Cover art for Map your workflows first to unlock AI's real value

Map your workflows first to unlock AI's real value

Companies that study how others redesigned around AI identify far more use cases and generate double the revenue, new research shows.

13 cards · 3m · tap to begin

From · · 3m

Map your workflows first to unlock AI's real value

Companies that study how others redesigned around AI identify far more use cases and generate double the revenue, new research shows.

In brief

Companies that study how others redesigned around AI identify far more use cases and generate double the revenue, new research shows. Research shows companies that learn from others' AI workflow redesigns identify 44% more use cases and generate nearly double the revenue. Originally reported by Harvard Business Review.

The AI adoption paradox

Nearly four years after large language models became widespread, most companies have adopted them but few have turned that investment into real productivity or performance gains.

The mapping problem

Researchers from INSEAD and Harvard Business School identified the core challenge: managers know AI can perform useful tasks but struggle to identify where and how it can reshape their businesses.

The field experiment

The field experiment

Five hundred fifteen early-stage companies worldwide received AI tools, training, and mentorship. Only half saw examples of how other organizations had redesigned workflows around AI.

Dramatic results from examples

Companies exposed to real-life examples identified 44% more AI use cases than the control group, completed 12% more tasks, were 18% more likely to acquire paying customers, and generated nearly double the revenue.

Map existing workflows end to end

Examine how changes in one activity affect every step that follows. Look for handoffs, approvals, information translations, and decision stalls. Those transition points often offer greater opportunities than the tasks themselves.

Find the real bottleneck

Ask what is really slowing you down.The answer usually is not the most obvious task but a handoff or decision that holds up everything that comes after.

“Finding the bottleneck that's closest to your customers or closest to the outcome you care about is often the first step.”

Bottlenecks reappear elsewhere

When one activity accelerates, another becomes an unexpected bottleneck. Faster product development increases customer feedback volume. Automated reporting exposes data quality weaknesses. Organizations discover yesterday's constraint disappeared only to reveal another one farther downstream.

Treat AI as ongoing process

Many AI projects initially appear successful before business performance plateaus. Organizations solve one problem but leave the next one untouched. Continuously identify and remove new constraints to capture value.

Redesign the entire process

The highest-performing companies did not stop at identifying bottlenecks. They redesigned the work surrounding them. AI changes the economics of a process only when organizations rethink how activities fit together.

Measure impact, not activity

Managers should ask whether AI-enabled workflows allow them to reach more customers, produce better products or services, or require fewer resources. Otherwise they are measuring activity rather than impact.

“They should ask themselves whether the AI-enabled workflow allows them to reach more customers, produce better products or services, or require fewer resources than before.”

Allow people to experiment

The best-performing companies learned how others reorganized around AI but did not merely copy their ideas. Managers considered what was possible inside their own businesses, leading to experimentation and new processes.

International Motors' approach

Robert Oh, chief digital and information officer at International Motors, a one hundred twenty-four-year-old Illinois-based manufacturer, initially saw dozens of AI experiments improve individual tasks but not the business itself.

From tasks to business value

International Motors mapped cross-functional workflows, identified twenty-one opportunities spanning sales, manufacturing, supply chain, and service, and prioritized five with the biggest impact.

What it means

Research shows companies that learn from others' AI workflow redesigns identify 44% more use cases and generate nearly double the revenue.

Read the original on Harvard Business Review

Read the TL;DR — this story on one card.

How we make summaries

React

Sign in to react and comment.

Comments (0)

Life is short. Keep it sweet. Respect others' opinions and be kind!

    Recommended next

    More decks on artificial intelligence and related topics.