AI First Imperative: Preventing Your Digital Kodak Moment
Imagine standing in a photo lab, holding a freshly printed picture, unaware that the very act of developing film is about to vanish. Disruption rarely announces itself with sirens—it creeps in, reshaping industries before most notice. AI isn’t just another innovation; it’s a tidal wave accelerating change. Will you harness it to stay ahead, or risk falling behind? What happens next is up to you.
Published Mar 18, 2025
Why did giants like Kodak, Blockbuster, and Nokia fail? The pattern is so recognizable it has a name: "The Innovator's Dilemma." You can do everything right according to traditional business wisdom and still find yourself disrupted out of existence when technological paradigms shift.
We all know these cautionary tales of corporate extinction, but let me share my personal journey with technological evolution. Back in high school, I would run 50-meter cables across rooftops just to play Doom with my neighbor. I'd make late-night phone calls to download emails from a local BBS. I still vividly remember my awe when I first experienced text chat through a 2400bps modem, thinking: "How amazing would it be if everyone communicated this way? We could have conversations with people on the other side of the planet. Continuously and essentially free."
Like many of you, I've witnessed wave after wave of technological revolution: the web, email, WiFi, IRC evolving into instant messaging, Stack Overflow transforming how we learn to program, mobile phones, smartphones, cloud computing, cryptocurrencies.
More than a decade ago, Rita McGrath highlighted in her HBR article "The Pace of Technology Adoption is Speeding Up" that industries were experiencing accelerated innovation cycles. She pointed to examples like automotive manufacturing, where design cycles had halved from 60 months to just 24-36 months, and consumer technology, where recent innovations were reaching mass adoption significantly faster than their predecessors.
With each of these advances, if you traveled back in time just five or ten years and described them, they would have sounded like science fiction. But Generative AI has changed that timeline drastically. Now you only need to go back a few months for it to sound like sci-fi.
As someone naturally enthusiastic about technology, I spent about two years trying to make sense of machine learning and generative AI before experiencing my own "wow" moment. When I rushed to share this epiphany with friends and colleagues, I struggled to convey the magnitude of what I was seeing.
I understand if you're not excited yet—but I believe with absolute conviction that Generative AI is creating a divide that will leave many behind. And I don't want you to be one of them. This isn't dark magic, though it often feels that way. It's not a silver bullet either. The key is keeping an open mind and learning how to use these tools effectively. The reward isn't just moving faster—it's reducing the risk of falling behind.
That's the core purpose of this AI First Imperative Series. Working from the assumption that some observations about these technological signals and trends are true, I suggest a conversation about preventing obsolescence and focusing on what truly matters.
My journey with AI began like many others - with healthy skepticism. In October 2022, perhaps influenced by media coverage or sensing the imminent launch of ChatGPT, I started exploring the field and discovered AWS DeepRacer—Amazon's 1/18th scale autonomous racing car designed to help developers learn reinforcement learning through a gamified experience. I even considered purchasing the physical car but hesitated when I couldn't connect it to practical, real-world use cases.
Simultaneously, I explored various ML courses but found most heavily weighted toward theory, requiring substantial time investment in mathematical models or significant budget for meaningful experimentation. I spent about a hundred dollars on DeepRacer credits alone—not insignificant for what was essentially just dipping my toes in the water. Though I successfully trained a virtual car using reinforcement learning, it remained an interesting but isolated experience.
As a non-native English speaker, I've always struggled with the subtleties of the language—distinctions between formal and informal tone, navigating UK versus US English conventions. When I discovered ChatGPT, my first use was simply checking if my English was correct. Useful, but hardly revolutionary.
The next level came when I asked it to rewrite a few paragraphs into "Pirate English" for my football team, aptly named "Los Piratas." This experiment was entertaining but still relatively superficial.
Eventually, I grew tired of writing match chronicles. After discovering Sonauto and Suno, I had the idea to create songs instead of written chronicles for the football matches—and that was my first genuine "wow" moment. Throughout the 2024 winter season, I generated about 18 different songs in multiple languages (English, Spanish, and Portuguese) across various genres (Tango, Blues, Rock and Roll, Pop, Hip Hop, Bossa Nova, Jazz, etc). Each song became part of the video highlights for a match, featuring lyrics and musical style tailored to that specific game. By season's end, I had released my first album—a collection of 18 original songs, each describing a different match.
While this might not sound groundbreaking, consider that I would never have been able to compose music or write rhyming lyrics in three different languages on my own. Even if we're talking about modest musical and lyrical quality, this represented something I simply couldn't have accomplished without AI assistance.
Do you suffer from impostor syndrome? I squeezed into John Williams and Hans Zimmer's table without paying the ticket price. Yeah, my job is fake, it isn't "entirely mine," but I was able to share some thoughts and emotions through sounds just like they do. I "compose" in at least seven genres, whereas John Williams seems to stick with one or two 😄.
This experience shifted my perspective: what initially seemed like a technological novelty revealed itself as a fundamental transformation in how we create, communicate, and solve problems.
By November 2024, I attended AI Supercharge in Wellington where AI adopters showcased how they're implementing GenAI. Three revelations left me astounded. First was the consensus among many attendees that "AI reduces software development effort by 20 times." The second was witnessing Generative AI for software development in action. And third was hearing a marketing graduate's inspiring story of how she secured her dream job.
Have you seen Cursor or Q Developer in action? If not, you should. The ability to interact with source code or a document through conversation is impressive. Even if you see it as an informed guess that sometimes gets things wrong, once you find the right flow, the feedback loop it creates for building something is incredible. Let me share one of my first attempts to capture the wow moments I experienced when discovering Cursor back in 2024.
The marketing graduate's story was equally compelling: "I had just graduated with no experience and found a job position I desperately wanted. So what did I do? Working with GenAI, I created a complete marketing campaign proposal and submitted it alongside my resume. I didn't just tell them I could do the job—I showed them." After all, if GenAI reduces the effort of generating code or content, why not leverage it for a job application? My thought was "wow, can you do that?" Yes, why not?.
One of the challenges with this GenAI wave is the overwhelming noise. Headlines scream "DeepShark launched a new chat and Nvidia stocks fall by 17%" or "AI will replace your job" daily. This constant barrage doesn't help clarify what matters. What truly helps is experimenting yourself and watching for specific signals—indicators that may reveal you're falling behind.
Here are three warning signs that your organization is at risk of AI disruption:
Consider the case of the legal profession, whose work is heavily influenced by a body of knowledge, interpretations, and applications of precedent. This doesn't mean lawyers will disappear—we'll always need legal expertise—but the nature of legal work is transforming rapidly.
The disruption isn't about elimination but evolution. Entry-level legal positions, where repetitive or non-billable work predominates, are particularly vulnerable. Document review, basic research, and standard contract drafting—traditionally assigned to junior associates—can now be completed by AI systems with increasing accuracy.
This doesn't mean law students should abandon their studies. Rather, it signals the need for a different approach: rapid technology adoption to learn and apply knowledge faster. Forward-thinking firms like LawCyborg in New Zealand demonstrate how AI integration enables lawyers to focus on higher-value work requiring human judgment, ethical consideration, and client relationship management.
Consider the transformation at Australia's Lawpath: Their AI-powered platform reduced customer service inquiries by 25%, increased document creation by 15%, and cut quote lead times from 3 days to half a day. This efficiency gap puts traditional legal service providers at a significant competitive disadvantage in both response time and service scale. Learn more
If your organization's decision-making process involves extensive human analysis of large datasets while competitors seem to make equally good or better decisions faster, they're likely augmenting human judgment with AI.
A Real-world example: In the financial services and e-commerce sectors, organizations implementing AI fraud detection systems are seeing dramatic improvements in decision quality and speed. For instance, FlightHub Group reduced their transaction abort rate from 5% to under 2% while simultaneously achieving their lowest chargeback rate ever, all without increasing operational costs. Similarly, CDKeys automated over 90% of transactions that previously required manual review, reducing their manual review rate from 10% to less than 1% while cutting fraudulent transactions by 6%. Companies still relying on traditional fraud detection methods are finding themselves at a competitive disadvantage, with slower transaction processing, higher fraud rates, and significantly more resources dedicated to manual reviews. Learn more about AI-powered fraud detection from AWS
Are you an AI skeptic? Let me suggest an exercise. Try it yourself. Imagine you're Michael Scott, working for Dunder Mifflin Paper Company, and at the beginning of your day, you need to deal with your email.
If you open http://robertoallende.github.io/dundermifflinmail, you'll open a fake email client with 30 emails. You have to prioritize, filter the important ones, write answers to those that need responses, and compile a list of tasks derived from these messages.
No need to do all the tasks I'm suggesting, but take a look at the emails and think how long it would take you.
Now, using Amazon Q Business, that job takes less than one minute. Just watch the video below.
Is it perfect? Probably not, although it's quite good. However, given such a high volume of information, it's a great starting point. This isn't a theoretical experiment—it's possible with the tools and models we have right now. The only part missing is the integration.
If your organization still manages information flow the traditional way while competitors are using AI-powered systems, you're likely spending 5-10x more time on administrative tasks - time that could be redirected to innovation and customer engagement.
By the way, did you listen to the music in the video? I "made" it myself. Isn't it perfect? 😀😀😀.
I've seen early versions of similar tools in Gmail and Outlook, but they're not quite there yet. This paragraph will likely be outdated very soon—at least I hope so—but if I were the product owner of these email platforms and not working on comprehensive AI integration, I'd be seriously concerned. They're falling behind.
All it takes is for products like Superhuman to gain momentum, and suddenly Gmail becomes obsolete. You don't want to wait until that happens to start considering adding AI. By then, it might be too late to catch up, and you'll be facing your own Kodak moment—watching as your market share evaporates because you failed to recognize and adapt to a fundamental shift in how people interact with email.
If you watch recent Y Combinator Demo Days, you'll notice a surge of startups working with GenAI. It increasingly looks like traditional SaaS will either be replaced by AI-native products or be forced to integrate AI capabilities to remain competitive. Honestly, I can't predict which scenario will prevail, but I'm willing to bet that the tools we use daily will become obsolete, replaced by solutions that truly understand the user.
Consider this: if an AI tool transforms my email into priorities, which then become actions in my todo list, I (almost) don't need email anymore. I might even be able to move beyond the EMAIL protocol entirely. I don't use email because I love its implementation—I use it because it solves problems beyond mere information exchange. If a product can fundamentally address that underlying need, boom. You become the king of email—or whatever they decide to call this new paradigm.
The competitive advantage isn't just about efficiency - it's about reimagining entire business processes and customer experiences. Companies approaching AI as merely a feature addition rather than a fundamental shift in how they solve problems will find themselves increasingly marginalized.
Generative AI is creating positive outcomes across numerous industries. Organizations of all sizes are discovering valuable applications that enhance productivity, spark innovation, and solve complex problems. No matter your field, there are practical, beneficial ways to implement this technology.
AWS has documented numerous success stories across sectors:
- Manufacturing companies like Holcim reduced manual invoice processing by 90%, while Samsung empowered business analysts to forecast demand without coding experience
- Oil & Gas firms like Bravante RRC developed AI solutions to identify deactivated offshore pipelines requiring removal, reducing planning errors and costly mistakes
- Transportation leaders like KONE use AWS IoT and generative AI to create smarter buildings, accelerating innovation and time to market
- Robotics innovators like KABAM created cloud-based management platforms that enhance robot intelligence with large language model capabilities
Explore these and many more AI transformation stories to discover applications relevant to your industry at AWS AI Case Studies.
If you prompt any GenAI to generate a watch showing 3:58 or 6:20 or 3:45, it will likely show 10:09 regardless of what time you request. This happens because most images it was trained on display watches at 10:09—the industry standard for showcasing timepieces in sales materials. Similarly, ask any GenAI to draw a person writing with their left hand, and it invariably depicts the person holding the pen in their right hand, reflecting the right-handed bias in training data.
I've heard many AI skeptics cite examples like these to argue that this isn't a trustworthy technology. That might be fair, and perhaps I'm overly optimistic. But consider this: even if Amazon Q Business's email classification isn't perfect, how long would it take you to read all thirty emails, prioritize them, and identify actions manually? What if you use AI as a first approach, then refine the results? Isn't that still useful and time-saving?
More critically, what happens if your competitor—whether another business or another person applying for the same job position—is using these tools while you resist them? In a free market economy, those providing better products at better prices usually win. The competitive advantage of AI adoption isn't theoretical—it's measurable in hours saved, opportunities seized, and innovation accelerated.
For many years, I was captivated by Björk's music. What drew me in was her unique fusion of haunting vocals with electronic beats and classical instruments like cellos and violins. Then, midway through her career, she pivoted—stripping away the electronic elements and orchestral arrangements. That change severed the connection I had with her music.
Last Christmas, I had a revealing conversation with a musician friend who proudly told me he performs only with instruments he can physically play during gigs. "I want to be authentic," he explained, rejecting technological assistance on principle.
His stance is admirable in its way. But as a music consumer, I realize that when I listen to music, I rarely care about the production methods behind it. What matters is the emotional resonance—the connection I feel.
This mirrors how we should think about product development and AI. In the same way an artist seeks connection with their audience, solving a real problem effectively matters more than the specific tools used to achieve that solution. Of course, you must ensure you're using compliant and safe technologies—that's non-negotiable. But once those requirements are met, refusing to leverage powerful new tools like generative AI isn't principled; it's self-limiting.
Deep down, you likely already know where your organization stands with AI adoption. That gnawing concern when you see competitors launching new capabilities in half the time it would take your team. The slight unease when reading about AI transformations in your industry while your own digital initiatives remain stalled in committees and pilot phases.
The signs are there if you're willing to see them. Your teams working the same way they did a year ago while the world accelerates around you. Customer expectations shifting faster than your ability to meet them. Competitors seemingly operating on a different timeline altogether.
Here's the thing, I bet you already know the answer or at least you may have the concern. In both cases, the only way forward is to open your mind and learn about it. No checklist or score will tell you what your instincts already know.
The real question isn't whether you're at risk—in a rapidly transforming landscape, every organization that stands still is. The question is whether you're ready to acknowledge it and take meaningful action before that risk transforms from potential disruption to irreversible decline.
- Start experimenting today. Choose one high-volume, low-risk process in your organization and test how AI could transform it. The email exercise above is a perfect example - implement something similar in your environment.
- Identify your own "wow" moment. As I shared with my music composition experience, the true power of AI often becomes clear through personal experimentation. Encourage key decision-makers to find their own breakthrough use case.
- Map your competitive landscape. Which of your competitors are already showing signs of AI adoption? What capabilities or efficiencies are they demonstrating that weren't present six months ago?
In the next post, I'll share my personal definition of what "AI First" means, backed by concrete examples of transformative impact when applied to the right use cases.
Many professionals feel like using AI is somehow "cheating" – as if the machine is doing their job and they're adding no value. Others remain skeptical, pointing out that AI isn't always right. These are two sides of the same misguided coin.
Think about it: If you ask AI for assistance and it provides the right answer – then what? If you ask AI something and it's wrong – then what? The real transformation lies in how you engage with these outcomes, and that's precisely what I'll explore in upcoming posts.
In the meantime, don't wait to get started – the gap between AI adopters and laggards is widening every day. The most dangerous position is believing you have time to wait and see. By the time AI's impact becomes obvious in your industry's metrics, the leaders will have already established nearly insurmountable advantages.
Will you be among them, or will you be the next cautionary tale?
Thank you very much for reading. Please, note that the opinions written in this article are my own and don't necessarily reflect my employer, AWS and any other company mentioned in this post.