AWS technology partners are rearchitecting products around agentic AI as enterprise customers demand more flexible, outcome-driven software experiences.
In three interviews with Channel Insider, leaders from AWS, Zendesk, and Dynatrace described a market moving beyond early AI pilots and into production-grade deployments, where customers expect agents to support workflows, resolve issues, optimize systems, and integrate across business applications.
Agentic AI pushes software partners beyond basic chatbots
Allison Johnson, who leads Americas technology partnerships at AWS, said the shift marks another major software transformation for the cloud provider, following the move from on-premises infrastructure to cloud and then to SaaS.
Now, she said, partners are reconfiguring products so customers can interact with them “in an agentic fashion.”
For many ISVs, that means more than adding a chatbot. Johnson said partners are rethinking development strategies, the balance between AI and humans, and how tooling interacts with environments as AI becomes embedded across product portfolios.
“It really is truly rethinking the customer experience working backwards from the customer,” Johnson said.
The emergence of outcomes-based pricing and Marketplace selling
Beyond product architecture, vendors are also rethinking how AI software is sold. Johnson said AWS partners are increasingly exploring consumption- and outcome-based pricing models.
Tim Marsden, who leads Zendesk’s AWS partnership, said customers are under pressure to adopt AI to improve customer service, but pricing models must also evolve.
Zendesk introduced outcome-based pricing in 2024 for AI-driven resolutions, meaning customers pay per resolution conducted by an agent rather than treating autonomous agents like traditional seats.
“If you’re charging for an agent’s seat, and suddenly you’re rolling out 100 or 1,000 AI agents, does that qualify as a seat? It doesn’t make sense anymore,” Marsden said. “Customers only pay for verified resolved issues.”
Distribution channels are also evolving alongside the technology.
Johnson said AWS Marketplace’s AI agents and tools category has expanded to approximately 3,000 listings within a year of launch, while consulting partners increasingly package and sell agentic solutions alongside software vendors.
The trend suggests that channel opportunities extend beyond software resale into agent development, deployment, and ongoing optimization services.
READ MORE: We spoke with AWS and two of its channel partners about the demand for agentic AI tooling on the Marketplace in March.
Zendesk leans into AWS partnership
Zendesk is one example of that shift. The customer experience company has launched AI-focused offerings including its Resolution Platform, autonomous service agents, Agent Builder, custom agents, and a contact center offering built with Amazon Connect.
Zendesk’s relationship with AWS has also expanded from a standard ISV relationship into a strategic partnership.
Marsden said Zendesk is built on AWS, uses Amazon Bedrock to power AI capabilities with Anthropic Claude models, and is one of the few partners reselling Amazon Connect as an embedded offering.
He said Zendesk expects roughly $200 million in AWS Marketplace transactions this year and has ambitions to become a $1 billion AWS Marketplace partner, an increasingly common yet still impressive milestone that would place it amongst the likes of ServiceNow, Zscaler, CrowdStrike, and Splunk.
Dynatrace says AI adoption depends on trust, cost, and observability
Dynatrace is approaching the AI shift from the perspectives of observability and reliability. Raj Rajaram, Dynatrace’s vice president of global partnerships, said agentic AI changes the operational model because traditional applications were largely deterministic, while AI-driven workflows are probabilistic.
That makes visibility, explainability, and transparency more important, particularly as enterprises move agents from experimentation into production.
“If I’m going to summarize, it’s really about reliability, safety, and cost efficiency,” Rajaram said. “I need to trust the system to deliver what I need.”
Rajaram said that the need for trust is becoming more urgent as agentic systems introduce more complexity into enterprise environments.
Unlike traditional applications, AI-driven systems can create downstream risk from a single poor model response, hallucination, or failed agent interaction. That makes observability less of a back-end monitoring function and more of a required layer for AI governance, he said.
Dynatrace is positioning its AWS work to help customers detect drift, surface root causes, and resolve problems before they reach production. The companies are working across agentic development and operations, including integrations around AWS DevOps Agent and AWS AgentCore.
Rajaram said Dynatrace’s role is to help customers auto-protect, auto-remediate, and auto-optimize AI-enabled systems.
Channel partners begin packaging agentic AI for customers
The shift also creates new opportunities for channel partners.
Rajaram said channel partners are beginning to package AWS and Dynatrace capabilities together to help customers “demystify the entirety of your agent development lifecycle.”
Johnson said AWS is also seeing consulting partners, including Accenture and Deloitte, bring agents to market through AWS Marketplace alongside ISVs.
For ISVs and other channel partners, the opportunity goes beyond building AI features. It is helping customers decide where AI belongs, where humans remain essential, how to control cost, and how to move from experimentation to production without losing trust.
Rather than replacing software vendors, AI is forcing them to rethink how they build, sell, and monetize their products.
“As far as the SaaS apocalypse, I think it’s just a huge opportunity now, and that’s what I hear from my partners,” Johnson said.