BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.com//pydata-amsterdam2026//speaker//JFHJZ9
BEGIN:VTIMEZONE
TZID:CET
BEGIN:STANDARD
DTSTART:20001029T040000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
TZNAME:CET
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20000326T030000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
TZNAME:CEST
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-pydata-amsterdam2026-KHXZLR@pretalx.com
DTSTART;TZID=CET:20260911T132500
DTEND;TZID=CET:20260911T135500
DESCRIPTION:Multi-agent LLM systems typically fall into one of two patterns
 : agents take fixed turns\, or a single orchestrator decides who speaks ne
 xt. Both are easy to implement\, but both sacrifice agent autonomy. As age
 nts become increasingly capable and specialized\, productive and optimized
  communication between them will be crucial to unlocking their full potent
 ial. A key part of this is deciding when it is time for an agent to speak.
  In this talk\, we will discuss whether agents can negotiate their own spe
 aking order\, and how we might evaluate their results. \n\nWe will use a S
 crum refinement meeting as our multi-agent playground. A team of role-base
 d agents\, representing classic roles from a Scrum team\, must collaborati
 vely turn a business request into a backlog of implementable user stories.
  We implement and compare different turn-taking strategies and evaluate th
 ese using LLM-as-a-judge across metrics including participation balance\, 
 unanswered questions\, redundancy\, topic drift\, and urgency realism. \n\
 nAttendees will learn about multi-agent communication patterns\, concrete 
 turn-taking strategies they can put into practice\, comparative evaluation
  results\, a reusable evaluation strategy\, and a public repository with c
 ode and notebooks they can adapt to their own multi-agent experiments.
DTSTAMP:20260710T150510Z
LOCATION:Unconference
SUMMARY:When Should AI Speak? Letting Agents Decide When It's Their Turn - 
 Emir Can\, Anna Pillar
URL:https://pretalx.com/pydata-amsterdam2026/talk/KHXZLR/
END:VEVENT
END:VCALENDAR