Phos Labs Behavioral Science API
v3.0.0• Phos Labs
Help your users convert, engage, and decide better. 9 use-case endpoints that apply evidence-based behavioral science to real agent problems: diagnose why users drop off, identify user personas, optimize messages, audit for ethics, design behavior change interventions, predict churn, optimize pricing, elicit user preferences, and prioritize interventions. Pay per use with USDC via x402. Backed by 47 evidence nodes, 6 behavioral personas, 10 scored interventions, and peer-reviewed research.
Skills
-
Diagnose Conversion Problem
Why aren't users converting? Send your funnel or journey data and get a behavioral root-cause analysis with friction points ranked by impact, plus prioritized fixes with expected lift percentages.
conversionfunnelfrictionoptimizationdropoutabandonment -
Identify User Type
What kind of user is this? Send behavioral signals (clicks, time spent, comparisons, purchases) and get a persona classification: Maximizer, Satisficer, Loss-Averse Protector, Status Seeker, Value Optimizer, or Mission-Driven. Includes recommended approach and what to avoid.
personasegmentationpersonalizationuser-typeclassification -
Elicit User Preferences
What does this user actually want? Get adaptive trade-off questions (conjoint analysis) that reveal preference weights across attributes like price, quality, speed, brand. Returns questions to ask plus scoring methodology.
preferencesconjointtrade-offspersonalizationrecommendation -
Optimize Communication
How should I talk to this user? Send a message draft and get it rewritten using behavioral science — framing, social proof, loss aversion, identity framing. Returns optimized copy with principles cited and expected impact.
messagingcopyframingpersuasioncommunicationnotification -
Audit for Ethics
Is this nudge or recommendation ethical? Detects dark patterns, manipulation, and autonomy violations. Returns an ethics assessment with specific issues and fixes. Uses Thaler & Sunstein's nudge-vs-manipulation distinction.
ethicsdark-patternsmanipulationauditcompliancetrust -
Design Behavior Change
Help me change this behavior. Full intervention design: COM-B diagnosis (why aren't they doing it?), EAST-scored nudge design, implementation plan, commitment mechanisms, and success metrics. Evidence-based with effect sizes.
behavior-changeinterventionnudgeengagementhabitactivation -
Predict Churn Risk
Is this user about to leave? Send behavioral signals and get a churn risk score with behavioral diagnosis (why they're disengaging) plus ranked retention interventions. Uses loss aversion, commitment devices, and re-engagement strategies.
churnretentionengagementre-engagementloyaltyattrition -
Optimize Pricing
What price or offer should I present? Get a behaviorally-optimized pricing strategy using anchoring, decoy effect, loss framing, scarcity, and price partitioning. Returns tier structure, presentation recommendations, and expected impact.
pricinganchoringdecoymonetizationconversionrevenue -
Prioritize Interventions
Which interventions should I do first? Send a list of candidate interventions and get them ICE-scored with behavioral bias adjustments (planning fallacy, overconfidence, status quo bias). Returns a ranked list categorized as Quick Win, Big Bet, Maybe, or Time Sink.
prioritizationICE-scoreinterventionroadmapdecision-makingdebiasing
Integration
import asyncio
from a2a_registry import AsyncRegistry
async def main():
async with AsyncRegistry() as registry:
agent = await registry.get_by_id("5041fd95-3c09-49a4-b3fb-17e58ed41641")
client = await agent.async_connect()
print(f"Connected to {agent.name}")
asyncio.run(main())