Agentic AI is flooding enterprise pipelines with possibility, but regulated industries are learning that the speed of deployment and the discipline of governance cannot be decoupled — and that AI process orchestration may be the only architecture disciplined enough to hold both. For financial services firms in particular, the pressure to modernize is running headlong into the obligations of oversight.

That tension is reshaping how technology companies position their platforms. Process-centric AI is emerging as the architectural answer — pairing the probabilistic power of AI with deterministic process structures that keep outcomes auditable and reversible, according to Matt Calkins (pictured, right), co-founder and chief executive officer of Appian Corp.

“Agents are an incredibly powerful tool, but they have to be used correctly,” Calkins said. “Agents actually need process more than any other form of AI. They’re the ones that go off the rails the easiest if you don’t have the regulations, the guardrails, the tracking [or] if you don’t have the structure of process. … You need [them] when you need adaptive intelligence. You don’t need [them] the rest of the time.”

Calkins and Mal Cullen (left), chief executive officer of CIBC Mellon, a joint venture between The Bank of New York Mellon and Canadian Imperial Bank of Commerce, spoke with theCUBE’s Dave Vellante and Alison Kosik at Appian World 2026, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed AI process orchestration, responsible agentic deployment and how CIBC Mellon is using the Appian platform to modernize operations without exposing clients to unmanaged risk. (* Disclosure below.)

AI process orchestration keeps financial services governance intact

For financial services firms, that governance imperative is not theoretical — it shows up in the actual architecture of how AI gets deployed. The stakes are particularly high in asset servicing, where a single ungoverned AI deployment can ripple across trillions in assets and client scrutiny of AI practices is intensifying week by week, Cullen noted.

“The risk appetite that we have to live with, with the regulators … if we had an ungoverned or irresponsible AI deployment, it is very, very high,” Cullen said. “The only place we’re actually using [AI] in a production mode is with Appian right now. The reason for that is because they have a very controlled governance model where I can provide transparency to how it’s being used. I can audit that. I can have documentation and it’s not impacting a client yet — it’s being delivered for an internal purpose.”

Appian’s approach to AI process orchestration centers on keeping humans in the loop before deployment goes live. Rather than allowing AI to build or execute autonomously, the platform asks AI to first surface every rule, interface and data structure for human review — then proceeds only on explicit approval, Calkins explained. That model proved critical for CIBC Mellon when a Canadian regulatory change required new cost-transparency reporting for mutual funds and exchange-traded funds by the end of 2025.

“If I had gone to the global BNY platforms and asked for changes to support a Canadian regulatory cost reporting challenge, the likelihood of getting that done in time would’ve been low,” Cullen said. “My team built that on Appian. The clients now have a client-facing application to deliver that. The team told me they would never have made that date if we didn’t do this on Appian.”

Here’s the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of Appian World 2026:

(* Disclosure: TheCUBE is a paid media partner for Appian World. Sponsors of theCUBE’s event coverage do not have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE