Bridging the AI Value Gap: Our Insights from IBM TechXchange 2025

We took a chance. WWT's Infrastructure Automation team had never attended  IBM TechXchange, and we neither knew what to expect nor if we would learn anything relevant to our daily work — helping organizations implement and gain value from practical, actionable IT automation. 

Our biggest takeaway? We would have missed out if we hadn't gone. We left with a clearer understanding of one of the most pressing challenges facing IT leaders today: the gap between their AI investment and making that juice worth the squeeze. More importantly, we learned that automation isn't just adjacent to AI success, it's fundamental to it. 

Without effective automation, AI remains an expensive experiment. With it, AI becomes a practical tool that delivers actual, consumable business value. Here are some highlights of what we learned.

The uncomfortable truth about AI adoption

While over 90% of organizations are increasing AI investments, only 25% report seeing usable, valuable results. Industry analysts predict 30% of generative AI projects will be abandoned after proof of concept by the end of 2025.

From our point of view, this isn't a technology problem; it's an automation problem. Organizations lack the automation foundations to operationalize AI insights. That means AI-driven analysis doesn't connect to business processes. The result: impressive demos that never reach production.

The real issue is, organizations are trying to add AI to manual processes rather than building automated workflows where AI drives action. Without automation to act on insights, nothing changes.

Making AI accessible through automation

The Model Context Protocol (MCP) emerged as the standard for connecting AI to automation workflows. Released by Anthropic in mid-2024, MCP provides a consistent interface that allows AI to feed directly into automation platforms, accessing data and triggering automated actions based on intelligent analysis.

Without MCP, every AI-driven automation requires custom integration with every data source — a scaling nightmare. With MCP connecting AI to automation platforms, organizations can automatically run remediation workflows, trigger pre-approved response procedures, and scale AI-driven automation across many practical use cases that can save IT staff vast amounts of time and toil while improving the health, stability, compliance and security of operational environments.

AI without automation is just expensive consulting. We saw demonstrations of MCP enabling AI to analyze monitoring data and automatically trigger infrastructure remediation, turning AI insights into automated business results.

Democratizing automation: From IT to everyone

The organizations seeing the most success have made automation accessible to domain experts across the business through platform engineering — self-service platforms with pre-built workflows, role-based access, built-in approvals and integration with existing tools.

Cloud-native scaffolding systems, whether hosted internally or externally, provide reusable templates with security controls and best practices baked in. When a network engineer can automate routine changes through a self-service portal, when a security analyst can trigger automated incident response, that's when automation's value compounds throughout an organization.

Closing the loop: From observability to action

Traditional monitoring creates alert fatigue. The breakthrough is connecting observability insights directly to automated workflows — event detection, AI enrichment that correlates events and assesses impact, automated response with human oversight, and continuous learning.

We saw systems detect degraded performance, correlate it with configuration changes, automatically roll back, verify the fix and create incident reports — all within minutes. 

Nearly three-quarters of companies struggle to scale AI value because they can't operationalize insights. Closing the loop between observability and automation solves this.

Practical steps for IT leaders

So, what can you do with these insights? Here are a few ideas that we think aren't far-fetched, even for organizations in the early stages of their automation and AI journeys:

• Start with automation, add AI for intelligence. Identify processes to automate first, then determine where AI can make those workflows smarter and more adaptive.
• Adopt standards that connect AI to automation. Implement MCP into your automated workflows now to bridge AI insights and automated actions. The overhead is low, and it will save you the pain of retrofitting later.
• Invest in platform engineering. Build self-service automation capabilities with templates, guardrails and approval workflows that make automation accessible and safe for domain experts.
• Create closed-loop operations. Connect observability to automated responses where AI-identified issues trigger tested, automated workflows with human oversight for exceptions.
• Measure business outcomes. Track reduced incident resolution time, improved deployment speed and increased the value of AI investments. Instead of asking, "How do we use AI to grow, scale, improve, etc.?" Ask, "Is our automation delivering value, and is AI making it more effective?"

Looking forward

IBM TechXchange 2025 reinforced a critical insight: organizations that succeed with AI do so on the back of their automation foundations. They make their automation more capable and adaptive while increasing the capacity for their teams to innovate and evolve. The gap between AI investment and value exists because many organizations try to implement AI without automation infrastructure to operationalize it.

Here's what we learned: AI can replace knowledge — facts, procedures, algorithms — but not wisdom: judgment, creativity, organizational context. The most successful implementations keep humans in the loop for judgment while AI-enhanced automation handles repetitive work. When AI recommendations flow into transparent, tested automation workflows, organizations trust AI in production.

Automation is the foundation, AI is the accelerant. For IT leaders, the opportunity isn't choosing between AI and automation — it's recognizing that automation enables AI to deliver value. Build your automation foundations first; make automation accessible to everyone; and then layer on AI to make those workflows smarter, more adaptive and more valuable.

The future of IT is humans empowered by AI-driven automation, working together to solve problems that neither could tackle alone.