Krafthaus prioritizes. decide.fyi verifies execution constraints.
Pilot baseline shown. Live metrics replace these after kickoff.
Krafthaus selects priority -> decide.fyi verifies policy -> action is routed.
Example: a $15 vs $20 pricing-anchor decision selected in Krafthaus is checked in decide.fyi for execution constraints before dispatch.
Two signal lanes in. Deterministic runtime verdict out. KPI outcome feeds the next sprint brief.
Signal Quality Score (SQS) rates evidence trustworthiness from four checks: reliability, freshness, noise, and consistency. Method details.
Krafthaus uses polling rigor, internet signal quality controls, and deterministic runtime execution in one decision loop.
Screeners, dedupe, source reliability, freshness, and noise checks keep both signal lanes usable.
Decision frameworks are used as constraints for interpretation, not as votes.
Decision Pack plus audit trace move directly into decide.fyi for a yes/no execution call.
Polling tells you what people think. Krafthaus tells you what to ship this sprint.
One constrained pricing question in, one runtime verdict out.
One active brief. One Decision Pack. One deterministic call.
/api/decision-packet-validate
{
"pack_type": "pricing_anchor",
"signal_lanes": ["human_polling", "internet_sentiment"],
"lane_weights": { "human_polling": 0.64, "internet_sentiment": 0.36 },
"signal_quality": {
"label": "Signal Quality Score",
"components": ["reliability", "freshness", "noise", "consistency"]
},
"candidate_option_id": "opt_a",
"lead_margin_pp": 24.2,
"gate_passed": true,
"runtime_provider": "decidefyi"
}Primary output is machine-readable Decision Pack + audit trace. PDF summary is optional.
Define one KPI-linked question, run one deterministic call, and feed results into the next sprint.
Frame one binary decision tied to one KPI and lock 2-5 options.
Capture targeted human polling and internet sentiment evidence.
Score quality and set adaptive weights by use case.
Package ranked options, confidence spread, Signal Quality Score, and guardrails.
Send packet to decide.fyi Decision API for yes/no gating and runtime call.
Track KPI impact and feed the result into the next brief.
Targeted polling plus source-scored internet sentiment
Adaptive blend based on reliability, freshness, and volatility.
Deterministic yes/no call with packet context
Outcome snapshot reused in the next sprint decision cycle
Built for teams that need one high-consequence call resolved quickly and executed with traceability.
Not a broad polling panel. This is decision operations for one high-impact call per sprint.
Example scorecard for launch narrative. Replace with live pilot telemetry before public claims.
Time from brief open to runtime dispatch.
Hours until one option establishes a stable lead.
Spread between yes/no gate before handoff.
Packets acted on in the same sprint window.
| Indicator | Sample baseline | Current | Delta |
|---|---|---|---|
| Decision cycle time | 12 days | 8 days | -33% |
| Time to first signal | 72 hours | 19 hours | -74% |
| Confidence separation | 9% | 42% | +33 pts |
| Same-sprint action rate | 22% | 61% | +39 pts |
Example only. Scope: pricing anchor pack. Window: 7-day cycle. Result shown for packet format reference.
Values are loaded via /api/site-case. Replace with your pilot feed. Decision Pack + audit trace is the system-of-record; PDF is optional.
Submit one pricing decision question. We return scope and kickoff options in one business day.
Your pilot request is in queue. Reference ID: pending.
We will reply with scope and kickoff options within one business day.
Operator-ready workflow with decision pack, runtime call, and outcome readout.
Define KPI owner, one binary question, and locked option set.
Collect human + internet signals, score quality, and compute confidence spread.
Ship packet to decide.fyi, execute the call, and record KPI outcome for the next cycle.