Partners begin to target agentic AI marketplace platforms
Cloud marketplaces are emerging as key venues for adoption of AI agents. Consultants and integrators aim to offer the technology but will need to work out pricing methods.
Cloud marketplaces could emerge as a prime mechanism for fulfilling customer demand for agentic AI, with IT service providers set to play an expanding role.
For example, AWS Marketplace, Google Cloud Marketplace, Microsoft's Azure Marketplace and Salesforce's AppExchange provide AI agents as a subcategory of their offerings. SaaS vendors are typically the key suppliers to those budding agentic AI marketplace platforms. Service providers, however, are also poised to sell their wares. This trend has happened before: Software technology companies first filled marketplaces with their cloud products, followed by consultants and systems integrators offering professional services. Indeed, service providers have made significant marketplace inroads over the last five years, generating significant revenue from cloud services and generative AI.
Agentic AI adds another element to service providers' marketplace business mix. The market for such offerings appears ripe for expansion. In an EY Technology Pulse Poll conducted in April, 48% of the more than 500 technology leaders surveyed said they were already adopting or fully deploying agentic AI. In addition, the EY poll revealed that 92% of respondents plan to increase their AI spending in the next 12 months, with 43% of this group allocating more than half of their AI budgets to agentic AI.
Agentic AI grows in marketplaces
Julie Linn Teigland, EY global vice chair for alliances and ecosystems, sees marketplaces as a key venue for enterprise adoption of AI agents.
"I think that is how you will see agentic AI being increasingly deployed," she said. "[Marketplaces] help reduce friction in how organizations discover, test and purchase AI tools. It is super easy to access, which means you can create the solutions faster and you can innovate faster."
[Marketplaces] help reduce friction in how organizations discover, test and purchase AI tools.
Julie Linn TeiglandGlobal vice chair for alliances and ecosystems at EY
Matt Yanchyshyn, vice president of AWS Marketplace and partner services at AWS, also acknowledged this potential.
"Agents are definitely already moving to the marketplace," he said. "We're exploring ways to add additional agentic capabilities. In the meantime, our partners are continuing to list a good variety of agents."
To date, agentic AI listings on the AWS Marketplace typically complement a software partner's previously available SaaS offerings.
That software angle is also present in Salesforce's AgentExchange, a subset of the company's AppExchange marketplace that's oriented to agentic AI. AgentExchange debuted in March.
"Right now, we are mostly focused on technology partners on the AgentExchange," said Brian Landsman, AppExchange's CEO and executive vice president of global partnerships at Salesforce.
Currently, those partners participate in AgentExchange in two ways. First, they use the company's AgentForce platform to build and deploy AI agents or create tasks that agents can perform. Second, partners focus on integrating other technologies, such as Box and DocuSign, with AgentForce.
Landsman said the first approach provides a marketplace on-ramp for professional services partners.
"I do think we will see more consulting partners start to list capabilities on the AgentExchange over time, just as they did on the AppExchange," he said.
Service providers are branching out from generative AI to agentic AI, using online marketplaces to pursue autonomous action versus content creation.
IT services partners take on agentic AI
This shift has already begun, as consultants and integrators begin to offer agentic AI in marketplaces. Accenture, Presidio, Rackspace and Slalom are among the service providers providing AI agents on AWS Marketplace.
Presidio, a technology services and solutions provider based in New York, offers its Presidio AI Agent Offering in the marketplace and plans to add more agents in July. The company's Human AI Build, AI capabilities for software application development, is also on the marketplace.
The AI offerings "are starting to really pick up with our customers," said Kevin Corace, Presidio's senior vice president, software and services lifecycle.
Justin Copie, CEO at Innovative Solutions, a systems integrator and AWS partner based in Rochester, N.Y., said it has built thousands of AI agents for customers and internal uses in recent months. He expects the company's AI agent work to grow rapidly on the AWS Marketplace, overtaking its SaaS business.
"This time next year, the AI agents that we're building and enabling through the marketplace will be a bigger business for us than the software-as-a-service business that we resell," Copie said.
IT services partners, meanwhile, have also landed on Google Cloud's AI Agent Marketplace, which launched in November 2024 under its former name, AI Agent Space. That marketplace, a section of the Google Cloud Marketplace dedicated to AI agents, includes offerings from Accenture, Deloitte, Wipro and other companies.
Agent pricing in flux
IT service partners developing AI agents need to determine how to charge for their offerings, but pricing models in the agentic AI world have yet to become clear.
"There is a whole bunch of work being done on what is the perfect price model," Teigland said. "I don't think we have a great answer yet."
"There's lots of different ways that you can build the economic structure around it," Copie said of agentic AI.
Indeed, several methods are currently in play. Landsman said one approach is packaging agentic AI as part of a SaaS subscription. He also mentioned that agentic capabilities could be available in a premium pricing tier.
Another pricing method is consumption-based or metered pricing. Here, fees might be based on the number of tasks an agent completes.
Corace said such transaction-based pricing will need guardrails around it, noting customers don't want an open-ended checkbook for AI agent pricing. He thinks agentic AI pricing models will eventually build off approaches that customers are already familiar with, such as subscriptions. In that scenario, a subscription model could be designed to cover a designated number of transactions and provide burst pricing as the customer's use of agents grows, he added.
Teigland said she believes AI agent use will be tied to "value-driven outcomes" and kept under close control. In this approach, she noted that AI agents might be offered under a subscription-like service in which the agentic workforce is monitored for risk, governance and responsible AI.
John Moore is a writer for Informa TechTarget covering the CIO role, economic trends and the IT services industry.