Why Collaborative Intelligence Is Now a Core Delivery Skill

/ AI, Application Development, DevOps

The News

In Episode 93 of the AppDevANGLE podcast, Paul Nashwaty sat down with Barry Rosen, CEO of Interaction Associates, a training and consulting firm specializing in collaborative intelligence. The conversation centered on a striking statistic from ECI Research’s 2026 application development research: 65% of engineering organizations report that their teams spend 20% or less of their time on net new innovation. Rosen argues that the root cause isn’t tooling or technical debt alone, but the organizational collaboration gaps that no platform can fix on its own, and that the project manager role must evolve to address them.

Analyst Take

The Real Productivity Crisis Isn’t in the Code

If the vast majority of development capacity is consumed by maintenance, compliance overhead, and coordination friction, the AI coding assistant conversation is almost beside the point. You can automate 30% of your code generation and still hemorrhage velocity if your teams are spending half their sprints resolving stakeholder misalignment or waiting on decisions that never quite got made in the last meeting.

This is the tension Rosen surfaces clearly: organizations have been treating a people and process problem as a tooling problem. The instinct to reach for a new platform or a new CI/CD pipeline is understandable, and often partially correct. But ECI Research’s Google GovTech Survey found that 48.0% of respondents selected “Navigating compliance documentation and audit evidence collection” as the greatest source of cognitive load for their developers today. That is not a problem a better IDE solves. It’s a coordination and process problem that lands squarely in the domain Interaction Associates is addressing.

The Project Manager as Collaborative Infrastructure

Rosen’s framing of the modern project manager as “a manager of a collaborative process over time” rather than a schedule tracker is analytically useful, not just rhetorically interesting. What he’s describing is an emerging function that sits at the intersection of stakeholder politics, delivery velocity, and increasingly, human-AI teaming. As AI agents take over more of the administrative layer of project management, scheduling, summarization, status reporting, the residual value of the human PM concentrates in exactly the places AI performs worst: reading political dynamics, building offline agreements, and challenging group thinking before it hardens into bad decisions.

The career implications are real. Project managers who invest in facilitation as a core discipline are positioning themselves for roles that are structurally hard to automate. Those who don’t are managing tasks that increasingly can be automated. That’s a binary outcome, and organizations that treat facilitation training as a soft-skills elective rather than a delivery-critical investment are taking a concrete risk.

Why Procurement and Org Design Both Miss This

There’s a broader structural issue here worth naming. Government and enterprise IT shops alike tend to budget for technology and contractor labor, but rarely for organizational capability building of the kind Rosen describes. ECI Research’s Google GovTech Survey found that 56.0% of respondents said procurement or contractual requirements “frequently” force engineering teams to use suboptimal developer tools, because approved vendor lists lack modern developer platforms. The irony is that even when organizations do secure better tooling, the collaboration deficits Rosen describes mean teams often can’t capitalize on what those tools offer. A faster deployment pipeline doesn’t help if the feature requirements were never actually agreed upon by the right stakeholders.

Interaction Associates occupies a market niche that sits well outside the typical enterprise software buying motion, which is part of why it often goes unfunded. Training and facilitation consulting are discretionary budget items in most organizations, cut first when program budgets tighten. But the cost of not investing shows up directly in the innovation capacity data Nashwadi cited: teams spending 80% of their time on everything except building new things.

Looking Ahead

The near-term trajectory here points toward a structural revaluation of human facilitation skills as AI agents become embedded in meeting and delivery workflows. Rosen’s prediction that meeting participants will soon interact with AI avatars in real time is plausible within a two-to-three year horizon, and it sharpens the case for human facilitators who can manage that hybrid dynamic rather than cede it entirely to the machine. Organizations that have already invested in collaborative intelligence as a capability will adapt faster; those that haven’t will find their AI tools generating outputs that no one actually agrees on.

For Interaction Associates, the strategic opportunity is to position collaborative intelligence not as a training product but as a delivery accelerator with measurable ROI tied to decision cycle time and innovation throughput. The data is available to make that case. If 65% of engineering teams are spending most of their time on non-innovation work, and a meaningful portion of that waste traces to coordination failures rather than technical blockers, then the return on building facilitation capability becomes quantifiable. The firms that learn to frame it that way will find enterprise buyers far more receptive than those still selling soft-skills workshops.

Author