Despite the Transformative potential of tools like ChatGPT.most knowledge workers I've talked to don't use it at all. Those that do are mostly limited to basic tasks like summarizing. Only a little more than 5% of ChatGPT's user base pays plus (a small fraction of potential professional users), suggesting a shortage of power users leveraging ai for complex, high-value work.
After more than a decade of building ai products at companies from Google Brain to Shopify Ads, I've witnessed the field's evolution firsthand. With the rise of ChatGPT, ai has gone from nice-to-have enhancements like photo organizers to important productivity drivers for all knowledge workers.
Most executives understand that the current rumors are more than hype: they are desperate to move their companies forward in ai, knowing that it is more powerful and easier to use than ever. So why, despite the potential and excitement, is widespread adoption lagging? The real obstacle is the way organizations approach work. Systemic problems prevent these tools from becoming part of our daily routine.
Ultimately, the question executives need to ask is not “How can we use ai to get things done faster?” Or can this feature be created with ai? “ but rather “How can we use ai to create more value?” What are the questions we should ask ourselves but don't?
Recently, I leveraged large language models (LLMs), the technology behind tools like ChatGPT, to tackle a complex data structuring and analysis task that would traditionally have taken a cross-functional team of data analysts and content designers a month or so. further.
This is what I achieved in one day wearing Google ai Study:
- Transformed thousands of rows of unstructured data into a structured and labeled data set.
- Used ai to identify key user groups within this newly structured data.
- From these patterns, a new taxonomy was developed that can drive a better and more personalized end-user experience.
In particular, I did No Just press a button and let the ai do all the work.
It required intense concentration, detailed instructions, and multiple iterations. I spent hours crafting precise directions, providing feedback (like an intern, but with more direct language), and redirecting the ai when it veered off course.
In a sense, I was compressing a month's work into one day and it was mentally exhausting.
The result, however, was not just a faster process: it was a fundamentally better and different result. LLMs uncovered nuanced patterns and edge cases hidden within unstructured data, creating insights that traditional analysis of pre-existing structured data would have missed entirely.
Here's the trick, and the key to understanding our ai productivity paradox: my success in ai depended on having leadership support to dedicate a full day to rethink our data processes with ai as my thought partner.
This allowed for deep, strategic thinking – exploring connections and possibilities that would have otherwise taken weeks.
This type of quality-focused work is often sacrificed in the rush to meet deadlines, but it is precisely what drives breakthrough innovation. Paradoxically, most people don't have time to figure out how to save time.
Spending time exploring is a luxury most PMs can't afford. Under constant pressure to deliver immediate results, most rarely have even an hour for this type of strategic work; The only way many find time for this kind of exploratory work is by pretending to be sick. They are so overwhelmed with executive mandates and urgent customer requests that they don't feel ownership of their strategic direction. Additionally, recent layoffs and other industry cutbacks have intensified workloads, leaving many PMs working 12-hour days just to keep up with basic tasks.
This constant pressure also hinders the adoption of ai to improve execution. Developing robust test plans or proactively identifying potential issues with ai is considered a luxury, not a necessity. It sets up a counterproductive dynamic: why use ai to identify issues in your documentation if implementing fixes will only delay the release? Why conduct additional research on users and the problem space if the direction has already been set from the top?
Giving people time to “discover ai” is not enough; most need some training to understand how to make ChatGPT do more than summarize. However, the training required is often much less than people expect.
The market is saturated with ai training given by experts. While some classes sell snake oil, many instructors are reputable experts. Still, these classes are often not suitable for most people as a starting point. They are time-consuming, overly technical, and rarely tailored to specific lines of work.
I've gotten the best results by sitting down with people for 10-15 minutes, auditing their current workflows, and identifying areas where they could use LLM to get more done, faster. You don't need to understand the math behind token prediction to write a good message.
Don't fall for the myth that ai adoption is only for those with technical training under the age of forty. In my experience, attention to detail and a passion for doing the best job possible are much better indicators of success. Try to put aside your prejudices; You might be surprised to learn who will become your next ai champion.
My own father, a lawyer in his sixties, only needed five minutes before he understood what LLMs could do. The key was to adapt the examples to your domain. We came up with a somewhat complex legal gray area and I asked ai/” rel=”noopener ugc nofollow” target=”_blank”>claudius to explain this to a first year law student with extreme case examples. He saw the response and immediately understood how he could use the technology for a dozen different projects. Twenty minutes later, I was halfway through writing a new law review article I'd been meaning to write for months.
Chances are, your company already has some ai enthusiasts – hidden gems who have taken the initiative to explore LLMs in their work. These “LLM whisperers” could be anyone: an engineer, a marketer, a data scientist, a product manager, or a customer service manager. Put out a call for these innovators and take advantage of their expertise.
Once you've identified these internal experts, invite them to conduct one- or two-hour “ai audits,” reviewing your team's current workflows and identifying areas for improvement. They can also help create initial directions for specific use cases, share your ai workflows, and provide advice on how to troubleshoot and assess the future.
In addition to saving money on outside consultants, these experts are more likely to understand your company's systems and goals, making them more likely to spot practical and relevant opportunities. People who are hesitant to adopt are also more likely to experiment when they see colleagues using the technology compared to “ai experts.”
In addition to ensuring people have room to learn, make sure they have time to explore and experiment with these tools in their domain once they understand their capabilities. Companies can't just tell employees to “innovate with ai” and at the same time demand features for another month before Friday at 5 pm Make sure your teams have a few hours a month to explore.
Once you've cleared this first hurdle of ai adoption, your team should be able to identify the most promising areas for investment. At this point, you will be in a radically better position to assess the need for additional, more specialized training.
The ai productivity paradox has nothing to do with the complexity of the technology, but rather with the way organizations approach work and innovation. Harnessing the power of ai is easier than the “ai influencers” selling the latest certification would have you believe; often requiring only a few minutes of specific training. However, it requires a fundamental change in leadership mindset. Instead of accumulating short-term results, executives must create space for exploration and deep, open-ended, goal-oriented work. The real challenge is not teaching ai to your workforce; You are giving them the time and freedom to reinvent the way they work.
Want to dive deeper into effective ai implementation? Verify ai-product-managers-ae0f3f933910″ rel=”noopener”>We need to raise the bar for ai product managers and ai-agent-rethinking-the-pursuit-of-autonomy-547ab54f4995″ rel=”noopener”>What characterizes a true ai agent? Rethink the search for autonomy.
