Michigan United

We pointed the FC26 engine at 52 real Engagious.com focus groups. Here's what we found.

FC26 is the first bot created to monitor and analyze social media for the full story behind election threats in real time. This is naturally a hard pill to swallow — the obvious fear is a bot that just makes things up.

Let's address that head on. We're going to look at how the bot reads and analyzes text. Instead of thousands of social posts an hour, we pointed the FC26 engine at 52 real focus-group transcripts from Engagious.com's YouTube channel. We used Watson NLU to pull results you probably didn't expect to get from focus-group data. What makes that believable, not just possible, is grounded generation — it constrains the model's output to the real data it was actually given, which is what keeps a bot from inventing anything.

Below is the result — a real report, from real data, not a mockup. FC26 itself is still ahead of us: what's left to fund is running this same engine on an actual election, not a focus group. Pick a state.

What you're looking at

This isn't a summary — it's a full strategic report. Every topic that came up — the economy, immigration, foreign policy, and so on, what this report calls a domain — gets scored on its own, not blended into one average. That matters: a candidate can be completely off-limits on one issue and wide open on another, and a single number would hide that difference entirely. On top of that: real emotion scores pulled straight from Watson NLU (not a guessed mood), and message lines quoted verbatim from a real, poll-tested persuasion methodology (WFP), not invented on the spot. The impact: this is the exact level of analysis FC26 would run on live disinformation instead of a focus-group transcript — the difference is the data source, not the rigor.

Help Make FC26 Real
See how →

We are going out of our way to prove that FC26 technology works.

Something that other fundraising appeals won't do.

FC26 is still theoretical — built, architected, never yet run on a real election. Its deliverable isn't a report you'll personally read, it's a voter population less exposed to coordinated disinformation in 2026.

To immediately prove its ability to analyze text for emotion, context, intent and more, we created a stand-alone tool. We nicknamed this technology “Frasier” after the TV character.

Frasier is three of FC26's eight bots — Emotion & Persuasion, Reporting, Adversarial Shadow — the exact pipeline that produced the report above.

The technology part is no longer theoretical. What's left to fund is the harder half: helping fund a Meltwater or Brandwatch feed, getting FC26 up and ready to fight for the 2026 general election.

Frasier

Persuasion-readiness scoring, moral-map analysis, campaign-ready reports.

FC26

8-bot pipeline architecture, fully built and deployed — proven only against synthetic data so far. A real feed is the ask.

IBM Watson NLU extraction Persuasion-readiness scoring Grounded counter-drafting Deterministic grounding — never a guessed fact

Already built

  • 8-bot scoring pipelinethreat, coordination, and narrative-cluster detection — runs end to end today, on synthetic data
  • 4 deployed appsoperator console, public transparency, government intelligence, persuasion terrain — deployed, waiting on a real feed
  • Calibration sandbox100k synthetic posts, full claim journeys, zero Watson cost — proof the pipeline works, not yet proof it's protected anyone
  • A real governance frameworkclaim-level, not person-level labeling — drafted before a single real post is scored

Needs funding

  • Live monitoring, not just calibrationa real feed and the NLU volume it costs, sustained through 2026
  • Human review capacityrecruiting and maintaining a network of volunteer monitors behind every "escalate to human" the system is designed to surface
Not a surveillance tool. FC26 labels claims, not people — the same discipline that keeps Frasier from ever inventing a number.

You just watched the method work. FC26 is the same engine, aimed at protecting a real election instead of testing a message — and it's still waiting on the funding to do that for real. Partners who support it get direct access to Frasier itself.

Let's talk →

or try Frasier yourself →