Kandidaten-Kompass
An AI briefing tool that digests long German-language political interviews (FAZ, Süddeutsche, taz) and party press releases into plain-language position cards, so voters can quickly grasp where Berlin state-election candidates stand on housing, transit, schools, and debt.
Berlin residents and German citizens following state elections who don't have time to read 4,000-word interview transcripts
- One-paragraph AI summaries of every major candidate interview, with sources cited
- Side-by-side stance comparison on 8–10 local issues (Mietspiegel, Schulden, CSD, innere Sicherheit) that the user can weight by importance
- Local election calendar with registration deadlines, mail-in ballot how-tos, and push reminders
- A 'what changed since last week' digest tracking new candidate statements and policy reversals
Berlin's Abgeordnetenhauswahl is weeks away and a surprise CDU candidacy swap (Evers replacing Wegner) is dominating the news cycle; voters are hungry for shortcuts through dense German political coverage.
Berlin Abgeordnetenhauswahl confirmed for 20 Sep 2026 with tight 5-party race (CDU 20.5% / AfD 18% / Grüne 17.1% / Linke 16.1% / SPD 14.2%) and active Wegner/Evers candidacy swap story driving traffic to candidate comparison pieces.Abgeordnetenhauswahl Berlin 2026: CDU 20,5 % vor AfD 18,0 % ↗Berlin-Wahl 2026: Die Spitzenkandidaten im Schnell-Check ↗
Market is crowded: official state-backed Wahl-O-Mat, academic Wahl-Kompass (Kersting/Münster, used since 2009), Wahlomat.co, wahl-omat.geo.io, and Wahlkompass.online all serve Berlin voters — only the AI-digestion-of-long-interviews angle is novel, and incumbents could trivially add it.Wahl-O-Mat 2026 ↗Wahlkompass - Kommunalwahl-Tool für Parteien & Kandidierende ↗Politikwissenschaftler veröffentlichen Wahl-Kompass zur Bundestagswahl ↗
All major German election comparison tools (Wahl-O-Mat, Wahl-Kompass) are free, and the dominant one is government-funded by Bundeszentrale für politische Bildung; willingness to pay for one-off state-election briefings in Germany is unproven and likely weak.Wahl-O-Mat 2026 ↗Wahl-Kompass online - Behörden Spiegel ↗
Usage is spiky and pegged to elections; could be repurposed across 16 Bundesländer cycles and Bundestagswahl, but each event is a 4-6 week intense-use window followed by dormancy — no evergreen retention driver.
German-language LLMs (GPT-4-class) handle long-form political text well and transcription is solved; main friction is licensing/ingestion from paywalled FAZ/SZ plus the editorial work of turning interviews into balanced, non-misleading position cards (high reputational risk if quotes are misattributed).