What did the Apollo economist actually find?
The report from the Apollo economist, published on TechCrunch AI on October 25, 2026, drops a hard truth on the business world. The data suggests that despite the massive hype cycle and billions in capital expenditure, artificial intelligence has delivered zero financial payoff for the vast majority of corporate America. This finding excludes the tech sector, which is currently the only industry capturing tangible value from the AI boom. For the rest of the economy, the technology remains a significant cost center rather than a revenue driver.
This is not a vague sentiment but a direct observation of corporate balance sheets. Companies across retail, manufacturing, logistics, and professional services have poured capital into large language models and integration services. They expected immediate productivity gains that would translate into higher operating margins. Instead, they have found that the cost of implementation, maintenance, and the necessary talent to manage these systems often exceeds the value generated by the tools themselves. The Apollo economist is highlighting a disconnect between technological capability and economic utility.
The story is particularly damning because it separates the tech sector from the rest of the market. Tech companies are building the infrastructure, selling the API access, and providing the compute power. They are the ones harvesting the value. Every other business is essentially renting this infrastructure. When a business rents tools without a clear, direct revenue-generating application, they are simply moving money from their own pockets into the pockets of the AI providers. This is a transfer of wealth, not a creation of value.
For a solopreneur or a small agency owner, this news should be a wake-up call. It validates the suspicion that buying into the hype without a specific, measurable use case is a losing strategy. We have seen this pattern before in other technology cycles. Companies rush to adopt new software because they fear being left behind, only to realize that the software itself does not solve fundamental business problems. The Apollo report provides the data to back up what many of us have felt in the trenches.
The critical takeaway here is the distinction between 'tech' and 'non-tech' companies. If you are not in the business of selling AI or building the underlying models, your relationship with AI should be purely utilitarian. You are not an AI company. You are a business that uses tools. If those tools do not directly lower your customer acquisition cost, increase your lifetime value, or drastically reduce operational overhead, they are a liability. The Apollo data suggests that for most, they are currently just a liability.
Why is the tech sector the only exception?
The tech sector is the only industry capturing value from AI because it owns the means of production. Companies like Anthropic, OpenAI, and the major cloud providers control the infrastructure, the models, and the distribution channels. They are not just using AI to make their existing businesses more efficient; they are selling AI as a product. This fundamental difference allows them to monetize the technology directly, turning the AI hype into actual revenue and profit growth.
Consider the business model of an AI model provider. They sell API access, compute time, and enterprise licensing. Their revenue scales linearly with usage. Every time a non-tech company integrates their model into a workflow, the tech company collects a fee. This is the classic 'pick and shovel' strategy. During the gold rush, the people selling the shovels made the money, while the miners often struggled to find enough gold to cover their costs. The tech sector is currently selling the shovels.
Furthermore, tech companies have the engineering talent required to integrate these models deeply into their products. They are not just slapping a chatbot interface onto a website. They are building AI agents into the core of their software architecture. This integration creates a moat. It makes their products stickier, more valuable, and harder to replace. For a non-tech company, the barrier to entry for this level of integration is incredibly high, often requiring a dedicated engineering team that most small businesses cannot afford.
The tech sector also benefits from the network effects of data. Because they control the platforms, they collect the data generated by users. This data is then used to improve the models, which makes the products better, which attracts more users, which generates more data. This flywheel is the engine of their profitability. A non-tech company using a third-party model does not get this advantage. They are just a customer, not a participant in the value creation cycle.
Ultimately, the tech sector is capturing the value because it is the only sector that treats AI as a foundational asset rather than a feature. They have reorganized their entire business models around the capabilities of these models. For everyone else, AI is an add-on. And in the world of software and business, add-ons rarely generate the kind of financial payoff that justifies massive capital expenditure. The Apollo economist is simply pointing out that the value is flowing upstream to the source.
Is the hype cycle masking a lack of productivity?
The hype cycle surrounding artificial intelligence has created a dangerous illusion of productivity. Companies announce partnerships, launch AI-powered features, and boast about their 'AI strategy' to shareholders and the press. This creates a perception of progress and innovation. However, the Apollo economist report suggests that this is largely performative. The actual productivity gains, measured by output per employee or profit per customer, are not materializing in the way the market expected.
This phenomenon is common in the early stages of any transformative technology. There is a period of 'installation' where companies spend heavily on infrastructure and experimentation. During this phase, productivity often dips or stays flat because the old systems are being disrupted and the new systems are not yet optimized. The problem is that the AI hype cycle has been so intense that many companies have skipped the optimization phase and gone straight to full-scale deployment, hoping the technology would magically solve their inefficiencies.
When you look at the metrics, the disconnect becomes clear. Companies are measuring 'AI adoption' by the number of employees using tools or the number of processes touched by AI. These are vanity metrics. They do not correlate with financial payoff. A company could have 100% AI adoption and still be losing money if those tools are not driving revenue or cutting costs. The hype cycle encourages companies to focus on these vanity metrics rather than the bottom line.
The lack of productivity is also due to the 'garbage in, garbage out' principle. AI models are only as good as the data they are trained on and the prompts they receive. Many companies are deploying AI without fixing their underlying data architecture. They are feeding messy, disorganized data into powerful models and expecting clean, actionable insights. This is a recipe for failure. The hype cycle promises that AI can fix these problems, but the reality is that AI often just amplifies them.
We need to stop measuring success by the volume of AI activity and start measuring it by the impact on the P&L. If a company cannot point to a specific line item in their financial report that has improved because of AI, then the hype is just noise. The Apollo economist is stripping away that noise and showing the reality. It is a sobering reminder that innovation without execution is just a cost.
What defines a real financial payoff?
A real financial payoff from AI is defined by a measurable, net improvement in the bottom line. It is not about time saved on drafting emails or the number of tasks automated. It is about whether that automation directly leads to higher revenue, lower customer acquisition costs, or increased operational margins. If you cannot trace the AI implementation to a specific increase in profit, you are not seeing a payoff; you are just incurring an expense.
For a solopreneur, a financial payoff looks like a CRM automation that reduces the time to lead from hours to seconds, resulting in a higher conversion rate. It looks like an AI-driven pricing model that optimizes margins on every sale. It looks like a content pipeline that generates qualified leads at a lower cost than manual outreach. These are tangible, money-making outcomes. They are not 'efficiency' in the abstract; they are revenue-generating operations.
Contrast this with the common 'AI payoff' narrative, which focuses on productivity. Productivity is a trap. If you use AI to write 10 blog posts instead of one, but those blog posts do not generate more sales, you have not increased your productivity; you have increased your output of noise. You have spent more time managing AI tools and editing content without any increase in value. That is not a financial payoff.
To measure a real payoff, you need a baseline. You must know exactly what your costs and revenues were before you implemented the AI solution. You must measure the delta after implementation. If you are not tracking these metrics, you are guessing. And in business, guessing is a strategy for failure. The Apollo economist report is a call to be more rigorous about how we define and measure success.
Ultimately, a financial payoff is about leverage. AI should allow you to do more with less, but only if the 'more' is something that customers are willing to pay for. If you are using AI to do more of the same low-value work, you are just running faster on a treadmill. Real payoff comes from using AI to unlock new capabilities, reach new customers, or deliver value in a way that was previously impossible.
Why does custom AI implementation often fail for small businesses?
Custom AI implementation often fails for small businesses because it introduces complexity that they are not equipped to manage. A custom solution requires data engineering, model fine-tuning, and ongoing maintenance. These are specialized skills that require a dedicated team. When a small business attempts to build a custom AI solution, they often underestimate the resources required, leading to a project that is either abandoned or becomes a massive drain on time and money.
The allure of custom AI is that it promises a 'perfect fit' for the business. It suggests that you can build a tool that understands your specific processes, your specific data, and your specific customers. This is true, but it is also expensive. The cost of building and maintaining a custom model usually outweighs the benefits for a small business. There are almost always off-the-shelf solutions that can achieve 80% of the result for 10% of the cost.
Small businesses often fall into the trap of 'over-engineering'. They want to build an AI agent that does everything, from answering customer emails to managing inventory to writing marketing copy. This is a mistake. It is better to build a simple, focused tool that solves one specific problem well. A custom tool that solves one problem is manageable. A custom tool that tries to solve everything is a nightmare.
Furthermore, custom AI requires constant updates. The field of AI is moving incredibly fast. A model that is state-of-the-art today might be obsolete in six months. If you have built a custom solution, you are responsible for updating it. This is a never-ending task. For a small business, this is a distraction from their core mission. It is better to rely on platforms that handle the updates and maintenance for you.
The reality is that for most small businesses, the best AI strategy is to be a power user of existing platforms, not a developer of new ones. You should be looking for ways to integrate the tools that already exist into your workflow. You should be looking for APIs, not fine-tuned models. You should be looking for automation, not custom development. This is the path to a sustainable, profitable AI strategy.
How should a solopreneur approach AI investment?
A solopreneur should approach AI investment with a 'revenue-first' mindset. This means that every dollar spent on AI tools and every hour spent on AI implementation must be tied to a specific, revenue-generating outcome. If you cannot draw a straight line from the AI tool to a sale or a cost reduction, do not do it. This is the only way to avoid the trap of spending money on 'innovation' that does not pay the bills.
Start by auditing your current business processes. Identify the bottlenecks that are costing you money or limiting your growth. Are you losing leads because your response time is too slow? Are you spending too much time on manual data entry? Are you struggling to create content that converts? These are the problems that AI can solve. Choose one problem, find a tool that solves it, and implement it.
Once you have identified the problem, look for existing solutions. Do not build if you can buy. There are thousands of AI-powered SaaS tools available. Most of them are cheap and easy to integrate. Use these tools to build your 'stack'. Your goal is to create a system that automates the repetitive, low-value work so you can focus on the high-value, strategic work that only you can do.
Measure the results. If the tool works, keep it. If it doesn't, cut it. Be ruthless. Do not get attached to a tool because it is 'cool' or because everyone else is using it. If it is not contributing to your bottom line, it is a distraction. The goal is to build a lean, profitable business, not to have the most 'AI-enabled' tech stack.
Finally, keep your implementation simple. Do not try to build a complex, multi-agent system from day one. Start with simple automations. Connect your CRM to your email platform. Use AI to summarize meeting notes. Use AI to draft social media posts. As you get comfortable, you can add more complexity. But always keep the focus on the outcome, not the technology.
Is it better to build your own tools or buy existing ones?
For almost every solopreneur and small business, buying is better than building. Building a custom AI tool requires significant capital, time, and specialized expertise. It is a high-risk, high-reward strategy that is rarely appropriate for a business whose core competency is not software development. Buying an existing tool allows you to leverage the R&D of the tech giants and focus on your own business.
When you buy, you get the benefit of a product that is already tested, maintained, and supported. You get a team of engineers who are constantly updating the model and fixing bugs. You get a community of users who are sharing best practices and tips. You get a tool that is ready to use, so you can start seeing results immediately. This is the definition of leverage.
Building, on the other hand, is a commitment. You are responsible for the entire lifecycle of the product. You have to deal with infrastructure, security, compliance, and updates. You have to troubleshoot every issue that arises. And if the underlying technology changes, you have to rewrite your code. This is a massive drain on your resources that takes you away from your customers and your revenue.
There is a specific case for building, of course. If you have a unique process that cannot be handled by any existing tool, or if you are building a product that you intend to sell to others, then building makes sense. But for internal business processes, buying is almost always the right choice. It is the path of least resistance and the path to the fastest return on investment.
The distinction is simple: are you in the business of creating software, or are you in the business of using software to create value? If you are the latter, buy. Let the tech companies fight it out for the best models. You just need to pick the one that works best for your specific problem and get back to work. That is how you win.
How do you measure AI return on investment?
Measuring AI return on investment requires a clear before-and-after comparison of your key performance indicators. You cannot rely on vague feelings of 'efficiency'. You need hard numbers. Start by establishing a baseline for the process you are automating. How much does it cost? How long does it take? What is the output? Then, implement the AI solution and measure the same metrics. The difference is your ROI.
If you are using AI to improve customer service, measure the cost per ticket and the resolution time. If you are using AI to generate leads, measure the cost per lead and the conversion rate. If you are using AI to manage your marketing, measure the return on ad spend. These are the metrics that matter. If these numbers are not improving, then the AI is not providing a return.
It is also important to factor in the 'hidden' costs of AI. These include the subscription fees for the tools, the time you spend learning and managing them, and the cost of any integration services. These costs can add up quickly. If you are not tracking them, you might think you are making a profit when you are actually losing money. Always calculate your total cost of ownership.
Be realistic about the time it takes to see a return. AI implementation is not a magic switch. It takes time to set up, time to tune, and time to integrate into your workflow. Do not expect an immediate payoff. Give it a few months. But if you are not seeing progress after a reasonable period, do not be afraid to pull the plug. The best AI strategy is the one that you can pivot away from if it is not working.
Ultimately, the best measure of ROI is your bank account. If your business is more profitable after implementing AI, then it is a success. If it is not, then it is a failure. Do not let the complexity of the technology distract you from this simple truth. The goal of business is to make money, not to use the latest tools. Keep your eye on the prize.
What are the common pitfalls of AI adoption?
The most common pitfall of AI adoption is 'shiny object syndrome'. This is the tendency to chase every new AI tool, model, and feature as soon as it is released, regardless of whether it solves a real business problem. This leads to a fragmented tech stack, wasted money, and a lack of focus. It is the opposite of a strategic approach to technology.
Another major pitfall is ignoring the human element. AI is a tool, not a replacement for human judgment and creativity. Many businesses make the mistake of trying to automate everything, including the tasks that require a human touch. This leads to generic, low-quality output that turns off customers. You must use AI to augment your work, not to replace the parts of your business that make you unique.
A third pitfall is poor data management. AI models are only as good as the data they are trained on. If your internal data is messy, disorganized, or incomplete, your AI tools will produce poor results. You cannot fix a data problem with an AI solution. You must clean up your data first. If you don't, you are just building on a shaky foundation.
Finally, many businesses fail to train their team. Even the best AI tool is useless if your employees do not know how to use it effectively. You must invest in training and education. You must create clear guidelines and best practices. You must foster a culture of experimentation and learning. If you don't, you will not get the full value out of your AI investments.
Avoid these pitfalls by being disciplined, focused, and human-centric. Treat AI as a tool, not a savior. Focus on solving real problems, not on chasing hype. And always, always keep your eye on the bottom line. That is the only way to avoid the 'zero payoff' trap and build a business that actually benefits from the AI revolution.
What should you do this week to fix your AI strategy?
This week, your primary goal is to audit your AI usage and cut the fat. Go through every AI tool you are currently paying for and ask yourself: 'Does this tool directly contribute to a revenue-generating outcome?' If the answer is no, cancel the subscription. This is the fastest way to improve your margins and regain focus. Do not worry about missing out on the 'next big thing'. If it is truly important, you will hear about it later.
Next, identify one specific process in your business that is bottlenecking your growth. It could be lead response time, content creation, or customer onboarding. Choose one. Do not try to fix everything at once. Find one tool that can automate or optimize that process. Implement it, test it, and measure the results. If it works, great. If not, cut it and try something else.
Third, establish a clear metric for success for that process. What does 'better' look like? Is it a 10% increase in conversion? Is it a 20% reduction in time spent? Write it down. This is your baseline. Without it, you are just guessing. Make sure you are tracking this metric every week. This will give you the data you need to make informed decisions.
Fourth, spend some time cleaning your data. If you are using a CRM, make sure your customer data is accurate and up-to-date. If you are using an email platform, make sure your lists are segmented and clean. This is unglamorous work, but it is the foundation of any successful AI strategy. The cleaner your data, the better your AI tools will perform.
Finally, commit to a 'human-first' approach. For every task you automate, ask yourself: 'Does this improve the customer experience?' If the answer is no, do not automate it. Your customers are buying from you, not from your AI. Keep the human connection at the center of everything you do. That is your competitive advantage. That is how you win in an AI-driven world.
Sources
TechCrunch AI — AI delivers ‘zero financial payoff’ to most companies, Apollo economist says — https://techcrunch.com/2026/10/25/ai-delivers-zero-financial-payoff-to-most-companies-according-to-apollo-economist/