Unlocking Hidden Value: When it Comes to AI, System Boundaries Are Where the Money is
- David Golub
- Nov 5
- 6 min read
Updated: 4 days ago
For leaders looking to amp up AI strategy, an axiom from systems thinking holds a cautionary tale: every organization is perfectly designed to produce the results it gets.
This insight has driven decades of quality improvement, yet it also contains a difficult truth. Your current performance isn't (just) a creature of execution quality, it's the natural output of how your systems interact, or rather – how they don't.
This difference matters, because when it comes to AI, the biggest gains lie beyond your well-mapped processes. As we’ve written elsewhere, ROI lives in the white spaces, the system boundaries created back when coordination was too expensive to happen.
Agentic AI changes the game, but unlocking coordination opportunities begins by reflecting on how value got trapped in the first place.
The Blindness of Optimization
Six months into your AI initiative, you've automated twenty tasks. Hours saved: 500 annually. Revenue impact: not much.
Sound familiar?
Too many AI strategies initially start by asking, "Which can we automate?" Teams identify repetitive work, score it by time saved and deploy agents to execute faster. The problem isn't likely delivery. It's the question they started with.
Most AI strategies begin with an audit of existing processes. Teams identify repetitive tasks, score them by hours saved and deploy automation. The math seems compelling: automate 100 tasks saving 10 hours each, multiply by loaded cost, report ROI.
This approach may be necessary but it is not sufficient. When you examine what happens within systems, you can only improve what already exists. You speed invoice processing , reduce order fulfillment errors and accelerate report generation.
You’ve applied WD-40 to the current business model, but haven't invented a better one.
As Pascal Bornet and colleagues have explained, in an era of transformation, optimization plateaus. You’ve perfected what exists but haven't created new capabilities or opened pathways to getting stuff done in the real-world.
The ROI difference is categorical. Deloitte, for example, has reported 30% efficiency gains with process automation, while Amazon yielded a 25% increase in order processing speed thanks to warehouse robotics.
Approaching AI as yet another tech optimization misses the chance to think bigger. When you optimize, you aren't in the headspace of imagining the valuable activities that never happen at all. With AI, you need to be.
Traditional process analysis won’t surface this opportunity. It finds plenty to optimize within each system, such as faster quotes in sales, better routing in logistics, improved forecasting in inventory, but doesn’t take the next step.
Boundaries Are Where the Money Is
Disruptive capabilities exist at the boundaries, in the messy handoffs, human effort and format inconsistencies that keep systems separated.
Systems thinking reveals that value emerges from interactions, not components. Yet many leaders take the view that AI strategies should focus on improving individual pieces and miss the big picture.
Optimization is rational. System boundaries exist for good reasons. They encapsulate complexity, enable specialization and allow independent evolution. Sales can upgrade its CRM without waiting for logistics. Finance can refine forecasting without coordinating with product development.
This modularity enabled modern enterprise scale, after all.
But these boundaries also create what systems theorists call "structural coupling" where each system optimizes for its local environment while growing progressively disconnected from adjacent systems. The boundary becomes a barrier.
Within a system, information flows naturally. Your CRM knows everything about customer interactions. Your ERP captures all operational data. But between systems, information doesn't flow: it must be actively transferred, translated, reconciled.
This friction isn't just technical; it's economic and organizational, and lurking in the shadows, it shapes your entire business model.
Friction determines what questions you can answer (average deal size is easy; which operational constraints cost the most in lost deals is nearly impossible). It defines viable strategies (standardized offerings scale; dynamic customization doesn't) and it constrains your company to what's possible given coordination costs.
Most critically for AI strategy, friction becomes invisible.
Compounding Coordination Costs
Think how this plays out in your own organization: when coordination costs are high enough, organizations stop considering certain capabilities altogether.
Bold ideas vanish from strategic planning, competitive analysis and investment decisions, creating a kind of "possibility blindness" that sets leaders up to miss out on the biggest AI opportunities.
Ask your team what AI could do for your company. They'll likely mention automating customer service, accelerating data analysis or improving demand forecasting.
Now ask about enabling real-time coordination between customer support and product development, or dynamic pricing based on operational capacity. These capabilities often won't even surface, and not because they lack value, but because they've been impossible for so long no one dares believe in them.
Anopsia compounds. Sales develops "standard packages" that become "our business model." Training programs reinforce them. Compensation systems reward them. Marketing promotes them. What began as a constraint becomes a culture.
Consider a scenario every B2B company faces: Your sales team quotes standard delivery terms while competitors customize theirs, winning deals you lose. This isn't inefficiency within your sales process, it's an absence of capability.
The kind of real-time coordination between sales, inventory and logistics that would enable dynamic customization doesn't exist because historically it would have takes days of manual work to accomplish.
Each accommodation spawns secondary processes that further obscure the original coordination defect. The organization literally cannot see the opportunity anymore, since it's been institutionalized away.
Your AI strategy inherits these blind spots, and they're likely where the gold is.
Agents Change the ROI Equation
Agentic AI resets coordination economics. Agents have the ability to maintain continuous context across systems, translate between incompatible data models and orchestrate complex multi-system workflows in real-time.
Coordination that once required days happens in milliseconds.
But here's where the strategy comes in: agents don't just make coordination faster, they make previously impossible interaction patterns economically viable.
This isn't automation of existing processes; it's activation of latent capabilities.
Consider the difference: An optimization approach might deploy chatbots to answer questions faster. An agentic approach enables agents that correlate tickets, identify revenue impact, check development capacity and adjust product priorities.
The first approach saves costs. The second creates a new capability. The first improves your current business model. The second launches a better one.
This pattern repeats itself across traditional silos:
Sales-to-legal coordination enables instant custom contracts (new capability) vs. faster standard contracts (task automation)
Operations-to-finance coordination enables dynamic resource allocation (new capability) vs. faster monthly reporting (task automation)
Service-to-product coordination enables continuous evolution (new capability) vs. faster ticket summaries (task automation)
These are meaningful gains, to be sure, but they are not new ventures, Coordination across boundaries creates new revenue models with untapped multipliers.
Start Finding Your Hidden Value
To identify where agentic AI will generate the highest ROI, you need to map your organization, but differently than your might of done during your previous offsite white boarding sessions.
Instead of process flows within departments, map flows that should exist but don't. Instead of measuring task time, measure decisions made with incomplete information. Instead of identifying repetitive work, identify key actions that don't happen.
Start with competitive losses. Have competitors enabled coordination you haven't? Do they quote faster because sales and legal systems coordinate in real-time? Do they personalize more meaningfully because production and sales share context? These are your boundary opportunities.
Inspect “artificial rhythms” in your business. Quarterly business reviews exist because that's how often humans can synthesize cross-functional data. Monthly forecasting happens because that's the coordination cadence finance can sustain.
These temporal boundaries, dictated by coordination costs, represent continuous value loss. But agents could enable continuous rather than periodic coordination.
Unearth the familar workarounds that have become SOP. Standard packages, approval delays, information requests, escalation procedures: these often mask coordination failures that agents could solve.
Each compromise represents a hidden value waiting to be unleashed.
A Mirror for Your Blind Spots
Winning AI comes from enabling valuable coordination your competitors still consider impossible, moving past automation to discover capabilities that couldn't exist when coordination was expensive.
Enterprises need a structured approach that bridges the gap between AI experimentation and business transformation. This is where Agentic Foundry's FORGE framework earns its salt.
By deliberating over five core activities (which may depend on company readiness),
FORGE helps companies Find the Friction, Organize the Work, Replicate Intelligence, Generate Feedback and Empower Teams.
FORGE is a practical toolkit to dissolve yesterday's boundaries and refocus coordination where it creates the most value, visualize agentic intelligence and scale human-in-the-loop solutions across the enterprise.
The boundaries you've accepted as fixed are full of potential. Shouldn’t you find it before your competitors do?
Agentic Foundry: AI For Real-World Results
Learn how agentic AI boosts productivity, speeds decisions and drives growth
— while always keeping you in the loop.
