BEGIN:VCALENDAR
VERSION:2.0
PRODID:IEEE vTools.Events//EN
CALSCALE:GREGORIAN
BEGIN:VTIMEZONE
TZID:America/New_York
BEGIN:DAYLIGHT
DTSTART:20260308T030000
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
RRULE:FREQ=YEARLY;BYDAY=2SU;BYMONTH=3
TZNAME:EDT
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20261101T010000
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=11
TZNAME:EST
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260820T211507Z
UID:A11606D8-4ECB-4A77-A815-66EACB1AE6EA
DTSTART;TZID=America/New_York:20260830T140000
DTEND;TZID=America/New_York:20260830T150000
DESCRIPTION:Large language models are increasingly commoditized\, yet enter
 prise AI agents built on them routinely fail to produce accurate\, trustwo
 rthy\, or auditable outputs. This session argues that the missing piece is
  not a better model but a missing architectural layer: the Context Layer. 
 A three-tier framework for AI-readiness will be introduced — a governed 
 data foundation\, a semantic context layer\, and the agents themselves —
  showing that the majority of the effort required to stand up a reliable e
 nterprise agent lies in the middle layer\, not in the agent or its integra
 tions.\n\nThe talk will discuss a reference architecture for this Context 
 Layer\, comprising a semantic repository (ontologies\, business glossaries
 \, certified metrics\, and persona-aware dynamic context)\, a governance w
 orkflow for authoring\, certifying\, and publishing definitions with full 
 lineage and drift detection\, and a single open interface through which ev
 ery downstream tool and agent discovers and consumes certified context. A 
 five-step reasoning pipeline — interpret\, decompose\, match\, apply con
 text\, and assemble — will be explained\, showing how natural-language q
 uestions are grounded in this governed model before any query executes\, w
 ith role- and attribute-based access control and compliance guardrails enf
 orced on every answer.\n\nCo-sponsored by: AI-Native Networking & Systems 
 Technical Community \n\nSpeaker(s): Chitiz\n\nVirtual: https://events.vtoo
 ls.ieee.org/m/573167
LOCATION:Virtual: https://events.vtools.ieee.org/m/573167
ORGANIZER:nxsvks@rit.edu
SEQUENCE:30
SUMMARY: Data Product as backbone of AI 
URL;VALUE=URI:https://events.vtools.ieee.org/m/573167
X-ALT-DESC:Description: <br /><p dir="ltr">Large language models are increa
 singly commoditized\, yet enterprise AI agents built on them routinely fai
 l to produce accurate\, trustworthy\, or auditable outputs. This session a
 rgues that the missing piece is not a better model but a missing architect
 ural layer: the Context Layer. A three-tier framework for AI-readiness wil
 l be introduced &mdash\; a governed data foundation\, a semantic context l
 ayer\, and the agents themselves &mdash\; showing that the majority of the
  effort required to stand up a reliable enterprise agent lies in the middl
 e layer\, not in the agent or its integrations.</p>\n<p dir="ltr">The talk
  will discuss a reference architecture for this Context Layer\, comprising
  a semantic repository (ontologies\, business glossaries\, certified metri
 cs\, and persona-aware dynamic context)\, a governance workflow for author
 ing\, certifying\, and publishing definitions with full lineage and drift 
 detection\, and a single open interface through which every downstream too
 l and agent discovers and consumes certified context. A five-step reasonin
 g pipeline &mdash\; interpret\, decompose\, match\, apply context\, and as
 semble &mdash\; will be explained\, showing how natural-language questions
  are grounded in this governed model before any query executes\, with role
 - and attribute-based access control and compliance guardrails enforced on
  every answer.</p>\n<p>&nbsp\;</p>
END:VEVENT
END:VCALENDAR