Leveraging AI in the Workplace: Building a Culture of Innovation
Why purchasing ChatGPT Enterprise is not enough. To thrive, businesses need to build a culture around AI.
Introduction
The release of GPT Enterprise
A few weeks ago, on August 28th, 2023, OpenAI announced ChatGPT Enterprise, their new service targeted at large businesses. According to the post on their blog, it will feature enterprise-grade security, unlimited GPT-4 access, longer context windows, and advanced data analysis capabilities. That’s a lot.
Since then, I’ve refrained from writing anything regarding it, partly because I have been fervently studying for my final accounting exam, but also because I wanted to really think about what this means for businesses and what I can contribute to the conversation.
As I’ve pondered what kind of impact this level of service may have, I’ve come to the conclusion that simply purchasing access to the tool will not be enough for businesses to gain a competitive edge. It will be the minimum requirement to keep pace with the competition. Do you gain a competitive advantage from using Excel?
All businesses will purchase access to the tool, or soon-to-be-released competitors from other major tech companies, and any competitive advantage gained from a productivity standpoint will quickly dissolve.
However, over the long term, I still think an incredible opportunity exists for those companies that are able to see what these AI tools are truly capable of. That is, are you able to leverage the flexibility of these tools and the creativity of your workforce to innovate and find completely new use cases that no one else has yet thought of?
What it means to have AI in the workplace
The introduction of AI like ChatGPT into the workplace is fundamentally different than traditional software applications. Rather than performing a specific function, these AI systems are omni-purpose tools, general in nature, with capabilities we are only beginning to understand and uncover. To gain real competitive advantages, companies who want to thrive need to be leaders in figuring out how to use them. This will be a team effort, across the workforce, to uncover innovative ways to apply these tools in day-to-day operations.
Companies that successfully empower their employees to fully utilize these tools for results that are better, faster, cheaper, or safer will pull ahead of their competition. Aside from technical implementation, it will thus be critical to build a culture that encourages experimentation with AI and celebrates successes and failures alike. Essentially, the goal should be to create a critical mass of experimental AI culture and engagement that continuously surfaces new opportunities and solutions.
I’ve put together a 4-part framework called "M-A-S-S" to help companies build this culture of AI-powered innovation.
1. Meaningful Empowerment
People's brain power is the biggest limiting factor to productivity. According to Cal Newport’s book Deep Work, we only have the capacity for 2-4 hours of truly intensive, focused work per day. With Meaningful Empowerment, AI can act as a lever to maximize this capacity by taking on repetitive or low-value tasks and allowing employees to focus on true value-driving activities.
To do this, employees need training on AI basics and best practices - I’ve written a little bit about it here - but, the main things to focus on are effective prompting, understanding the basic capabilities of Python, as well as demonstrations of a few use cases. The goal is not mastery of the tool, or showing them what to use the tool for, but to give them the curiosity and confidence to experiment with it themselves. Rather than dictating uses, encourage people to find solutions tailored to their specific role. Once a basic level of comfort with the tool is achieved, it’s easy to start coming up with how it could be used in the workplace.
Some ideas I have for my own work context:
Note: Suggestions below are limited to legal and regulatory requirements, along with data security/confidentiality best practices.
Uploading long documents to "converse" with, surfacing insights vs. reading the whole thing.
Brainstorming approaches to complex projects, or brainstorming places your plan may have overlooked.
Automate writing memos, proposals, documentation, year-end reviews.
Data cleaning and formatting, performing advanced analytics, or even highlighting possible conclusions or statistical significance of results.
Summarize long documents or meeting minutes.
Perform routine testing activities or basic analytics.
Translate foreign text.
Edit, proofread and provide feedback on writing.
Empowering people to direct AI with their domain-specific expertise and current business context will uncover more innovative applications than top-down mandates or suggestions.
2. Alternative Delegation
This is similar to Meaningful Empowerment but targeted towards managers. A key skill for successful managers is learning to effectively delegate tasks that you could do yourself but would be better off passing along. With Alternative Delegation, we want to help build the secondary skill of habitually considering whether AI could perform tasks they would previously delegate to others. This is not necessarily to save time completing the task directly, but rather to avoid delays due to miscommunication.
How often have we delegated work, only to receive a completely different deliverable than intended? Since AI performs the task almost instantly, we gain immediate feedback to course-correct our instructions before long bouts of lead time are wasted.
At worst, in cases where the AI tool is unable to complete the task sufficiently, you've just spent the time fully and clearly articulating instructions that can then be passed on through your standard working practices.
3. Sharing and Collaboration
This is the most critical step in building a culture that spontaneously produces new AI use cases. Ensure that people are able to share their AI innovations across the company. Everyone needs to be encouraged to share their ideas, successes, and failures.
Building a cultural critical mass requires each team member to regularly share their experiences with the first 2 steps above. As these use cases are shared throughout the organization, they'll inspire others, who'll provide feedback and share their own ideas, repeating the cycle. The more this occurs, the more it will continue and the more unique and differentiated the ideas will get.
The goal here is raw creativity and collaboration. Maximize cross-pollination within, and importantly between, business units to surface creative applications from different contexts. According to David Epstein's book, Range, many of the most innovative discoveries come from people who have a broad base of knowledge to work with, rather than narrow expertise. Here it is no different. Lateral dissemination is what will spark the most innovative solutions and lead to a true competitive advantage.
Some considerations for how this can be done:
Lead by example. Be an early adopter yourself, use chatbots and other tools daily and share with your team how you are using them. Ask for input on new ways to use them. Your usage and curiosity will inspire others.
Ensure universal AI access and training. Innovation can come from anyone, so allow everyone to experiment.
Build communication channels to highlight successes and learnings. Whether it's a weekly or monthly blog or a dedicated Teams/Slack channel, deliberate facilitation of recognizing experimentation is vital.
Incentivize sharing via feedback loops, rewards, or even competitions. Teams and Slack channels allow for reactions which provide positive feedback to engaged employees and highlight posts that are the most popular which can then be additionally recognized and rewarded.
Provide resources. Create centralized resources like FAQs, user guides, policies, and lists of the most popular use cases, so that information is easy to access when needed.
The most important thing is providing a concerted effort to get people talking to each other about how they are using the tools in their work.
4. Safety
The crucial overarching topic. While pursuing innovation, companies must prioritize ethics, security, and thoughtfulness around AI. Concerns exist around data privacy, algorithmic bias, legal liability, and more.
I’ve discussed this before, but prior to deploying AI, assess risks specific to your industry and put appropriate guardrails in place. Be aware of governing regulations and be transparent with employees who may be worried about the technology. AI presents amazing opportunities but also responsibilities we must take seriously.
Conclusion
GPT Enterprise and similar AI services will soon be commonplace. To gain a real advantage, companies must do more than simply adopt the technology.
The key to a true competitive advantage is unleashing human creativity to uncover innovative applications - and that requires building a culture of free experimentation and sharing, with ethical oversight. The M-A-S-S framework provided here aims to help companies achieve the cultural critical mass required to fulfill the promise of AI while avoiding potential pitfalls.
By empowering your greatest asset - your people - you can build an engine for continuous innovation that will drive a real competitive edge. The future will belong to the companies that embrace AI as a collaborative effort, not just a technological solution.