When Personalization Matters: A Decision Matrix For Generative AI
A Simple Framework To Help You Prioritize Generative AI Use Cases
The pace of the latest AI hype has been accelerating since the beginning of this year. Not only is it hard to stay on top of the latest news, but it has also become increasingly difficult to determine what can you actually do with it? Because, as you follow the AI news, you get the impression that generative AI is the new silver bullet. It’s actually not. It’s a deja vu for anyone who’s been through the most recent AI hype just a few years ago. Generative AI has distinct strengths (and weaknesses). You can use it to write your emails, your poems, your speeches. And so, it is very easy to use it more and more and to succumb to the convenience factor. But depending upon the context in which you write, the sender might have different expectations about what they receive from you.
The level of personalization and authenticity, for example. Efficiency vs. authenticity is a conundrum only few have had to ponder up until this point — those with a professional speech- or ghostwriter. But as anyone now has access to a personal ghostwriter, it’s an important one to consider. In that context, we’ve asked the question: When should you actually use generative AI? Let’s explore…
» Watch the latest episodes on YouTube or listen wherever you get your podcasts. «
Increasing Personal Productivity With Generative AI
Generative AI is the latest technology that promises a significant increase in personal productivity, especially when working with text. From summarizing information in documents to drafting sales emails, blog posts, and more — even triggering more complex operations autonomously — generative AI applications like ChatGPT already help automate many tasks that have previously taken us a considerable amount of time. But while most tools are either free or available for just a few dollars per month, relying on generative AI might cost you more in the end — namely: trust and authenticity.
When should you use it?
That cold sales email? — Sure!
Your company blog post? — Alright!
How about a mentee asking for advice? — Hmm.
What about an apology to a co-worker? — Ehm.
Your daughter’s graduation speech? — I don’t know.
Her wedding speech? Or a eulogy for a family member? — Probably not.
Generative AI and the large language models at the heart of it are good at predicting the next word in a sentence, but these models need additional context and specificity to mimic your tone and style. It seems very efficient to have AI generate the text for you. Challenge yourself to view it from the perspective of the person reading your AI-generated message and noticing that it doesn’t 100% sound like you — and how it ultimately makes them feel when they realize it was an AI that wrote it; not you.
Have you ever sent a long and personal note along with an invitation on LinkedIn — and received the pre-canned answer ”Likewise”? Did the other person not even invest 30 seconds to type a personal reply? That’s what it feels like. (You’ll see more of that as Microsoft Copilot rolls out and AI-generated responses become more frequent and more accessible.)
But how can you quickly determine in which scenarios generative AI can already safely increase your personal productivity?
Deciding when to use generative AI
What's a structured way for leaders to assess generative AI, leverage its strengths, and avoid shortcomings? We discuss a methodology below to evaluate the possibility at a use case level based on four key dimensions: Complexity, personalization, impact, and technology readiness.
The Generative AI Decision Matrix
Complexity involves two main factors: adoption and governance, and is critical in B2B environments. It's noteworthy to remember that change management is just as important as finding the right solution. Even great solutions can lead to disappointment if they are not adopted effectively. Additionally, governance is vital in ensuring data privacy, availability, and the ability to validate the data source.
Personalization refers to the ability to customize a narrative to suit your preferences. Depending on the context, personalization may be more or less important. For example, it might not be critical in a supply chain context, but it could be a make-or-break factor in marketing. It's essential to grasp the level of personalization required to implement an effective solution.
The size of the bubble represents the Impact and ROI. You may ask why it's not on the X or Y axis. Although ROI is essential early on in Generative AI, it's more important to consider the leading enablers when making selection decisions.
Technology Readiness is the capacity to implement a solution successfully. This could be impeded either by limitations in the current technology or by the need to train the system to achieve the desired result. Despite the rapid advancement of technology in recent months, it's important to note that the potential roadblocks might not be the tech itself but the need to properly maintain and support it.
A few examples can help bring this to bear
Suppose you're considering writing a book about Product Management using your own experiences. You might adopt a narrative style that includes anecdotes, analogies, and anonymized personal stories. The target would be a medium level of personalization, implying a low complexity. The potential impact of your book could be substantial, with the potential to become a bestseller and make you a millionaire (woo hoo!). Additionally, the tech readiness for such a project is high, making it an ideal time to pursue it.
Personalization: M | Complexity: L | Impact: H | Technology Readiness: H
If you're interested in implementing generative AI to summarize medical notes, there are a few factors to consider. Firstly, it's vital to ensure that your doctors trust the summary generated by the AI. Additionally, the data used to train the system will contain personally identifiable information (PII), which presents a challenge. You'll also need to ensure that the summaries are consistent across all doctors and that you have enough data to train the AI effectively. However, the impact of implementing generative AI is significant, while the level of personalization is relatively low (it is not a love letter!).
Personalization: L | Complexity: H | Impact: H | Technology Readiness: M
Summary
Generative AI already delivers significant personal productivity increases today. As the tools and technologies further improve in quality and performance, they will become even more and more attractive to use. However, there are exceptions to the general rule of using generative AI. When your message is personal, it should either come from your head or your heart — and the recipient of your message should know that, Generative AI, and the foundation models that power this technology, are evolving rapidly. Today’s assessment might be rejected within days or weeks. And as we are getting use to this new technology and are adapting our expectations, we’re curious…
How do you decide what to use generative AI for?
This post has been co-authored by Harish Natarahjan, Strategy Consulting Executive for High-Tech industries.
» Watch the latest episodes on YouTube or listen wherever you get your podcasts. «
What’s next?
Appearances
June 8 - Panel discussion with Transatlantic AI eXchange on Web 3.0 Generative and Synthetic Data Application
Join us for the upcoming episodes of “What’s the BUZZ?”:
May 23 - Rod Schatz, Data & Digital Transformation Executive, will join as we explore how you can Prepare Your Business Against AI-generated Misinformation.
June 8 - Ravit Dotan, Director The Collaborative AI Responsibility Lab at University of Pittsburgh, will join when we cover how responsible AI practices evolve in times of generative AI.
Follow me on LinkedIn for daily posts about how you can set up & scale your AI program in the enterprise. Activate notifications (🔔) and never miss an update.
Together, let’s turn hype into outcome. 👍🏻
—Andreas
 | A guest post by
|
nice framework for Generative AI Use Cases