Understanding AI's Impact Through What is Created: Labor Units in Focus

Listen to the conversation around AI, and you’ll hear plenty of talk that has a familiar ring from past eras of technology.

We hear about how AI will enable new levels of efficiency, and lower costs. It will automate menial tasks, freeing up more time to solve high-value problems, lead, and take on the next challenge.

In each of these areas, AI is already demonstrating the capability to produce massive gains, and it is improving rapidly. The implications for the workplace are relevant and worthy of consideration, just as they were at the dawn of the Industrial Revolution, the digital age, and the internet era.

But the truth is that we can only tell part of today’s story with the metrics of the past. Previous eras of technology measured the productivity of humans, with companies selling seats and subscriptions to ascribe value to the software applications that served as enabling tools. In the AI era, the machines will be measured, both in the output they produce and their ability to augment humans. It is in this latter category — machines’ ability to do more than humans ever thought possible —where AI will make its most transformative impact. 

Unlike past eras, we must understand the value of AI through what is created, not what is saved. Through the labor unit offset, we will capture how AI replaces and enhances the work of humans, in turn driving an increase in the volume of AI applications, and their effectiveness. This will result in new distribution models that better balance value and business results than those of past eras, while creating more measurable outcomes. In this article, we will break down the labor unit offset, and explore new modes of value creation in the AI era.

THE NEW PARADIGM OF KNOWLEDGE WORK

With the ability to reason, solve problems, engage in conversation, and create, AI presents a 100-year technology transformation that will expand the bounds of knowledge work. According to McKinsey & Company, generative AI could automate and augment 60 to 70% of the activities of knowledge workers. Already, it is driving gains for workers at all levels.

Lower-skill employees will be able to leverage AI to skill up faster. Upon introduction of a conversational chatbot for customer support agents, a recent NBER working paper by MIT and Stanford researchers found a 34% increase in productivity among novice and low-skilled workers. AI helped to disseminate the best practices to others, while also improving customer sentiment. 

For highly skilled workers, the ability to make decisions and take on complex tasks will create new AI teammates. A 2023 study from the MIT Sloan School of Management found that using AI with its current capabilities increased productivity by 40% among highly skilled workers. The potential to unlock new levels of value is profound. Already, OpenAI is developing a “high-income knowledge worker” agent for a reported price of $2,000 monthly, and a “PhD-level” agent for $20,000 monthly. 

This will have a particular impact on the areas where we focus at Grit Capital Partners, including commerce, media, martech, and fintech which today are the most fertile ground for AI disruption.

With a recent memo, Shopify CEO Tobi Lutke revealed how AI is already augmenting the commerce software company’s 10X employees as the “tools become 10X themselves.” This is resulting in “reflexive and brilliant usage of AI to get 100X the work done.”

Shopify’s workforce stands to be reshaped. Before receiving approval for hiring, teams must now prove that the proposed new job can’t be done using AI. Employees should consider, “What would this area look like if autonomous AI agents were already part of the team?”

But the point is not to reduce resources. AI is producing gains, and they go beyond productivity. Lutke reports that Shopify team members are using AI to “approach implausible tasks, ones we wouldn’t even have chosen to tackle before.” 

The introduction of AI multiplies not only the amount of work that gets done, but increases the complexity and effectiveness of that work, which in turn increases customer satisfaction. In the process, new pockets of demand will form as customers seek more AI-enabled capabilities, leading to new use cases and applications. 

Capabilities improve, demand keeps growing. This leads to more use of AI, more work being accomplished, and more new products. In the end, this cycle more than offsets any reduction in headcount. Shopify is attempting to create an environment where it repeats, over and over. 

This arrives at an opportune time. Breakthroughs in the efficiency of LLM training from DeepSeek and ByteDance are showing the way to lower infrastructure costs. Following Jevons Paradox, increased affordability and availability of this resource will drive up the adoption of AI. Increase demand, and more entrepreneurship at the application layer will follow. Shopify won’t be alone.

THE VALUE OF AI

Along with increasing the autonomy of applications, AI will usher in an era where software can be measured on its own merits. If it can work on its own, its output can be quantified. This is different from past eras of technology, where the exchange of value was based on providing access to technology, rather than the value it provided.

As this evolution begins, software leaders are exploring outcome-based pricing models, oriented around the work performed by AI. In one early example, Salesforce is charging per-conversation for its Agentforce product. Each time an autonomous Salesforce “agent” performs work instead of a human, there is a per-unit cost. 

The economics of AI have yet to be fully worked out, especially at the application layer, where most customers interact with the technology. As EY notes, determining the value of an outcome carries additional complexity. Applied AI companies that work closely with customers to test new models will be best positioned to unlock value.

In past eras, we’ve seen how category-defining companies emerged to create sticky business models that are synonymous with technology capabilities. Amazon and eBay created marketplaces in the internet era, while the rise of social media and on-demand were foundational to the mobile economy. In the cloud era, Salesforce itself invented SaaS and the per-seat model, ushering in a new era of subscription-based growth. Today, some are proposing to flip that model to Services as Software, as workflows and processes are designed around AI as the primary “users,” rather than humans.

At Grit Capital Partners, we’ve seen how companies working at the application layer seize opportunity to align technology capability, delivery, pricing, and ROI. In previous eras of technology, we played a pioneering role in developing the App Store business model, and the development of the product-led growth strategy that helped to galvanize SaaS growth. The next generation of entrepreneurs will have to be just as forward-thinking. As an early-stage venture firm, investing in Applied AI, we value future markets not just for their software revenue potential, but also for the labor unit value of what is created in the human-machine combination of future output in our target markets.