---
title: "Slides: The Future Is Domain-Specific Agents - Justin Schroeder, StandardAgents"
category: "slides"
video_id: "spNAUEgq_A8"
sourceLabels: ["Public YouTube video frames", "Public YouTube metadata"]
---

# Slides: The Future Is Domain-Specific Agents - Justin Schroeder, StandardAgents

## Source Video
[The Future Is Domain-Specific Agents - Justin Schroeder, StandardAgents](https://www.youtube.com/watch?v=spNAUEgq_A8)

## Relationship To World's Fair 2026
These slides are extracted from a public AI Engineer YouTube video connected to World's Fair 2026. Speaker-matched clips are supporting context unless later confirmed as exact session recordings; official livestream recordings are day-level/event-level source material.

## Related Scheduled Sessions
- No individual scheduled session mapping has been assigned yet; treat this as an event livestream deck.

## Extracted Slides
![[assets/slides/spNAUEgq_A8/slide-001.jpg]]

OCR text:

> A hatem mellem tS
> Domain-Specific Agents
> or
> Nene re | 4

![[assets/slides/spNAUEgq_A8/slide-002.jpg]]

OCR text:

> oo vento (@) StandardAgents |
> 2 FFEEILE arrou:s ® jsonreader
> ZODOLIEI * FormkKit > AutoAnimate
> TEMPO MSU COT
> X. eipschrosder — 7

![[assets/slides/spNAUEgq_A8/slide-003.jpg]]

OCR text:

> ry
> 
> ~~

![[assets/slides/spNAUEgq_A8/slide-004.jpg]]

OCR text:

> Ce UM elastic texas eam termes
> a

![[assets/slides/spNAUEgq_A8/slide-005.jpg]]

OCR text:

> What is an agent?
> 
> X @ipschroeder

![[assets/slides/spNAUEgq_A8/slide-006.jpg]]

OCR text:

> % Claude ®
> @ oh 7
> X. @ipschroeder A

![[assets/slides/spNAUEgq_A8/slide-007.jpg]]

OCR text:

> But everyone 1s
> ob NKebberee: asl eras

![[assets/slides/spNAUEgq_A8/slide-008.jpg]]

OCR text:

> Why:

![[assets/slides/spNAUEgq_A8/slide-009.jpg]]

OCR text:

> Agents are hard
> ¢ Agentic loop orchestration
> ¢ Provider abstraction
> * Durable execution

![[assets/slides/spNAUEgq_A8/slide-010.jpg]]

OCR text:

> >
> It’s amess out there
> ¢ Building robust agents is hard.
> ¢ There is no defined way to do it.
> ¢ Telemetry/observability is hard.

![[assets/slides/spNAUEgq_A8/slide-011.jpg]]

OCR text:

> Client Resources Prompts Toots Discovery Sampling Roots Elicitation Instructions ‘
> Sire x x & ? x x ? ?
> AgentAl x x uw ? x x 2 ?
> AgenticFlow u iv) ta ts x x ? ?
> AIQL TUUI iv] iv) iv] ta iv] x 4 ?
> Amazon Q CLI x ud iV] 2 x x 2 ?
> Amazon Q IDE x x G x x x 2 ?
> Amp iv} ta Ga x iV} x ? ?
> Apify MCP x x i % x x 2 ?
> Tester
> Augment Code x x iv] x x x ? ?
> Avatar Shell iv] x 4 x x x ? ?
> BeeAl x x Wi x x x
> Framework
> BoltAl x x iv] ? x x & .
> Call Chirp x iV} iV] x x x . y
> Chatbox x x G x x x
> ChatFrame x x iv] x x x
> ChatGPT 4 4 td Ww +4 a ? >?

![[assets/slides/spNAUEgq_A8/slide-012.jpg]]

OCR text:

> i aie
> 
> _— rz a , fin .
> 
> * vA y _ Psi - @ : . : x Ze
> - ob a ef Ne as
> ne ny ES . 4 oh Mee i : By ay
> 
> q a at ae ¥ <p ‘ “a 7 + ;
> 
> * = ee : | | i ; g ¢ re "A " f i a fl
> N J SS =" Fae ea , Y & oo
> 
> aaa i -—) oe S| aL om es
> 
> i] a , a SS ee ay (mri iN ae ead ys iN
> d I . +8 DOP: ne | ’ AA aga on % ff i
> 
> ' 7 Sh ™ . # re , sad % a or
> 
> mea roy a) OO" ee 3 a ae “et a
> 
> y tj 1 or ro ty ae a2 54 rho re
> 1 2 el AC ; ee
> 
> i ee. | ls el
> 
> iene en. Mi ae ° a u ; Amr
> 
> f Fy oe tw et fr ae a i < - ry od ] F Py z
> 
> po i, ee tees a ae . x , F
> rr er ne a a a =
> ar eae tee ice ae, a i" WV a .
> ne eo eee ar G a
> ee ey = 7 nics
> a ea
> * nite —_ ee! le . t : i
> iin aid
> oe sg. a - ee ae ww ~~
> By Re afar Ce Ste en ak

![[assets/slides/spNAUEgq_A8/slide-013.jpg]]

OCR text:

> System Prompt
> Poses aa ;
> /

![[assets/slides/spNAUEgq_A8/slide-014.jpg]]

OCR text:

> Messages
> MCP
> Skill
> Context
> Tools
> SystemPrompt
> Model

![[assets/slides/spNAUEgq_A8/slide-015.jpg]]

OCR text:

> Messages
> MCP
> KAYAK
> NAVAN
> Figma
> Playwright
> Gmail
> Google Sheets
> Skill
> Tools
> SystemPrompt
> Model

![[assets/slides/spNAUEgq_A8/slide-016.jpg]]

OCR text:

> Pe ,
> Own NAVAN
> FFigMa & Playwright
> mm Gmail B Google Sheets ‘
> aus 7 enance
> Ay aT |
> OR feat Code Fix btinter |
> nt
> DO tithe
> ae

![[assets/slides/spNAUEgq_A8/slide-017.jpg]]

OCR text:

> System Prompt
> i FFigma : om
> *oncoscencusecscnsceconccee® ss x
> =.

![[assets/slides/spNAUEgq_A8/slide-018.jpg]]

OCR text:

> / | ae \
> System Prompt System Prompt System Prompt SNES Gg
> i FFigma : : Gmail : ; gnavan He = Keree [=
> *wacccrevccscecerscesenacen® Pasecccccvscuccessvcouaces® A eeceecccsssevacsascccasas® *ouscceecces ‘es *
> cD Cx Ci r

![[assets/slides/spNAUEgq_A8/slide-019.jpg]]

OCR text:

> ae =
> eee eee _
> a 0 \
> or a
> aR Mee rt ag eRe ta.
> SN eet Ont ae ee a are |
> ? eet + if ; oA
> a \: ae ss int } FANG ; aha \* 4 e. & NS "oe
> “, <a ‘ as | \i A ‘9 “E SN ZS . ‘
> ah\. KAS \N ar oN eee aa
> MS OS ec a)
> a YY ‘ ea 3 me ;
> wr : a rr se " As . fend
> . a 5 Oe 4 Le e
> " cee We y N | * -
> en # \* ‘
> ee ‘ 7 ae

![[assets/slides/spNAUEgq_A8/slide-020.jpg]]

OCR text:

> Domain Specific Agents...
> ¢ Far more efficient with tokens

![[assets/slides/spNAUEgq_A8/slide-021.jpg]]

OCR text:

> Domain Specific Agents...
> ¢ Far more efficient with tokens
> = ky

![[assets/slides/spNAUEgq_A8/slide-022.jpg]]

OCR text:

> Domain Specific Agents...
> ¢ Far more efficient with tokens
> * Make small language models practical
> DeepSeek v4 flash is
> 137x cheaper than Fable &
> per task A 7
> e" if

![[assets/slides/spNAUEgq_A8/slide-023.jpg]]

OCR text:

> Domain Specific Agents...
> ¢ Far more efficient with tokens
> ¢ Make small language models practical
> * Can enforce strict limits on capabilities

![[assets/slides/spNAUEgq_A8/slide-024.jpg]]

OCR text:

> Domain Specific Agents...
> ¢ Far more efficient with tokens
> * Make small language models practical
> * Can enforce strict limits on capabilities
> ¢ Have excellent scaling characteristics .
> my


## Slide-Derived Subjects To Review
Subject extraction uses video title, related session titles/descriptions, transcript context, and OCR text when available. OCR is best-effort and should be reviewed against the embedded slide images.