Planning Is Doing
The gap between strategy and execution has collapsed. AI agents eliminated the distance between thinking and shipping. Here's what that means for everyone.
There used to be a gap. A big, annoying, soul-crushing gap.
You'd get the idea on Monday. Cool. By Tuesday, you'd sketch a plan. Wednesday through Friday? Shopping for tools, briefing a contractor, waiting on a proposal, refreshing your inbox. Next week, maybe something starts. Three weeks later, a first draft shows up. Six weeks in, you launch - if you're lucky, and if scope hasn't crept into oblivion.
Sound familiar? Yeah.
That gap is gone. The distance between "I want this to exist" and "it exists" has collapsed to the length of a conversation. If you haven't felt that collapse yet, you're already behind the people who have.
This is a status report.
The Shift Nobody Named
We've been calling it "AI productivity." That undersells it by an order of magnitude.
Here's what actually happened: planning became doing. The act of describing what you want - clearly, specifically, with intent - is now the act of building it. Strategy and execution used to be two different phases. Now they're the same phase.
Let that sink in for a second. Every framework you've ever learned about business -- the separation of thinkers and doers, the handoff from strategy to operations, the entire consulting-industrial complex built on the assumption that ideas need armies to execute -- all of it rests on a foundation that just cracked.
Researchers at MIT Sloan and Harvard Business School studied over 700 consultants at BCG using generative AI on real consulting tasks. Performance improved by 40% within AI's capabilities. Less-skilled workers? They saw gains of 43%. The tool compressed the entire cycle of thinking, drafting, iterating, and delivering into a single motion (Dell'Acqua et al., 2023).
That's a phase change.
The Evidence Is Not Subtle
Let's get specific, because vague AI claims are cheap. The data? Anything but vague.
Erik Brynjolfsson and his team at Stanford studied 5,179 customer support agents who got access to a generative AI assistant. Productivity jumped 14% overall. For novice workers -- people in their first months on the job -- the gain was 34%. The AI basically downloaded the tacit knowledge of top performers into everyone else's brain, instantly (Brynjolfsson et al., 2023).
GitHub ran a controlled experiment on developers building software with and without Copilot, their AI coding assistant. The group with AI access completed the task 55.8% faster (Peng et al., 2023). Fifty-five percent. Let that number breathe.
Goldman Sachs projects generative AI could raise global GDP by 7% -- roughly $7 trillion in added economic value (Hatzius et al., 2023). The IMF frames it differently: nearly 40% of global employment is exposed to AI, with the figure rising to 60% in advanced economies (Georgieva, 2024). "Exposed" means the nature of the work is changing, right now, whether the workers change with it or not.
And adoption? Anything but gradual. A 2024 NBER study found that nearly 40% of U.S. adults aged 18-64 are already using generative AI, with 23% of employed workers using it at least weekly. The researchers' conclusion: generative AI adoption has been as fast as the personal computer and overall adoption has been faster than either PCs or the internet (Bick et al., 2024).
Read that again. Faster than the internet.
What "Planning Is Doing" Looks Like in Practice
Here's the before-and-after. Real stuff.
Content marketing, before: Hire a content strategist. Develop a content calendar. Brief writers. Wait for drafts. Edit. Brief a designer. Wait for graphics. Schedule posts. Manage a publishing tool. Elapsed time: 2-4 weeks for a single campaign. Cost: $3,000-$10,000.
Content marketing, after: Describe the campaign to your AI agent. It writes the posts, adapts them for each channel, generates the visuals, schedules the publishing. You review, adjust, approve. Elapsed time: an afternoon. Cost: your existing subscription.
Market research, before: Hire an analyst or a boutique research firm. Define the scope. Wait 2-6 weeks. Receive a 40-page PDF. Try to extract the three insights that actually matter. Budget: $5,000-$50,000.
Market research, after: Ask your AI to analyze the competitive landscape, pull recent data, synthesize findings, and present the top insights with sources. Review and refine. Elapsed time: an hour. The 40-page PDF is still there if you want it, generated in minutes, but the insights came first.
Building an internal tool, before: Write a product requirements document. Get engineering time allocated (wait for the next sprint, or the one after that). Build. Test. Deploy. Elapsed time: 4-12 weeks.
Building an internal tool, after: Describe what you need. The AI generates the code. You test it, iterate through conversation, deploy. Elapsed time: a day, maybe less. GitLab's 2025 Global DevSecOps Survey found that 34% of developers' code is now AI-generated, and 83% of teams using AI in their development lifecycle achieve multiple daily deployments (GitLab, 2025).
The people who plan are now the people who ship -- because the tools no longer require an army between the idea and the output.
The Solo Operator Is the New Enterprise
Here's the implication nobody in corporate strategy wants to say out loud: a single person with AI tools can now produce output that previously required a team of ten.
That's arithmetic. AI handles the first draft of your writing, your code, your research, your design, your data analysis, your outreach. You handle the taste, the judgment, the strategy, the relationships. The only bottleneck left? Your clarity of thought.
The McKinsey Global Institute estimated that generative AI could add $2.6 to $4.4 trillion annually in value across the use cases they analyzed. Most of that value comes from augmenting knowledge work -- the exact work that founders, operators, and small teams do every day (McKinsey Global Institute, 2023).
The Fortune 500 will adopt AI eventually, but slowly -- encumbered by procurement processes, legal reviews, change management committees, and the organizational antibodies that large companies produce against anything threatening the status quo. Accenture's 2024 Technology Vision found that 96% of executives see AI agent ecosystems as a significant opportunity over the next three years. Seeing it and capturing it in a company with 50,000 employees? Good luck (Accenture, 2024).
The advantage belongs to the fast. To the small. To the solo operator who reads about a new capability at 9 AM and has it deployed by lunch.
The Objections, Addressed
"AI output isn't good enough."
It wasn't. Now it often is -- and when it's off, it's still a better starting point than a blank page. The Harvard/BCG study found that even when AI was used outside its capabilities, trained users learned to recognize the boundaries and route around them (Dell'Acqua et al., 2023). The real skill is knowing when to use AI and when to override it. That's called judgment. You already have it.
"This is just hype. Remember crypto? Remember the metaverse?"
Crypto didn't make anyone's Tuesday meeting more productive. The metaverse didn't write anyone's quarterly report. Generative AI is being adopted faster than the personal computer (Bick et al., 2024). The right comparison here is electricity. And this time, the productivity gains are measurable within months, not decades.
"What about quality? What about accuracy?"
Valid concern. AI hallucinates. It makes mistakes. So do interns, contractors, and that agency you're paying $8,000 a month. The real question is: is AI plus your judgment better than the alternative, at a fraction of the cost and time? For a growing number of tasks, the answer is unambiguously yes.
"Won't everyone have these tools? Where's the advantage?"
Everyone has access to a gym, too. The advantage is adoption speed, integration depth, and the willingness to restructure how you work. The Stanford HAI AI Index has tracked this consistently: the gap between AI awareness and AI fluency is enormous (Stanford HAI, 2024). Most people know AI exists. Far fewer have rebuilt their workflows around it. That gap is your window.
"This threatens jobs."
It does. Let's not sugarcoat it. The IMF projects that in advanced economies, roughly half of AI-exposed jobs could see tasks automated, while the other half could be augmented (Georgieva, 2024). The honest answer? This is a restructuring. Work changes shape. The people who adapt will capture disproportionate value. The people who don't will struggle. That has been true of every major technological transition in history, and pretending this one is different doesn't help anyone.
The Real Risk Is Inaction
Every week you spend doing things the old way is a week your competitor spends compounding their advantage. And every hour you spend copying outputs from a chat window into six different tabs is an hour AI spent thinking but not doing. That's the uncomfortable math.
Nobody is running a flawless AI-augmented operation right now. The tools are new, the workflows are rough, and most people's AI is still disconnected from the tools they actually use to ship. The trajectory, though, is unmistakable. Sequoia Capital's David Cahn noted that while AI investment has raced ahead of revenue -- the gap between infrastructure spending and actual returns expanded from $200 billion to $600 billion between 2023 and 2024 -- the founders building real products with genuine user value are the ones who will survive the shakeout (Cahn, 2024). The infrastructure is being built. The question is whether you'll be building on it or watching from the sidelines.
The people who move first get compounding returns. Every workflow you automate this month frees up hours next month. Every AI skill you develop now becomes the foundation for the next capability. Every connection between your AI and your actual marketing stack -- your scheduler, your analytics, your CRM -- is a seam that disappears forever. Every process you reimagine today becomes institutional knowledge that new hires inherit tomorrow.
Sam Altman wrote that "the most successful people I know are primarily internally driven; they do what they do to impress themselves and because they feel compelled to make something happen in the world." That internal drive is what separates the people reading this and nodding from the people who will actually change how they work tomorrow morning (Altman, 2024).
What Comes Next
This post is a starting line.
In the posts that follow, we're getting specific. Painfully specific. How to set up AI agents that actually run your marketing -- not just draft it, but publish it, schedule it, measure it. How to build workflows that compound. How to close the gap between what your AI can think and what it can execute. How to think about this shift as a permanent operating advantage.
The foundation is this: planning is doing. The gap between thinking and shipping has closed. But only for the people who've wired their AI into the places where work actually happens. The ones still toggling between a chat window and a dozen open tabs? They're using AI. The ones whose agents touch every tool in their stack? They've deployed it. That's the difference. And those people are never going back.
And the market will not wait for everyone else to catch up.
References
Accenture. (2024). Technology vision 2024: Human by design. Accenture. https://www.accenture.com/us-en/insights/technology/technology-trends-2024
Altman, S. (2024). What I wish someone had told me. Sam Altman's Blog. https://blog.samaltman.com/what-i-wish-someone-had-told-me
Bick, A., Blandin, A., & Deming, D. J. (2024). The rapid adoption of generative AI (NBER Working Paper No. 32966). National Bureau of Economic Research. https://www.nber.org/papers/w32966
Brynjolfsson, E., Li, D., & Raymond, L. R. (2023). Generative AI at work (NBER Working Paper No. 31161). National Bureau of Economic Research. https://www.nber.org/papers/w31161
Cahn, D. (2024, June 20). AI's $600B question. Sequoia Capital. https://www.sequoiacap.com/article/ais-600b-question/
Dell'Acqua, F., McFowland, E., Mollick, E., Lifshitz-Assaf, H., Kellogg, K., Rajendran, S., Krayer, L., Candelon, F., & Lakhani, K. R. (2023). Navigating the jagged technological frontier: Field experimental evidence of the effects of AI on knowledge worker productivity and quality (Harvard Business School Working Paper No. 24-013). Harvard Business School. https://mitsloan.mit.edu/ideas-made-to-matter/how-generative-ai-can-boost-highly-skilled-workers-productivity
Georgieva, K. (2024, January 14). AI will transform the global economy. Let's make sure it benefits humanity. International Monetary Fund Blog. https://www.imf.org/en/Blogs/Articles/2024/01/14/ai-will-transform-the-global-economy-lets-make-sure-it-benefits-humanity
GitLab. (2025). The 9th annual global DevSecOps report. GitLab. https://about.gitlab.com/developer-survey/
Hatzius, J., Briggs, J., Kodnani, D., & Pierdomenico, G. (2023, April 5). The potentially large effects of artificial intelligence on economic growth. Goldman Sachs Global Investment Research. https://www.goldmansachs.com/insights/articles/generative-ai-could-raise-global-gdp-by-7-percent
McKinsey Global Institute. (2023). The economic potential of generative AI: The next productivity frontier. McKinsey & Company. https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/the-economic-potential-of-generative-ai-the-next-productivity-frontier
Peng, S., Kalliamvakou, E., Cihon, P., & Demirer, M. (2023). The impact of AI on developer productivity: Evidence from GitHub Copilot. arXiv preprint arXiv:2302.06590. https://arxiv.org/abs/2302.06590
Stanford University Human-Centered Artificial Intelligence. (2024). The AI Index 2024 annual report. Stanford HAI. https://hai.stanford.edu/ai-index