Disparate Integrations, Compliance Gaps: Why AI Blueprints Are Struggling Against Agent Sprawl

In an effort to secure a future rooted in AI, many enterprises are inadvertently rewinding back to the past. Before cloud transformation and SaaS eclipsed on-premise software and infrastructure, company data was often siloed by default because interconnectivity between departmental functions was either too cumbersome, or impossible.

Once cloud tools became dominant, however, point solutions by necessity morphed into point solutions imposed by vendor lock-in and inflexibility by design. In the pursuit of fast utility, many businesses set themselves up to incur increased costs down the road, with few options to roll back the tech sprawl that had become so ingrained.

For all of the infinite potential agentic AI has to offer, the current rush to deploy intelligent agents risks the very same information silos that created a generation of technical debt and enterprise-wide fragmentation.

This warning comes from Manoj Mohan who joined us for an interview to discuss the risks of siloed agent sprawl in the enterprise. Mohan spent over two decades on the front lines of automation as Head of Engineering for Data Platforms and Insights at Intuit, and previously Meta, Apple, and Cloudera.

• Debt by a thousand agents: "It's like the early days of the cloud. You feel incredible velocity at first, but after a year or two, your costs have exploded and it's too late to easily fix the architecture. The same problem is now set for a ripple effect, and the risk in integrating disparate systems later comes with a steep price."
• Out of touch, out of compliance: Mohan cautioned against a new and insidious form of technical debt, as well as a governance disconnect that can lead to direct compliance violations. "For example, if a customer makes a ‘do not contact’ request under GDPR or CCPA and that information doesn't reach a siloed sales agent, the company faces a breach of trust, or worse."

Mohan's solution is a proactive architectural approach centered around an Enterprise Architecture team that is given enough latitude to define a blueprint the entire company sticks to. The three-layer solve prevents agent chaos by first cataloging all agents in a shared registry for visibility, then standardizing their communication through a common protocol, and finally controlling their actions with automated governance-as-code. But the architectural blueprint cannot exist in hypothetical isolation; it requires reusable technical foundations for agents to interoperate.

• The common tongue: Frameworks like the Model Context Protocol (MCP) and Agent-to-Agent are emerging to create a shared standard for agents to communicate and coordinate. "The great advantage of agentic workflows is their ability to loosely couple and exchange information across systems," he notes. "You don't want to tie that to a hard-coupled API. In the future, I see more systems moving towards flexible, semi-structured event payloads, which will allow them to evolve and adapt more easily."
• Build once: Mohan drew a parallel to workflows he saw firsthand at Intuit, where business units like QuickBooks and Credit Karma all needed common functions like payment processing. Instead of each building their own version, they built a single, configurable platform with fewer engineers. "That’s the model for AI as well," he said. "While the business context of each agent is unique, the technical implementation, the high-level design, and the architectural flows must be consistent and seamless." This efficiency reduced developer workload to the point of a task that previously required up to 25 engineers could now be accomplished with 10.

Inevitably, internal pre-planning eventually needs to meet customers where they are.

• Productivity's two faces: Mohan argues that a proper AI strategy delivers a powerful one-two punch by solving for two distinct personas. For the internal "developer persona", AI tools can amplify productivity, allowing teams to build and ship features faster. But the real transformation comes from serving the "customer persona" with that newfound engineering productivity. In one use case, Mohan's team used agentic workflows to automate complex manual tasks that customers were previously forced to handle themselves. "If we can build automation for something like financial transaction reconciliation that was a manual burden, it will instantly make the customer happy." When groundwork is laid in a repeatable fashion, engineering teams benefit from efficiency while customer benefit from faster features being shipped.

Despite a clear path forward, Mohan offered a sober prediction. Most companies, he believes, will get it wrong first. "AI transformation is happening at such a rapid pace that even hardcore technologists are finding it hard to keep up. If engineering leaders can't keep up, how can we anticipate that larger enterprises will get this right on the first iteration?"

The pressure is immense, especially as prominent leaders like Salesforce’s Marc Benioff make bold claims about massive AI-driven productivity gains and human worker replacement. Mohan predicts many companies will try to "put some numbers out there to buy time" as they learn. "My hunch is that most non-tier-one enterprises will only solve for agent sprawl in a second or third iteration, after some problems have already started to percolate."

It should be clear by now that AI for the sake of AI is not a viable strategy, but the obvious is not always obvious when complex externalities pressure businesses to act less than rationally. "The pragmatic answer, unfortunately, is many leaders are still trying to find how to prove AI as a core part of their strategy. But AI is not just part of your strategy. If you bring that lens, you're never going to get the best value. It has to be holistic," Mohan said. Proactively planning the chain of ownership, managing stakeholder communication, and getting the fundamentals of orchestration right from the outset is a recipe for minimal risk and maximum probability of success.