AI’s valley of despair: Why reengineering workflows is critical for ROI
Companies investing in AI are encountering a harsh reality: inflated expectations are colliding with modest returns, creating what industry observers call a "valley of despair." The disconnect stems from a focus on automating tasks rather than reengineering entire workflows to align with AI’s capabilities. Without addressing bottlenecks and measuring tangible outcomes, AI investments risk failing to deliver measurable business value.
Editor, Lazyfounder
TABLE OF CONTENTS
Why AI investments are falling short
A recent discussion on SiliconANGLE highlights a growing concern among AI decision-makers: only 15% reported an increase in EBITDA over the past year, despite widespread AI adoption. Fewer than one-third could directly link AI initiatives to bottom-line improvements.
The issue is rooted in Amdahl’s Law, a principle originally applied to computing but now relevant to business processes. It explains why accelerating one part of a workflow—such as drafting a proposal in seconds—fails to improve overall speed if bottlenecks like reviews, approvals, or deployment remain unchanged.
This gap between expectation and reality has led to what industry experts describe as the "valley of despair," where AI’s potential is undercut by inefficiencies in surrounding processes.
Beyond automation: The case for reengineering
The solution, according to Jeremiah Stone, Chief Technology Officer at SnapLogic Inc., lies in reengineering business processes rather than merely automating existing tasks. Automation speeds up individual steps, but reengineering questions whether those steps should exist at all.
Organizations often default to automating what they already do, rather than reinventing workflows to eliminate inefficiencies. For example, a coding tool might generate software in minutes, but if testing, security, and deployment take weeks, the overall process remains slow.
AI initiatives must target measurable improvements in revenue, efficiency, or risk. Stone emphasizes that enterprises should expect firm ROI data within six months, with pilots designed to test sustained use—not just capabilities in controlled environments.
How to measure and manage AI-driven transformation
To avoid the valley of despair, companies must treat AI as a series of targeted business transformations, not just a technology deployment. This means identifying bottlenecks, handoffs, and sources of rework before applying AI to a process.
Key metrics to track include cycle time, quality, adoption rates, cost per outcome, and financial impact. For AI-driven systems, particularly those involving agentic processes, monitoring both the range and average of operating costs is critical, as complex cases can consume disproportionate resources.
Early adopters are finding success by assigning business leaders ownership of AI outcomes, setting hard timelines for impact data, and continuously refining strategies based on cost and performance metrics.
Lessons from AI adoption struggles
Earlier Lazyfounder reporting on AI infrastructure partnerships, such as IBM and CoreWeave’s collaboration on workload controls, underscores the importance of aligning AI deployments with operational realities. Similarly, TOTVS’s expansion of its enterprise AI foundation, LYNN, highlights how integrating governed operational data can improve AI’s business impact.
These examples show that AI success depends on more than just technological advancements—it requires rethinking how work gets done. Companies that focus solely on automating tasks, without addressing underlying process inefficiencies, risk repeating the same mistakes driving the current disillusionment.
What this means
Lazyfounder analysis — our interpretation, not reported fact.
For founders and operators, the takeaway is clear: AI’s promise hinges on reengineering workflows, not just deploying tools. The 15% of companies seeing EBITDA gains from AI aren’t achieving results by accident—they’re treating AI as a catalyst for broader business transformation.
This means auditing processes to identify bottlenecks before implementing AI, setting clear ROI targets, and holding business leaders accountable for outcomes. Pilots should simulate real-world conditions, not idealized scenarios, and metrics like cost per outcome or financial impact should take precedence over vanity metrics like adoption rates.
The valley of despair isn’t inevitable, but avoiding it requires moving beyond incremental improvements and embracing fundamental change.
Related on Lazyfounder
Sources
- SiliconANGLE · 2026-10-23
How to avoid AI’s ‘valley of despair’
This story is an original summary drafted with AI by Lazyfounder from the reporting listed above and checked by automated validation. Facts are attributed to their original publishers; sections marked as analysis are Lazyfounder's. Where a source is in another language, facts were machine-translated and quotations are reported, not reproduced. Read the original coverage via the links, and see our AI policy and corrections policy.
About the author
Editor, Lazyfounder
Tarun Mottlia edits LazyFounders, covering Indian startups, funding rounds, AI and product launches. Every story on the site is AI-assisted and checked against its cited sources before publication.
More stories by Tarun MottliaGet the LazyFounder Brief
Startup, funding and AI news in a five-minute read. Join the early-access list.
