data only
- You can commit to six 3-hour live sessions plus guided work between them
- You'll set up ChatGPT, Claude, GitHub, and Streamlit access before Session 1
Not a fit if
- You want a production system built for you, not to learn how to build one
- Your project requires live sensitive, confidential, or personal data
- You're looking for ongoing implementation or private consulting support
By the end, you can defend the decision, not just the demo
You'll turn a genuine operational problem into a tested AI-assisted proof of concept, explain how it works, evaluate its limitations and risks, and make an evidence-based call on whether and how to operate it. A well-reasoned decision not to deploy counts as a successful outcome when deployment would be unsafe or premature.
- Scope a real problem into a bounded, buildable project
- Turn domain knowledge into clear, testable requirements
- Use ChatGPT and Claude efficiently, and read the code they generate instead of trusting it blindly
- Call APIs, handle errors, and keep credentials out of your code
- Deploy, version, and roll back a working application
- Test against acceptance criteria, and evaluate security, privacy, and operational risk before deciding to operate it
See all 14 assessed competencies
The kind of thing this approach builds
These come from Jeff’s 1:1 applied AI advisory work, not the course pilot, which hasn’t run its first cohort yet. They show the range of everyday problems the same method can solve.
A prioritized daily task list
Pulls together a business inbox, a personal inbox, a calendar, and a CRM into one list of what actually needs attention today, and flags client relationships that have gone quiet.
A solar power monitor
A Raspberry Pi connected to a home solar system logs how much power it’s using and sends an alert the moment something goes offline.
A multi-account calendar and email assistant
Brings several calendars and multiple email accounts across providers into one place and prioritizes what needs to happen each day.
Six live sessions. One real project. A steady build toward a working prototype.
Each week combines live teaching with structured guided work you complete on your own between sessions. Every task has written instructions, an expected result, and a time allocation, so you always know what "done" looks like.
01Problem and responsible use
Select your primary and fallback problem, define users and workflow, compare tools, and set ground rules for prompts, cost, and risk.
Live 3h · Guided 1.5h02Requirements and acceptance
Map the process, separate fact from assumption, and write requirements and acceptance criteria you can actually test against.
Live 3h · Guided 2h03Architecture, code, and versions
Design the components, learn to read AI-generated code, and set up a private repository with a known-good version and a rollback plan.
Live 3h · Guided 2hA one-on-one checklist review of your architecture, scope, safety, and integration readiness before you build.
04APIs, deployment, and rollback
Call public and authenticated APIs, handle errors and secrets properly, deploy in Streamlit, and practice restoring a known-good version.
Live 3h · Guided 2h05Debugging, testing, and risk
Diagnose and fix a real error, test the normal, boundary, and failure cases, and review security, privacy, and peer feedback.
Live 3h · Guided 2h06Demonstration and operation
Demonstrate and explain your proof of concept, decide whether it's ready, and document maintenance, rollback, and stop conditions.
Live 3h · Guided 2h18 hours of live instruction, 11.5 hours of guided project work, and one 0.5-hour private review: 30 documented hours in total, delivered over six weekly sessions (up to eight weeks if public holidays require it).
Everyone builds the same starting point: a client inquiry and follow-up tracker
The shared application
Using made-up, safe example data, everyone builds the same small tool: something that reads a new client inquiry, drafts a helpful summary using AI, and reminds you when to follow up, automatically skipping public holidays. Along the way, you connect it to an AI service, save your work with a full history you can step back through, and turn it into a simple web page you can click through. If something breaks, you'll practice undoing the change and getting back to a version that worked.
The starter gives you structure, not the answer. What you build, test, and can explain is what gets assessed.
Your own project
Alongside the common build, you apply the same method to your own genuine business problem: you define it, build it, test it, and explain it. A project that turns out to be unsuitable is narrowed, switched to your fallback problem, or completed through the common exercise instead.
Either way, the required result is a working proof of concept for one complete workflow, not a production system.
The same tools you can keep using afterward
You don't need to know what any of these mean yet, that's what the course teaches. No proprietary lock-in either: everything here is something you can keep a subscription to and keep building with.
ChatGPT & Claude
The AI assistants that do the actual writing and coding, with you directing them.
Python
The programming language the AI writes in. You'll learn to read it, not memorize it.
Google Colab
A notebook in your browser for trying code out. Nothing to install.
GitHub
Where your work is saved, with a full history you can undo mistakes from.
Streamlit
Turns your code into a simple, clickable web page anyone can use.
An AI service connection
The link that lets your tool talk to ChatGPT or Claude automatically.
A public holiday lookup
A free, no-login data source your tool checks automatically.
Google Meet
Where the live sessions happen.
Before Session 1, you'll need ChatGPT and Claude access (including a low-cost paid subscription to at least one), plus Google, GitHub, and Streamlit accounts. A mandatory 60-minute group onboarding session confirms everything works before the course starts.
We teach the method. You make the decisions.
This is a group Weiterbildung (structured professional development), not individual consulting. You choose and scope your problem, write the requirements, generate and review the code, test it, and decide whether it's ready. HammNet does not build or operate participant-specific production systems.
Within the course
- Teaching repeatable methods and demonstrating common tasks
- Asking diagnostic questions and reviewing your evidence
- Explaining errors and assessing competence
Outside the course
- Building your specific solution for you
- Configuring production systems or importing live business data
- Operating your solution or providing private implementation services
What stays out of the classroom
- Real sensitive, confidential, or personal data (made-up or non-sensitive data only)
- Automated legal, medical, employment, credit, or other high-impact decisions
- Financial transactions
- Unsupervised external communications
- Scraping or automation that could violate a platform's terms
- Access to production systems where failure could materially harm the business or someone else
Sessions are recorded for the cohort only, kept restricted, and deleted 30 days after the course ends. You'll accept a cohort confidentiality commitment, and camera use is optional.
What showing up actually looks like
Cohort size
3 to 6 participants per cohort, kept small on purpose.
Sessions
Six weekly live-online sessions, 3 hours each, 18 hours total.
Guided work
11.5 hours between sessions, plus one 0.5-hour private review.
Schedule
Normally six weeks, up to eight if public holidays require it.
Language
Taught in English. Certificate and transcript issued in English and German.
Delivery
Google Meet, a private GitHub repository, and a controlled Google Drive folder.
Attendance
You can miss one of six sessions if you complete the make-up work within seven days.
Assessment
Competency-based, not graded. All 14 competencies are required for a completion certificate; an attendance and assessment record is issued otherwise.
Support
Questions go through GitHub Discussions, checked twice a week, continuing for 30 days after the course ends.
Taught by Jeff Hamm
Jeff Hamm is the course's sole instructor, with more than 25 years of technical experience and a teaching record going back to 2004.
- Teaching
- Assistant Professor and Adjunct Lecturer at NTNU (2011 to 2025); current Cellebrite technical instructor; IACIS instructional service (2004 to 2009)
- Applied AI
- DBA study focused on AI and machine learning, in progress; completed the University of Texas at Austin's AI and Machine Learning Business Applications program in 2026
- Credentials
- M.S. in Cybersecurity and Information Assurance; B.S. in Information Technology; Cellebrite Certified Operator and Physical Analyst for Inseyets
- Published
- Co-author of Digital Forensics (Wiley, 2017)
What working with Jeff looks like
Nicola Blanchard Wine Export and Distribution Strategist, MainzWorking with Jeff has been genuinely transformative for my business. He helped me critically evaluate how I work, identify where my time was being lost, and put practical AI tools in place to improve my performance.
As a solo entrepreneur, it is imperative that I remain flexible, move quickly, and make the most of every minute. With Jeff's guidance, I have automated a number of workflows that were previously taking up substantial amounts of time. I am now free to focus on creative, high-value work and to move my business forward with more clarity and positivity.
I would strongly recommend Jeff to any entrepreneur looking to work smarter, not harder. I will be booking regular sessions to make sure my systems continue to evolve and support the way I want to work.
From Jeff's 1:1 applied AI advisory work, not the course pilot, which hasn't run its first cohort yet. Shared here as an example of the kind of thinking this course teaches you to do yourself.
Pilot cohort pricing
+ VAT per participant
- All six live sessions and structured guided project work
- The mandatory pre-course onboarding session
- Your private project review
- A digitally signed certificate and competency transcript on successful completion
- Access to GitHub Discussions support for 30 days after the course ends
This is the pilot cohort price. Required AI subscriptions and API usage are maintained and paid separately by each participant.
Register your interest
This course is still forming its first pilot cohort, and seats will be limited to 3 to 6 participants. Tell us a bit about your operational problem and your general availability (weekday morning, afternoon, or early evening), and we'll keep you posted as details firm up. No application, no commitment, just a note that you're interested.
If you're looking for custom development or work with live sensitive data, HammNet's consulting services may be a better fit than this course.