The leader track 🤝 · for research, insights & strategy leads · 6 x 45 min · standards, not tools
Judgement mode: what genuinely accelerates in the research pipeline and what only appears to, the evidence standard your whole team must share, the documented failure modes of AI-assisted research, the consent and confidentiality rails around participant and client data, how to stand up a stack, and how to decide - and defend the decision - when the evidence arrived fast.
What AI changes in the pipeline
Where the hours actually go in research, which stages compress, which stages must not - and why "we finished the review in a day" is not by itself good news.
⚖️ Leader 2 · easyThe evidence standard you share
Source tiers, certainty levels, what a sample size buys you, and the one-page standard that stops every review being re-litigated from scratch.
⚠️ Leader 3 · mediumWhere AI fails research
Corpus gaps, agreeable answers that follow your framing, confident summaries of thin studies, and survivorship in whatever ranked first - the failure modes to design against.
🔐 Leader 4 · mediumConsent, confidentiality, governance
Participant transcripts and client material in third-party tools, ethics and data-protection duties, peer-review confidentiality, and what your team may paste where.
🧰 Leader 5 · decidingStanding up the research stack
Search and screening tools, extraction, reference management, qualitative coding - what each layer must do, and what to insist on before you buy anything.
🎯 Leader 6 · decidingDeciding on fast evidence
Confidence language that means something, the audit trail that makes a decision defensible months later, and how to say "we do not know yet" without losing the room.
The researcher track 🔬 · for analysts, academics, UX & market researchers · 10 x 45 min · no coding needed
One running decision, taken through the whole pipeline: should Cadence build a lecture-notes product for higher education. The evidence lab climbs 0 → 1 → 2 → 4 → 7 → 9 surviving claims as the levers come on, and b10 does not ship a memo until every load-bearing claim survives the audit or is honestly marked unknown.
The researcher's AI loop
Question, gather, screen, extract, challenge, synthesise - and the audit that decides whether any of it holds. Start at a humbling 0 of 10 claims surviving.
❓ Researcher 2 · easyDecision, question, claims
Work backwards from the decision to claims that could actually be false - the single change that makes everything downstream checkable.
🗂️ Researcher 3 · mediumSearch strategy and source tiers
Search on purpose instead of accepting what ranked first: databases, query design, and a tier system your team can apply without arguing.
🧹 Researcher 4 · mediumScreening at scale
Inclusion rules written before you look, dedupe, retraction checks, and AI-assisted triage that stays auditable - plus why more sources can make you worse.
📋 Researcher 5 · hands-onExtraction tables
One row per source: sample, measure, number, page, funder. The unglamorous artifact that makes a synthesis reproducible - built with AI, checked by you.
🔬 Researcher 6 · the heartThe evidence lab
The signature session: run the hostile-reviewer audit on all ten claims, read every failure rule, and learn which failures are fatal and which are just a qualifier away.
🧮 Researcher 7 · hands-onContradictions and statistics
Go looking for the source that disagrees, then read numbers properly: effect size versus significance, intervals, publication bias, and what replication really tells you.
🪡 Researcher 8 · hands-onSynthesis you can defend
From extracted rows to a narrative with confidence language that matches the evidence - and disagreement kept visible instead of averaged away.
🎙️ Researcher 9 · hardestInterviews and coding at scale
AI-assisted transcript coding without laundering your own assumptions: codebooks, agreement checks, saturation, and the quotes that must stay verbatim.
🏆 Researcher 10 · hardestCapstone: the research memo
Assemble the whole thing - question, tiers, screening log, extraction table, contradictions, replication - into a memo with an audit trail, and pass the audit.
Choose your path 🗺️
Two doors into the same conviction: speed is only worth having if the conclusion survives someone reading it closely.
The knowledge map 🧠
Both tracks at a glance - hover a session to spotlight its concepts, click any node to jump in.