When is a prediction reliable enough to use?
When can we trust a prediction? I develop research workflows that connect qualitative evidence with quantitative data, expose contradictions and assess the evidence behind forecasts.
From client workflows to independent research
For a private-equity firm, I developed a workshop and n8n workflows supporting alpha analysis. The workflow combined qualitative transcripts with quantitative inputs, highlighting changes through early-warning signals and supporting preparation for calls. Client identities, internal methods and scoring logic remain confidential.
How much evidence does a prediction need?
I examine where accounts converge, where they conflict and which interests shape them. Source provenance, independent corroboration, recency and counterarguments matter more than volume alone. The working collection contains around 24,000 derived hypotheses; these are not 24,000 individually verified reports.
Three formats
The Prediction Report examines forecasts and uncertainty. The AI Bubble Tracker follows expectations and risks in the AI market. The Daily Checker reviews changes and predefined triggers. The research pipeline is ongoing; outputs remain working analyses requiring verification before use.
Read and review
Email me if you would like to read a current report and provide feedback on the ongoing research and its three reporting formats.
Commission qualitative research
I investigate market and company questions through targeted conversations, trade-fair visits and public-source research, including work at Internorga. Each assignment starts with a clear question and distinguishes evidence, interpretation and open questions.
Commissioned research: possible formats
| Question | Method | Output |
|---|---|---|
| What do potential customers need? | Targeted qualitative interviews | Needs and open questions |
| How is a market changing? | Trade-fair conversations and public sources | Comparison of perspectives |
| How strong is a claim? | Source comparison and verification | Evidence and uncertainty |
Proposed formats; scope and access to interview partners are agreed in advance.