AIAgentBible:Theultimateguidetoagentdisruption011AIAgentBibleTheAIagentlandscapeisevolvingrapidly,with500+startupsfoundedsince2023.Hereisourtop

AIagentresearch—featuringthetrends,keyplayers,and6predictionstowatchallbasedon

CBInsightspredictiveintelligence.AIagentsarede?ningthenextwaveof

techinnovation.Everybig

techcompanyandarapidlygrowingprivatemarketlandscapearebuildingagent

offerings

targetinghorizontalusecasesandindustriesfrom?nancialservices

tomanufacturing.Forenterprisesacrosssectors,onequestionisbecomingunavoidable:WhichAIagent

strategieswillseparatemarketleadersfromthoseleftbehind?Enterprisesareunderpressure

tobuildandimplementagentsas

theseLLM-basedsystems

changehowcompaniesoperate,hire,andscale.Across9reports,discoverwherestartupinnovationispointing,promisingpartnershipand

acquisition

targets,andkey

trends

towatchbasedonCBInsightspredictiveintelligence.?2025CB

Insights

|

1Sevenyearsago,Iwatchedcompaniesstrugglewithprimitivechatbotsthatbroke

the

momentcustomerswentoff-script.Today,those

early

systems

have

given

way

to

AI

agents,which

in2short

years

have

movedfrom

experiment

to

enterprise

priority.I’m

seeing

agent

mentions

on

earnings

calls

10xsince2023.Thevelocityisunlikeanythingwe’ve

trackedbefore.Whatstrikesmemost:agentsareclimbingthevaluechainfasterthananytechnologyI’ve

seen.Theygainedtractionincustomerservice–whichmadesense,given

thesector’searly

AIadoption.

82%oforganizationswesurveyedinJune2025

said

theywillbeusingAIagents

incustomersupportin

thenext12months.Butnow

they’re

supportingclinicaldecisions

at

hospitals,

assessing?nancialrisk

atbanks,and

draftinglegalmemos

atlaw?rms.Thisiswhy

25%ofAIagentstartupsfoundedsince2023

arealreadydeploying,achieving

levelsofcommercialdistributionthatpreviouslytookstartups5+years,accordingtoour

CommercialMaturityscores

.Themomentumisaccelerating.Among

the

1,500+

techmarkets

CBInsights

tracks,

?ve

ofthetop10bydealactivityin2025areagent-related.Onein?venewunicorns

isbuildingagenttechnology.Halfofthe

top20revenue-generatingagentstartups

didn’tevenexist

three

yearsago.But

here’s

what

I

keep

telling

our

enterprise

clients:agents

are

only

as

good

as

the

data

they

runon.Therealcompetitiveadvantageisn’tintheAI–it’sinoperationalizingaccurate,comprehensivedataand

theorchestrationlayers

thatmakeagentsactuallyuseful.That’swhere

theroadblocksare.Enterprisesunderstandagentsmatter,but

they’restruggling

withintegrationcomplexity,securityconcerns,andvendorsprawl.Theopportunityismassivefor

thosewhocannavigate

this

transition.Theresearchanddatathatfollowsbreaksdownthisrapidlyevolvinglandscapebyvertical,maturity,and

momentum.Remember:in

the

age

of

AI

agents,the

edge

doesn’t

live

in

themodel.Itlivesinconnectingpurpose-builtagents

topurpose-speci?cdata.Thosewhoget

thisrightwillshape

theeconomyagentsarecreating.ForewordManlioCarrelliCEOofCBInsights?2025CB

Insights

|

2AIAgentBiblePredictiveIntelligenceonPrivateCompaniesWithCBInsights,youseeyournextmove–?rst.Rare

data

straightfrom

companies

buildingthefuture.PowerfulAIthatautomatestheanalysis.?2025CB

Insights

|

3AIAgentBible6AIagentpredictionslookinginto20267TheAIagentecosystem13TheAIagentmarketmap13TheAIagent

techstack19TheAIagentrevenuerace29AIagentsmakeinroadsacrossenterprisework?ows31YCombinator’s2025SpringbatchrevealsthefutureofagenticAI31Buildingtheagenteconomy:HowcloudleadersareshapingAI’snextfrontier36Thesummerofvibecodingisover43Industry-speci?capplicationsgainmomentum513marketsfueling

theshift

toagenticcommerce51TheindustrialAIagents&copilotsmarketmap58100real-worldapplicationsofgenAIacross?nancialservicesandinsurance64TableofContentsOutlookonAIagents

5Additionalresources68?2025CB

Insights

|

4AIAgentBibleOutlookonAIagents2025:AgentswithguardrailsAnAIagentisanLLM-basedsystemdesigned

toindependentlyperform

tasksonauser's

behalfbyreasoning,planning,leveragingmemory,andinteractingwithexternal

toolsand

otheragents.Mostagentstodayoperatewithinconstrainedenvironments,utilizingstructuredwork?ows

and“guardrails”

tocompletespeci?cobjectiveswhileretainingsomedecision-makingcontrol.Asfoundationmodelcapabilitiesimprove,agentsareexpected

tobecomeincreasingly

autonomous.Beyond2025:FullyautonomousagentsAgentswilloperatewithouthumaninvolvement,withmoresophisticateddecision-making,

adaptability,and

taskexecution.Lookinginto2026,watchfornewformfactorsoutsideof

thecopilot/chatbotinterfacethat

willpush

theboundariesofwhatan“agent”is.?2025CB

Insights

|

5AIAgentBible“AIagentsarenowhereneartheir?nalform.Whatwe

mean

isthis:todaywe’re

ina“copilot”erawherewegetoutputsviaAIchat.Butagentsarealreadytransformingintosomething

more

likeasuperpoweredinstrumentortool.Toolsshapetheworkwedo,soanAIsupertoolwillallowourworktogoinallkindsof

previously

impossible

directions.Inenterprise,thecopilotmodelhashelpedinareaswhere

knowledgework

is

both

routine

andhigh-stakes,

like

support,training,onboarding,and

compliance.But

the

next

profound

shift

willbeforagentstoceasewaitingforinput,insteadactingasproactivecollaboratorsandextensionsofourminds.Interactingwithanagentwillnot

justhelpsolveone

problem;

itwillactuallyshapehowanorganizationfunctionsandreactstochange.Overtime,enterprisesthatembraceAIsupertoolsascorelearningsystemswill?ndtheir

culturesandcapabilitiestransformedinwaysthatgofarbeyonde?ciencygains;theywill

unlockentirelynewideas,solutionsandways

ofworking.MichaelMignanoPartnerat

LightspeedEarlyindicationsofthisinclude“AI-native”workspaces—toolsandplatforms

builtfromthe

grounduparoundAIcapabilities,ratherthanlayeringAIfeaturesontopof

atraditionalproduct.Otherexamplesincludewebbrowsinginterfacesandvoice-drivenapplications.?2025

CB

Insights

|6AIAgentBible6AIagentpredictionslookinginto20261)VoiceAIsprintsaheadEarly-stagegenAIcompanieswith

thefastest-growingheadcounts

areconcentratedinAI

agentapplications—andmorespeci?callyinvoiceAIdevelopment.CompaniesarenowpositioningthemselvesforafuturewherehumansinteractwithAIvia

conversationratherthantextinterfaces.ForagenticAIdevelopment,implicationsaresigni?cantacrosscustomerservice,sales,andITsupportwherevoiceagentscanhandle

complexconversationswithouthumanintervention.?2025CB

Insights

|

7AIAgentBibleWatch

voiceAIdevelopmentstartups

forpartnership,investment,andacquisitionopportunities.SignaIing

thepotentiaIforincreasingconsoIidation,

Metaacquired

voiceAI

startups

PlayAI

and

WaveFormsAIin2025.2)WaveofAIM&AcomesforagentsAgentic

solutions

led

the

top

AI

exits

to

kick

off

the

year

in

Q1,25.Moveworks,

Weights&Biases,and

OfferFit

secured

the3IargestdeaIsamong85acquisitions—estabIishingagents

as

theprimaryfocusofindustryconsoIidation.OveraII,therehavebeen

35+acquisitionsinthebroaderAIagent&copilotspacein2025so

far.IntheIastquarter,thatincIudesOrbyAI

(acquiredbyUniphore),Windsurf

(acquiredby

Cognition),Cognigy

(acquiredbyNICE),and

Flowise

and

SanaLabs

(acquiredbyWorkday).EnterprisebuyersareincreasingIyseeking

tobuiIdcomprehensiveagentsoIutions

togaina

competitiveedge.Twofurther

areasofconsolidationwepredict:1.SalesandmarketingAIagents

offerIow-hangingfruitforSaaSIeaders’agentpIans.2.ThecodingAIagent&copilots

marketisripeforconsoIidationamidexpIosivegrowth,soaringvaIuations,andafracturedIandscape.Partnershipswithmajor

techandAI

pIayerssetapart

topcodingM&A

targets.3)ThemarginsqueezehitsbeyondcodingAIagentstartupsacrossverticaIswiIIface

thesameeconomicpressures

that

reshapedcodingAImarginsin2025.ReasoningmodeIs,whichsparked

theriseof“vibecoding”(giving

ahigh-IeveIgoaIanddeIegatingmuIti-stepimpIementationto

theAI),aIso

transIatedintohighercomputecostsastheyin?atedoutput-tokenvoIumeroughIy20x.CBInsightsAcquisitionHunterIdenti?es

promising

early-stage

disruptors

and

acquisition

ready

assets

precisely

and

before

competitors

even

noticethem.Learn

More?2025CB

Insights

|

8AIAgentBibleExpectotheragentcategorieswithgrowingusagetoreworkpricingandcontractsasreasoningcostsmount.Forexample,inMay2025,Salesforce’sAgentforceshiftedprices

from$2perconversationtoahybrid-usageFlexCreditssystem,tyingcreditstonecessary

actionsforanoutcome.Asmarginscompress,strugglingplayerswillseekexits—seeforexample

Cognition’s

acquisitionofWindsurfafterthestartup’slicensingdealwithGoogle.4)ThefoundationsforagenticcommercesolidifyOneof

thebiggestbarrierstofullyautonomousshoppingisenablingsecure,real-timetransactions.Anewclassofstartupsis

tackling

thischallengehead-on,building

AI-nativepaymentrailsanddigitalwallets

thatletusersauthorize—andlimit—spendingbyAIagents.Overall,the

AIagentpaymentsinfrastructure

marketisoneof

themorenascentmarketsin

theAIagent

techstack,withanaverageCommercialMaturityof2.4(validating)andanaverageMosaicscoreof697.MosaicscoreEvaluatestheoverall

health

and

growth

potential

of

privatecompanies

based

on

performance,?nancial

stability,

market

conditions,

and

management

strength.It

combinesthesefactors

into

a

single

score(out

of1,000).Inanindicationof

the

tech’spotentialandthe

technicalhurdlesthatmustbeovercome,

establishedcardandpaymentnetworks

areinvestingandpartneringwithstartupsin

the

market(seegraphic).?2025CB

Insights

|

9AIAgentBibleStripe

meanwhileannouncedinSeptember2025it’slaunchinganAPIforagenticpayments.

The$107Bpaymentsgiantalsolaunchedwith

OpenAI

theAgenticCommerceProtocol,whichwilleventuallyenableAIagents

tocompletepurchasesonbehalfofusersbyproviding

astandardizedcommunicationframeworkbetweenbuyers,AIagents,andbusinesses.As?ntechgiants,AIstartups,andcommerceplatformsconvergeon

theagenticpayments

challenge,thisearlycollaborationwillshapehowfuture

transactionswillhappen:securely,

autonomously,andinside

theAIinterfacesconsumersalready

trust.4)“Datamoatwars”reshapeenterprisesoftwareSoftwareincumbentsarewallingoffaccess

to

theircustomerdataasAIagentsbecome

morecapable.

Salesforce’snewSlackAPIratelimitsin2025,whichpreventexternalapps

fromaccessingchatdatainbulkorstoringitlong-term,areaclearexample.

AtlassiansimilarlyannouncedadditionalratelimitingforitsCon?uenceandJiraAPIsin2025.?2025CB

Insights

|

10AIAgentBibleThiscreatesproblemsforAIstartups

thatdependonaccessingdataacrossmultiple

systemsofrecord

toautomatework?ows(e.g.,knowledgemanagementplatformslikeGlean).Butitalsocreatesfrictionforenterpriseswhowanttointegratedataacrosstheir

apps.Acountermovementisemerging.InSeptember2025,

Snow?ake

launchedaconsortiumwithoveradozenproviders(includingSalesforce)committing

tostandardizeddataformats

thatenableAI

toaccessinformationacrossmultipleapplications.Whilerepresentingasmall

fractionof

themarket,itsignalsenterprisedemandforopennessovervendorlock-in.Thistug-of-warwillplayoutinafewways.WeexpectincumbentswillcontinuerestrictingAPIaccesstoprotecttheirAIproductrevenues.Enterprisesthatwantmorecontrolovertheir

datawillpushback,drivingdemandforsolutionsthathelpthemown

theirdatainfrastructure

directly.Somestartupswillgo“upthestack”tobuildagentsthatworkdespitelimitedAPIaccess(e.g.,viacomputeruse),whileotherswillmoveclosertowardwheredataisstored

andprocessed.5)AgentmonitoringtoolsemergeasanessentialenterprisecategoryAIagentreliability

remainsamajorchallengein

thelandscape.Agents

thatfail,hallucinate,orbehaveunpredictablycreateimmediateoperationalproblems.Thisisdrivingactivityinoversightcapabilities

tomanageagentrisk.Analyzing

the

7early-stagedeals

($30.9M

totalfunding)in

the

AIagentobservability,evaluation,&governancemarket

sofar

thisyearhighlightsemerging

technicalneeds:●Voice

agent

testing:Cekura

($2.4M

seed)and

Coval

($3.3M

seed)both

focus

onvoiceAI

testingandsimulation,signaling

thatconversationalagentsrequiredistinct

observabilityapproaches.●Synthetic

user

generation:Multiple

companies(Cekura,Coval,Janus)focus

onsyntheticusergenerationforagent

testing.●AI

productivity

measurement:Larridin’s$17M

seed

round

led

by

Andreessen

Horowitz

showsenterprisedemandforquantifyingAIagentROI,withafocusonmeasuringhuman-AIworkforceproductivity.?2025CB

Insights

|

11AIAgentBible“Assecurityagentsbecomemoreprevalent,thenextgenerationofcybersecurityprofessionalswillevolveinto“generals”managingarmiesofAIagentsand“specialforces”handlingtier-3incidentresponseandlargecross-functionalarchitecturechanges.Tosupport

theserolesandaddressexplainabilitychallengesincriticalsecuritydecisions,mostAI

securityagentsleverageLLMsnotonly

toperformtasksbutalso

toconvertcomplexactions

andenrichmentintoeasy-to-understandreportswithcitationsonkeyinformationsources.EdwardWuFounder&CEO,DropzoneAICompaniesLatestroundDealsizeDealdateRoundinvestorsCekuraSeed

VC-II$2.4M7/1/2025AlanRutledge,AllenWu,AllportCapitalTraceloopSeed

VC$6.1M5/27/2025GrandVentures,IbexInvestors,SamsungNELarridinSeed

VC$17.0M4/2/2025AndreessenHorowitz,BloombergJanusConvertiblenote$0.5M4/1/2025YCombinatorLangWatchPre-Seed-II$1.1M2/25/2025Antler,PassionCapital,VoltaVenturesCovalSeed

VC-II$3.3M1/23/2025FortitudeVentures,GeneralCatalystLucidicAIConvertiblenote$0.5M1/1/2025YCombinatorSource:CBInsightsBusinessGraph.Data

as

of

8/26/2025.Companiesdeployingagentsatscalewillneedproperoversight

tools

toavoidcostlyfailures

andmaintainoperationalcontrol.2025early-stagedealstoAIagentobservability,evaluation,andgovernancecompanies?2025CB

Insights

|

12AIAgentBibleKEYTAKEAWAYS●AIagentfunding

isexplodingas“digitalcoworkers”becomereality:AI

agent

startups

raised$3.8B

in2024

(nearly

tripling2023's

total),with

every

major

tech

player

now

developing

agents

oragenttooling.TheshiftfromAIcopilotstoautonomousagentswithgreaterdecision-makingscopeishappeningquickly,promisingtoalterworkforcecompositionand

maximizeoperational

e?ciencythroughtaskautomation.●Customerserviceandsoftwaredevelopmentleadcommercialadoption:Thesetwo

horizontalmarketsshowthestrongesttractionbasedonmedianMosaichealthscores(714and737respectively),withtwo-thirdsofsurveyedorganizationsalreadyusingorplanningtouseAIagents

incustomersupportwithin12months.Thesemarketsbene?tfromwell-de?nedwork?owsand

testableenvironmentsthatmakeagentdeploymentmorestraightforward.●Trust

remains

the

biggest

barrier

to

full

autonomy:Despite

rapid

growth,

fully

autonomous

agentsremainlimitedduetoreliability,reasoning,andaccessissues.Startupsare

addressingthisthrough

5primarytrust-buildingmethods:transparency,humanoversight,technicalsafeguards,security&

compliance,andcontinuousimprovement.Mostcurrentapplicationsstilloperatewith“guardrails”

withinconstrainedarchitecturesratherthantrueautonomyTheAIagentecosystemMappingthe

landscapeof

startups,infrastructure

providers,and

emerging

revenue

leaders

reshapingthe

marketTheAIagentmarketmapWeminedtheCBInsightsdatabasetomap170+AIagentstartupsacross

26categories.We

alsoprovideanoutlookonAIagents’progress,limitations,

andfuture

directions.Published3/6/2025—

Link

to

reportWe

selected

companiesfor

inclusion

based

on

Mosaic

health

scores

(500+)and/orfunding

recency

(since

2022).We

included

privatecompaniesonly

andorganized

themaccordingto

their

primary

focus.This

market

map

is

notexhaustiveof

the

space.Wantto

beconsideredforfuture

AIagent

research?

Brief

our

analysts

to

ensurewe

have

the

mostup-to-datedataon

your

company.?2025

CB

Insights

|13AIAgentBible?2025CB

Insights

|

14AIAgentBibleAIagentinfrastructureThissegmentcoverscompaniesbuildingagent-speci?cinfrastructure.(Weexcluded

general

genAIinfrastructuremarkets

likefoundationmodelsandvectordatabasesfrom

themap.)DevelopmenttoolsAdiverseecosystemof

toolshasemerged

tosupportagents’development.Theserange

frommemoryframeworkslike

Letta

thatenablepersistent,retrievablememoryacross

interactions;to

tools

thatallowagentsto

takeactionviaintegration(e.g.,

Composio),authentication(e.g.,

Anon),andbrowserautomation(e.g.,

Browserbase).Anothersetofcompaniesisgivingagentsmoreutilityacross

payments

(whichincludes

companiesdevelopingcryptowalletsforagentsaswellasvirtualcards)and

voice(developmentplatformsand

toolsfor

testingAIvoiceapplicationsaswellasspeech

models).Meanwhile,demandforsimpli?ed,comprehensivedeploymentoptionsisdrivingtheriseofAIagentdevelopmentplatforms

—themostcrowdedinfrastructuremarketonourmap.LLMdevelopersincludingCohere

(withitsNorthAIworkspace)and

Mistral

havelaunchedtheirownagentdevelopmentframeworks,while

Amazon,

Microsoft,

Google,and

Nvidia

allofferAIagentdevelopment

tooling.With

manyenterprisesfavoringestablishedvendorsduetolowerrisk,big

techcompanieshavesigni?cantadvantageshere.Trust&performanceConcernsaroundreliabilityandsecurityhavehelpedestablishamarketforagent

evaluation

&observability

tools

.Early-stagecompaniesare

targetingapplicationssuchasautomated

testing(e.g.,

HaizeLabs)andperformance

tracking(e.g.,

Langfuse).Multi-agentsystems,where

specializedsub-agentswork

together

tocomplete

tasks,also

showpromiseinimprovingaccuracy.InsightPartners-backed

CrewAI’smulti-agentorchestrationplatformisreportedlyalreadyusedby40%oftheFortune500.?2025CB

Insights

|

15AIAgentBible“Jo?oMouraCEO,CrewAIVendorsarealsotacklingreliabilityconcernsdirectly.Basedonourbrie?ngswith20+AI

agentstartupsinQ1’25,companiesareusing5primarymethodstobuildusertrust:1.

Transparency2.

Humanoversight3.Technicalsafeguards4.

Security&compliance5.

ContinuousimprovementMulti-agentorchestrationwillevolveintoanOSforenterprises—governingthousandsofagentswithrulesforautonomy,oversight,andauditability.Pre-builtagentsmayhelp

earlyadopters,butlong-term,enterpriseswillfavorplatformsthatletthemcomposeand

governcustomagentsatscale.?2025CB

Insights

|

16AIAgentBibleHorizontalapplications&

jobfunctionsHorizontalAIagentstartupsmakeupnearlyhalfof

themapandoveralllandscape.Thissegmentprimarilyfeaturesstartups

targetingenterprises,withindustry-agnostic

applicationsacross

jobfunctionslike

HR/recruiting,

marketing,and

securityoperations.Companiesin

theproductivity&

personalassistants

market,including

OpenAI

withits

Operatoragent,are

targetingconsumersandemployeesdirectly.TheAIagentmarketswiththemosttraction—basedoncompanies’medianMosaichealth

scores—are

customerservice

andsoftwaredevelopment(whichincludes

coding

and

code

review&

testing

agents).Thesemarketsarealsoamong

themostcrowded

due

to

thevalue

agentsbring

towell-de?nedwork?owsand

testableenvironments.Weseethisre?ectedinadoption,particularlyatthecustomerservicelayer:Among64organizationssurveyedbyCBInsightsinDecember2024,

two-thirdsindicatedtheyareusing

orwillbeusingAIagentsincustomersupportinthenext12months.Overall,horizontalAIagentapplicationsaremorecommerciallymaturecomparedtotheinfrastructureandverticalsegments,withovertwo-thirdsof

themarketdeployingorscaling

theirsolutionsbasedon

CBICommercialMaturityscores.?2025CB

Insights

|

17AIAgentBibleVertical(industry-speci?c)applicationsWeexpectincreasingverticalizationasstartupscarveoutnichesbysolving

industry-speci?c

customerproblems,especiallyinareaswithstrictregulatoryscrutinyand

data

sensitivity.Thiscategoryfeaturescompaniescateringtoindustriesincluding:●Financialservices&insurance:

Themostcrowdedverticalcategoryonthe

mapwith

11companies,startupsherearetargetingavarietyof?nservwork?owssuchas?nancialresearch(Boosted.ai

andWokelo),insurancesales&support

(Alltius

and

Indemn),andwealthadvisoryprospecting&operations(FinnyAI

andPowder).●Healthcare:Solutionsinthismarketaimtoreducethevolumeof

manualtasksforhealthcareprofessionalsacrossusecaseslikeclinicaldocumentation,revenuecycle

operations,callcenters,andvirtualtriage.SolutionsfromcompanieslikeThoughtful

AI

(revenuecycleoperations)andHippocraticAI(sta?ngmarketplace)aretargeting

end-to-endhealthcarework?ows.●Industrials:Thesecompanieslooktooptimizeprocessesandequipment

—

including

controlsystems,robots,andotherindustrialmachines

—without

relyingon

consistent

humanintervention.Forexample,Composabl

launchedanagent

platform

in

May2024thatusesLLMstocreateskillsandgoalsforagentsthatcancontrolindustrial

equipment.PubliccompanieslikePalantir

arealsoactiveinthisspace.Learn

more

in

ourindustrialAIagents&copilots

market

map

.“Thepromiseofautonomousagentscannotbeful?lledwithoutsolvingcontextengineering.DanielChalefFounder

and

CEO,Zep

AIAgentsneedtherightcontextattherighttimeandintherightformatto

make

reliabledecisions.Enterprisesalsorequireprovenanceandgovernance:

knowingwhere

information

originated,howitevolved,andensuringcomplianceacrosssensitivedomains.

Currentapproachestoagentmemorystilltreateachuserinteractioninisolation,with

little

connection

toothertouchpointssuchasemail,supportcases,CRMdata,or

SaaS

usage.

Enterpriseswill

needcontextsystemsthatunifythesesignals,trackevolvingrelationshipsandbusinessstates,andmaintaincontrolsandaudittrailsthatmakeagentstrustworthyatscale.?2025

CB

Insights

|18AIAgentBibleKEYTAKEAWAYS●Voice

AI

emerges

as

a

new

battleground

with

an

average

Mosaic

score

of756and$400M

raisedin2025,whileMetaacquiredits?rstcompaniessince2022(PlayAIandWaveFormsAI)in

thisspace.●TechgiantsracetoownagentcommunicationstandardswithAnthropic'sMCP,Google'sA2AProtocol,andIBM'sAgentCommunicationProtocolalllaunchingwithinayear.Professionalservices?rmsandcloudprovidersarerapidlyadopting

theseprotocols,making

them

tablestakes

across

thevaluechain.●Criticalwhitespaceopportunitiesemergeacrossmonetization,marketplaces,andcost

managementastheAIagentstackmaturesInunderayear,

theAIagentlandscapehasgrownfromroughly

300players

to

thousands

.

Agentsaremakingtheirwayinto

work?ows

acrossverticals,from

e-commerce

toindustrials

.Underpinningthismomentumisanemergingtechstack.Infrastructurelayers—from

foundationmodelstooversight—arehelpingenterprisesbuild,deploy,andmanageAI

agentsmoreeffectively.Using

theCBInsightsBusinessGraphandproprietarysignals,wemapped135+promising

privatecompaniesbuildinginfrastructureforAIagents.Below

themap,weoutline

theemergingmarketsandtrendsinvestorsandstrategyleaders

shouldbewatching.TheAIagenttechstackWeminedtheCBInsightsBusinessGraphtomapthe135+startupsacro