Melt Research/Executive brief

Everyone is building AI.
Who is rebuilding the business?

Building AI is getting easier. Organizing a business around it is not.

October 2026

AI can now build in a night what organizations take months to absorb. Companies are experimenting, building agents and automating tasks at unprecedented speed. But building an agent is not the same as making AI work across a business.

The real test comes when AI has to operate in the complexity of the business itself. Can it give reliable answers when conditions change? Can it help people make better decisions, faster? Can the organization demonstrate that it is improving performance? And who is accountable for the results?

These questions become particularly important in manufacturing and supply chain, where a change in customer demand, a supplier delay or a production constraint can trigger decisions across multiple teams. Responding requires more than access to data. It requires understanding how the operation works, what is affected and what can be done about it.

01The knowledge gap

The business runs on experience that its systems don’t contain.

Enterprise systems hold orders, inventory, transactions and operational updates. But they rarely capture the reasoning behind decisions: what triggered a change, which alternatives were considered, or why one option was chosen over another.

In manufacturing, for example, changing sales forecasts can trigger thousands of adjustments to orders, delivery dates and supplier coordination. When materials become constrained, experienced planners decide which customers to prioritize, which orders can wait and how production should be reshuffled.

The systems contain the data. The judgment behind the decisions often lives with the people who make them.

02The value gap

Finding opportunities is easy. Proving business impact is harder.

Organizations can identify AI opportunities across functions, from automating routine tasks to improving complex operational processes. But deciding where to invest, estimating the full cost of deployment and proving the resulting business impact are much harder.

A successful prototype doesn’t establish whether a solution is reliable enough for production or whether it will improve the measures that matter: cost, margin, revenue, delivery performance or service levels.

The challenge is connecting AI initiatives to business priorities and measurable outcomes, rather than stopping at adoption, usage or hours saved.

03The ownership gap

AI is changing how work gets done. Who is responsible for the results?

Central AI teams can provide expertise, direction and governance, but they cannot own every business process. The people closest to the work understand its priorities, constraints and the consequences when something goes wrong.

In manufacturing, an AI recommendation can affect purchasing, production schedules, inventory and customer commitments. As AI moves across departments, organizations need clarity on who approves its decisions, who validates its recommendations, who oversees its performance and who is accountable for the outcome.

Without clear ownership, deploying AI can create new risks rather than improve how the business operates.

The next phase is an organizational challenge.

The companies that make the most of AI will not necessarily be those that build the most agents. They will be the ones that make operational knowledge accessible, connect AI investment to measurable business performance, and establish clear ownership for decisions and outcomes.

That requires more than a technology rollout. It requires a different way of organizing the business around AI.

Go beyond the executive brief.

Explore the full research into how organizations are structuring AI transformation, assigning responsibility and building the capabilities to make AI work across the business.

Get the full research report

About Melt

Melt is building the decision layer for AI in Operations, Manufacturing & Supply Chain.

Today, decisions about suppliers, materials, and production determine whether factories run on schedule or lose time and money. Yet the context needed to make those decisions is scattered across ERP systems, spreadsheets, emails, and employees’ experience. As a result, disruptions take too long to resolve, costly problems recur, and critical expertise remains trapped with individuals.

Melt reconstructs how operational decisions happen, identifies where intervention can make the biggest difference, and deploys AI agents to help teams act before small issues become costly delays. We’re turning the operational know-how that runs complex businesses into an AI workforce that can scale across the enterprise.

About this research

This brief draws on Melt Research conducted in 2026, combining analysis of public material on AI transformation roles with conversations with AI transformation leaders and independent consultants across Israel, the United States and Europe. The research includes perspectives from operations-heavy industries such as manufacturing, supply chain, food and agriculture, electronics and semiconductors, alongside technology, finance and professional services.

The findings reflect patterns observed across the research and are not intended as a representative survey of all organizations.