About This Episode
In this episode of Stratola Spectrum, Dinesh Chandrasekhar, Chief Analyst, Stratola, sits down with Massimo Pezzini, one of the most influential analyst voices in application integration. Massimo spent decades at Gartner, helped define iPaaS early, and now leads research at Workato.
The conversation frames a key shift happening right now: for nearly 30 years, integration and automation have progressed in incremental steps. New tooling arrived (ESBs, API management, BPM, RPA, iPaaS), but the rules of the game stayed largely the same. Massimo argues that agentic automation changes that pattern. For the first time, the industry has a genuine chance to automate non-deterministic processes, meaning workflows where the same input can lead to different outcomes depending on context, history, and dynamic conditions.
Dinesh challenges the “new kills old” narrative. Massimo explains why legacy and established automation stacks rarely die due to install base realities. Instead, markets accumulate capabilities, then converge. He lays out a clean integration model with three core problems: data consistency, process automation, and composition (APIs). Historically, different products served each problem, forcing enterprises to manage many platforms. Over the last decade, the industry has moved toward convergence, creating Swiss-army-knife platforms (especially iPaaS) that cover multiple use cases.
The episode then moves into AI’s role. Massimo positions AI and integration as symbiotic: integration enables AI by feeding it enterprise data and embedding it into business processes, while AI enables integration by reducing learning curves, accelerating testing, automating documentation, and improving reliability and governance. The future, in his view, is not the disappearance of APIs or integration platforms, but their evolution into more discoverable, agent-friendly interfaces with richer manifests and clearer semantics. Agent interoperability standards are emerging, but the winning protocol is still uncertain.
Finally, Dinesh and Massimo discuss what this means for system integrators. Massimo predicts SI value shifts away from selling coders and toward architecture, governance, delivery models, business case development, and deep vertical expertise. The episode ends on an open question: the real breakthrough will come from the “killer application” for agents, not just doing today’s work faster.
Key Takeaways
Agentic + orchestration is the first real “revolution” in this space because it enables automation of non deterministic processes. Classic automation is great for deterministic flows (same input, same output), but agents can handle context and variance in processes like sales or onboarding
Legacy integration tech is not getting killed, it is being layered and accumulated on top of the install base. RPA, BPM, API management, ESBs remain alive because enterprises have invested heavily and still need deterministic compliance grade automation.
Integration is still the same 3 problems, but platforms are converging into “Swiss army knife” stacks. Data consistency, process automation, and composition (APIs) used to require many tools, but iPaaS style convergence is simplifying the landscape.
Integration and AI are symbiotic: integration enables AI with the right data and workflows, while AI improves integration quality and usability. GenAI can shorten learning curves, auto document flows, accelerate testing, and push integration from “basement team” to more democratized use.
The hardest upcoming gap is cross-vendor agent interoperability, because we are heading into “agentic bubbles.” There is no clear winning standard yet; likely it will stay API and event based, but APIs must become more discoverable with richer manifests so agents can understand capabilities.
Big opportunity on the data side is governance and security automation, not just natural language SQL. Example: auditing agents that verify whether security and data governance policies are actually implemented across systems.
SI and consulting stays relevant but shifts from coding to architecture, governance, delivery models, and vertical playbooks. Enterprises will need help designing agent plus integration architectures, proving business value, and mapping workflows to real systems and APIs.
Killer app is still unknown, but it will be something that lets enterprises do things they could not do before at a reasonable time and cost. Efficiency gains alone are not enough; the winner will unlock new capabilities, not just faster execution of existing tasks.