Introduction

AI’s promise is being priced in, and it is happening fast. McKinsey Global Instituteⁱ estimates that by 2030, AI-powered agents and robots could unlock about $2.9 trillion in annual economic value in the United States alone (with a big catch: it depends on redesigning work, not sprinkling tools on top…more on that later). And Gartner forecasts $2.52 trillion in worldwide AI spending in 2026ⁱⁱ, a 44% year-over-year jump, signaling just how aggressively organizations are investing in the foundations.

But the “value at scale” story is still stubbornly uneven. McKinsey’s 2025 global surveyⁱⁱⁱ finds that while 88% of organizations report regular AI use in at least one business function, only about one-third say they’ve begun scaling AI programs enterprise-wide. Deloitte’s 2026 survey^iv puts a sharper point on the execution gap: only 25% of respondents have moved 40% or more of AI pilots into production, and 37% report using AI at a surface level with little or no change to underlying processes. Gartner^v offers the bluntest reality check: by the end of last year, at least 50% of generative AI projects were abandoned after the proof-of-concept stage.

That gap between breathtaking potential and inconsistent realization isn’t a technology problem. It’s an execution problem.

A Brief History of Technology Revolutions 

We believe that the foundational obstacles to AI delivering business value at scale have nothing to do with technology. Instead, we assert that they have everything to do with optimizing the processes and organizations that must leverage AI technology to effect tangible business value.

This is nothing new; every tech revolution experiences a similar lag between the emergence of pure technology and its application to produce clear value. Think about these tech revolutions:

• Steam power: early adopters used steam mainly to replace human or animal manual labor to pump water out of mines. The real productivity boom came later, when engineers converted steam into reliable rotary power, and businesses redesigned factories and logistics around continuous mechanical energy, including (new layouts, new skills, and new infrastructure).
• Electricity in factories: early factories swapped steam engines for electric motors but kept the same line-shaft layouts—real productivity arrived when they redesigned the factory around distributed power (new layouts, new flow, new roles).
• Computers in offices: companies bought computers for years before productivity visibly popped, because value required process digitization + management changes, not hardware on desks.
• Internet + retail: having a website wasn’t the win. The real winners rewired fulfillment, inventory visibility, last‑mile delivery, customer service, and merchandising.
• Mobile telecom: the true power of this tech that's been evolving since the late 80s has only started to become apparent as an enabler of processes ranging from field maintenance to ecommerce to media consumption as a result of those processes evolving to make optimal use of (and guide evolution of) the form factor.

While most of the hype around AI focuses on the speeds and feeds of raw tech, much of the reality we see reported focuses on its inability to produce business results to commensurate with expectations. However, case studies are starting to emerge that demonstrate the importance of process and org adaptation in achieving expected results. Take the case^vi of UPS ORION, an early AI route optimization project. 

UPS built ORION^vii to compute optimized daily delivery routes using advanced algorithms and massive operational data. Early on, the math wasn’t the hard part—the routes were inconsistent and not “drivable” within real-world constraints, and frontline skepticism was high. UPS treated adoption as the main engineering problem: it integrated ORION into existing workflows, retrained leaders and drivers, and—crucially—changed performance measurement so teams weren’t judged against yesterday’s habits but against a new “ideal route” standard. The result wasn’t a clever pilot; it was a new operating rhythm. ORION ultimately saved 10 million gallons of fuel annually, cut CO2 emissions by 100,000 metric tons, and delivered $300–$400 million in yearly savings/cost avoidance once fully deployed.

What's Different This Time

The difference between past tech revolutions and AI is that past ones gave humans generations to catch up. That catch-up time is being compressed toward a lower limit of zero by AI. Consider these recent data points:

• In 2024,^viii  Google reported that more than a quarter of its new code was AI-generated and accepted into production. In 2025, ^ix Microsoft estimated that 20–30% of its codebase was written by AI. By 2026, ^x Anthropic indicated that AI was responsible for 70–90% of code company-wide, with its Claude Code system generating nearly 90% of its own code.

• On SWE-bench, a benchmark that measures how well models solve real-world coding problems, the ^xi Stanford AI Index recorded a dramatic surge in performance, from just 4.4% of problems solved in 2023 to 71.7% in 2024. That 16x leap in a single year is the kind of acceleration rarely seen in prior technology cycles.
• Benchmarks show that ^xii Claude Mythos correctly solves 83% of security scenarios—a 25% improvement over Opus 4.6, released just two months earlier, underscoring the pace at which model capabilities are advancing.

The bottom line for companies is that those that are quickest to adapt to get the most out of AI, literally, if they are just months or, eventually, weeks ahead of their competition, will be the winners in the AI tech revolution. The slow ones will be left behind with their owners and employees.

How to Be on the Winning Side of the AI Wave

We’ll have lots more to say about all of the above in the rest of this series. In the meantime, view AI through Business Creativity, where human insight meets intelligent systems. Rather than AI-enabling your business, how will you rebuild it around AI?

• Evaluate your end-to-end value chain: Where do value pools exist, and what are the constraints preventing you from expanding your share of them?
• Rethink your approach—ignoring constraints: What is the ideal process to access a given value pool, without concern for technical or human constraints?
• Now bring in the AI: How can AI be used to get as close to that ideal process as possible?
• Uplevel it with human collaboration: Forget what the human process masters are doing today; design your new process to embed human/AI collaboration at its core.
• Make it future-tolerant: How will you be able to get even closer to ideal with the technology you can expect to have access to in 6-12 months? How can you design the process to adapt?

References

i. A new year’s resolution for leaders: Redesign work for people and AI, Kweilin Ellingrud, McKinsey Global Institute, January 8, 2026: https://www.mckinsey.com/mgi/media-center/a-new-years-resolution-for-leaders-redesign-work-for-people-and-ai

ii. Gartner Says Worldwide AI Spending Will Total $2.5 Trillion in 2026, Gartner, January 15, 2026: https://www.gartner.com/en/newsroom/press-releases/2026-1-15-gartner-says-worldwide-ai-spending-will-total-2-point-5-trillion-dollars-in-2026

iii. The state of AI in 2025: Agents, innovation, and transformation, McKinsey, November 5, 2025: https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai

iv. From Ambition to Activation: Organizations Stand at the Untapped Edge of AI’s Potential, Reveals Deloitte Survey, DAVOS, Deloitte, January 21, 2026: https://www.deloitte.com/us/en/about/press-room/state-of-ai-report-2026.html

v. Why 50% of GenAI Projects Fail — And How to Beat the Odds, Arun Chandrasekaran, Gartner, January 26, 2026: https://www.gartner.com/en/articles/genai-project-failure

vi. Looking Under the Hood: ORION Technology Adoption at UPS, BSR: https://www.bsr.org/en/case-studies/center-for-technology-and-sustainability-orion-technology-ups

vii. UPS ORION AI Routes Save 100M Miles, $400M Yearly, Reruption: https://reruption.com/en/knowledge/industry-cases/ups-orion-ai-routes-save-100m-miles-400m-yearly

viii. Q3 earnings call: CEO’s remarks, Sundar Pichai, Google, October 29, 2024: https://blog.google/company-news/inside-google/message-ceo/alphabet-earnings-q3-2024/#full-stack-approach

ix. Microsoft CEO says up to 30% of the company’s code was written by AI, Maxwell Zeff, TechCrunch, April 29, 2025: https://techcrunch.com/2025/04/29/microsoft-ceo-says-up-to-30-of-the-companys-code-was-written-by-ai/

x. Top engineers at Anthropic, OpenAI say AI now writes 100% of their code—with big implications for the future of software development jobs, Beatrice Nolan, Fortune, January 29, 2026: https://fortune.com/2026/01/29/100-percent-of-code-at-anthropic-and-openai-is-now-ai-written-boris-cherny-roon/

xi. AI masters new benchmarks faster than ever, HAI Stanford University, 2025: https://hai.stanford.edu/ai-index/2025-ai-index-report/technical-performance

xii. Claude Mythos Might Break Cybersecurity. But Not in the Way You Think, Ricardo Garcês, Medium, April 12, 2026: https://medium.com/@ricardomsgarces/claude-mythos-might-break-cybersecurity-but-not-in-the-way-you-think-d5c64ecbbd3b