Decision science
from two AI tools at work.
How Phronesis and Synthesis actually rank options, merge model outputs, and pressure-test your calls. One field note per week.
-
7 min read →
Building a personal decision journal with AI help
A decision journal is the highest-leverage self-improvement practice most people skip. Here is how to build one, and how AI tools make it practical.
decision-journal habits self-improvement Phronesis -
8 min read →
The chatbot trap — why a conversation isn't research
Chat interfaces optimize for engagement, not output quality. Here is why conversational AI fails at real research, and what to use instead.
synthesis chatbot research Synthesis -
8 min read →
Decision audit trails — what regulators and boards really want
Boards, auditors, and regulators all ask the same question about decisions. Here is what a real audit trail looks like, and how AI decision tools help produce one.
compliance audit governance -
8 min read →
Comparing 3+ SaaS tiers objectively — without the spreadsheet hell
A framework for comparing SaaS pricing tiers that accounts for feature overlap, usage-based surprises, and vendor psychology. Run the compare in 10 minutes, not 10 hours.
saas pricing buying-guide Phronesis -
9 min read →
Hiring decision templates — senior engineer, PM, and designer
Most hiring decisions are gut-feel theater dressed up as analysis. Here are three weighted-criteria templates for senior roles, and how to run them blind.
hiring templates decision-framework Phronesis -
8 min read →
How Synthesis merges Claude, GPT, and Gemini into one answer
A behind-the-scenes look at how Synthesis runs five models in parallel and reconciles their outputs into a single coherent research brief.
synthesis architecture orchestration Synthesis -
9 min read →
How to run a literature review with 5 AI models in parallel
A practical workflow for doing deep literature review using Synthesis. When 5 models in parallel beat 10 hours of solo reading, and how to structure the prompt.
research literature-review synthesis Synthesis -
8 min read →
How to run a vendor RFP without anchoring bias
Most vendor RFPs are anchored by the first vendor to present. Here is a debiased RFP protocol that pairs structured evaluation with blind AI scoring.
procurement rfp vendor-selection Phronesis -
7 min read →
Cross-scoring — why analyst disagreement is a feature, not a bug
Most AI decision tools hide disagreement. Phronesis surfaces it. Here is why the split matters more than the winner, and what to do when analysts disagree.
methodology disagreement scoring Phronesis -
8 min read →
Phronesis vs spreadsheet decision matrices — a showdown
Decision matrices in Google Sheets are the default. Here is what they do well, what they quietly hide, and when a purpose-built tool wins.
decision-matrix spreadsheet phronesis Phronesis -
9 min read →
10 red flags in consultant reports — and how AI audits them
Consultant reports have a predictable pattern of rhetorical cover. Here are 10 red flags and a prompt you can run in Synthesis to audit any deck you receive.
consulting audit synthesis Synthesis -
8 min read →
Decision paralysis — a cognitive science guide for product teams
Why smart teams freeze on important calls, what's happening in the brain, and five interventions that unstick decision-paralyzed product teams.
psychology cognitive-science decision-making Phronesis -
7 min read →
When to use AI orchestration vs a single Claude call
Orchestration is not always the right tool. Here is the decision rule — by question shape, cost, and latency — for when a single call wins.
orchestration single-llm synthesis Synthesis -
10 min read →
HubSpot vs Salesforce vs Pipedrive — an objective breakdown
A vendor-neutral comparison of HubSpot, Salesforce, and Pipedrive. Where each wins, where each loses, and the buyer persona each actually fits.
crm comparison hubspot Phronesis -
9 min read →
Prompt engineering for multi-model research synthesis
Writing prompts for a single LLM is a different skill from writing prompts for an orchestrator. Here is what changes and why.
prompt-engineering orchestration synthesis Synthesis -
8 min read →
A weighted-criteria approach to buying a laptop
Laptop reviews optimize for views, not your use case. A weighted-criteria framework produces better buys in 15 minutes than hours of review-watching.
buying-guide laptops weighted-criteria Phronesis -
8 min read →
The hidden cost of rushed strategic decisions
Rushed strategic decisions look cheap. The rework, reversals, and team churn they create are the expensive part. Here is how to price the whole decision.
strategy decision-cost risk -
9 min read →
How to choose a CRM without letting sales demos bias you
Sales demos are optimized to sell you a CRM, not to help you choose one. Here is a demo-proof framework for comparing HubSpot, Salesforce, Pipedrive, and the rest.
crm buying-guide vendor-selection Phronesis -
9 min read →
Multi-model AI research vs single-LLM — what actually wins
Head-to-head benchmark of multi-model orchestration against a single LLM on 20 research prompts. The orchestrator wins 65%. Here is why.
multi-model orchestration benchmark Synthesis -
8 min read →
Why blind multi-analyst scoring beats consensus voting
Blind AI analyst scoring produces more honest rankings than consensus voting. Here is why disagreement is a feature, and how Phronesis uses it.
decision-science methodology blind-jury Phronesis
Get new field notes in your inbox.
We ship one piece a week on decision science, multi-model AI, and the protocols our tools run under the hood.
Subscribe via Pro →RSS also available · /blog.xml