Whether it’s a cocktail party or a networking event, the inevitable question always comes, “What do you do for a living?” My oft-repeated response is to rattle off some variation of I help organizations implement technology to create business value, blah, blah, blah.
And then, as a preemptive attempt to keep them from being bored to tears, I throw in that I also do “fun stuff” like emerging tech. And that is the usual spot in the conversation where their eyes shift from glazed-over to twinkling. People love to gab about the exciting new stuff like self-driving cars, smart homes, and quantum machines. If I am not careful, I will get trapped in an endless conversation and miss the hors d’oeuvres.
Yes, emerging tech makes an amazingly good small talk subject. But for all its shiny newness, it is still a technology tool. My job is the same whether I am dealing with ERP systems or Emerging Tech – help clients envision its use, implement into their organization and optimize their investment.
In this series, I will share how companies can envision, implement, and optimize in three different Emerging Tech areas – Artificial Intelligence (AI), Internet of Things (IoT), and Quantum Computing.
So, grab a canape or two and let’s get started.
It is difficult to think of a major industry that AI will not transform. This includes healthcare, education, transportation, retail, communications, and agriculture. There are surprisingly clear paths for AI to make a big difference in all of these industries. Andrew Ng
We are only beginning to unleash AI’s potential.
As a mother of two tweens, I am paying close attention to the rise of self-driving cars. But there are many other interesting AI applications on the near horizon, too – including diagnosing medical conditions, predictive maintenance in manufacturing, and virtual reality in insurance.
That said, with great power comes great responsibility. Keeping bias out of our AI algorithms and applications needs to be a top priority now before AI’s soon-to-be ubiquity makes it impossible to correct.
As breathless as I get when I think about AI’s accomplishments so far, I still want to throw my Amazon Echo across the room when Alexa cannot understand the simplest of my commands. We are at an interesting point with AI. We use it regularly enough in normal life that we have a taste of its potential. But, we also know it is still rudimentary and ripe for improvement. So, the question is not, “what business problems could you solve with natural language processing, image recognition, and deep problem-solving?”
Instead, the question becomes, “what business problems could you solve with high performing natural language processing, image recognition, and deep problem-solving?”
From practical to policy, there is a lot to consider when implementing AI. On the practical side, will your company leverage off-the-shelf natural language processing, image recognition, and machine learning tools? Will you design and implement your own custom algorithms? Or, will you fall somewhere in between?
Defining your policies requires more decisions, and has a more significant impact, than your tooling. How will you handle data integrity and privacy? What processes do you have in place to prevent bias from entering your algorithms? This is why AI must also live outside the walls of your IT organization. You need the lawyers. You need your HR folks. You need to include the customer experience design team. You need to include people who have rich experiences in realms that do not involve zeros and ones.
You need people more than technology when it comes to AI. And don’t get me started on the need for diversification in AI.
AI without optimization is nothing more than a fun parlor trick. If you don’t integrate and institutionalize your organization’s AI capabilities, you will miss out on true business value.
Alas, given the evolving nature of AI, chasing optimization can feel like being trapped in a house of mirrors. There are a lot of seemingly good options forward but only a few yield positive outcomes. Which processes should stay “as-is” and which should become “smarter” with AI enablement? Who on your team will be trained in these new technologies, and how?
Pragmatism works, even with something as seemingly magical as AI. A disciplined PMO process, that evaluates things like impact, cost, and risk, can help you understand where to optimize your AI investment.
Internet of Things
“When we talk about the Internet of Things, it’s not just putting RFID tags on some dumb thing so we smart people know where that dumb thing is. It’s about embedding intelligence so things become smarter and do more than they were proposed to do.” — Nicholas Negroponte
The Internet of Things is the worker bee of the Emerging Tech scene. It may not have the wow of AI or the hype of Blockchain, but IoT gets things done. On the factory floor, in the hospital unit, or from your wrist, IoT captures and ingests all the information. And when you have all the information, you also have the honeypot of opportunity.
In my opinion, one of the most interesting facets of IoT is the technology’s ability to quantify what humans know inherently. For example, Sally, a seasoned technician, knows in her gut that when machine 237 starts performing a bit sluggishly, it will likely fail in the coming week. Thus, she will keep a closer eye on the machine, and perhaps even order some spare parts in advance. She has an attuned intuition after over thirty years of living and breathing the machines on the factory floor.
But, Sally is human. Humans take lunch breaks. Humans like beach vacations. And eventually, humans like to cash in their 401Ks and retire to the aforementioned beach so they can have endless lunch breaks. Humans simply cannot monitor, analyze, and act twenty-four, seven.
Additionally, humans do not scale. Sally, and her ilk, are incredibly valuable to their employers. But with today’s changing workforce, it is rare to find institutional knowledge cultivated by longevity.
Enter IoT. What do you measure and monitor today? How do you currently apply human intitution? IoT allows you to replicate and expand your current capabilities, whether it’s realtime machine monitoring, observing patient interactions, or making your home smarter.
Remember that cringey hype slogan: “Go big or go home?” It is oddly relevant when we talk the Internet of Things. IoT, by nature, becomes most valuable when it “goes big.” More devices mean more data, more data means more business insights.
But going big has it challenges, the first of which, obviously, is scalability. You may want a modest IoT implementation now, but as soon as you start receiving some data, you will want more and more. Data is an addiction. Scalability is the only cure. Two areas to be particularly mindful of during scalability planning are your hardware capacity and your data processing (databases, machine learning, analytics dashboarding, and more.)
Another important consideration is security. Security is always necessary, of course. But once you start scaling, you become a juicy target for nefarious nerds. There is a lot of media chatter about eavesdropping surrounding the Amazon Echo and Google Home. But, our connected thermostats, vehicles, and industrial equipment are just as, if not more, vulnerable.
Data dies on a spreadsheet. IoT gives us the capability to capture a tremendous amount of data, but it’s merely data until you do something with it. And, I am not talking slapping it into Excel every few weeks. Data lives when it is accessible. Real-time monitoring is the first step towards letting your IoT data live. Whether you use PowerBI, Tableau or a homegrown dashboarding tool, seeing your data is an important first step towards accessibility.
But I would take it a step further. Dashboarding is reactive analytics. Yes, it’s nice to see what has happened in the past, and we can probably make some good future inferences once we sit down and really digest all the pie graphs and bar charts. But, in the end, we are reactionary to something that happened yesterday, three days ago, or a month ago.
What if we could react to something that will happen to tomorrow, three days from now, or a month in the future? What if we use predictive analytics? That is where IoT and Machine Learning come in. ML thrives when it has a lot of data to use for learning and predicting. IoT creates a lot of data. IoT and ML are the peanut butter and jelly of emerging tech.
By applying ML to your IoT data, you can solve future problems such as when your machines will need replacement parts or the expected optimal staffing levels for your hospital.
“As we all know, in our everyday life, a cat can only either be in an alive or dead state. A cat in the quantum world can be in a coherent superposition of alive and dead states.” – Chinese physicist Pan Jian-Wei, dubbed “the Father of Quantum.”
I first studied Quantum Computing in grad school, which was almost two decades ago. At that time, it was a fun academic exercise, far removed from reality. It swelled with potential, but the practical implications were less tangible. Today, quantum is more real and requires us to rethink computing, security, and the seemingly impossible.
Without going into the gory details, Quantum computers allows us to go beyond zeros and ones. Quantum bits (or Qubits) make this happen. Qubits can be both zero and one at the same time (mind blown!), allowing for substantially more processing power. This power means you can process significantly more complex algorithms in the fields of machine learning, physical sciences, and financial services. And, oh yeah, they will completely destroy our current encryption practices.
So, about that last line ‘completely destroy our current encryption practices” … According to the Washington Post, Quantum computers might someday be able to crack all existing forms of encryption. In fact, the Chinese credit notorious American spy Edward Snowden for jump-starting their investment in Quantum research. Snowden’s leaks motivated China to double down on next-generation cybersecurity.
While we are still a few years away from mass decryption chaos, organizations should start considering, now, what their future security practices may look like and developing nimble, responsive processes.
Another important, albeit grossly unsexy, consideration: cost. At the onset of their commercial availability, Quantum computers will, undoubtedly, be more expensive than your run-of-the-mill traditional cloud computing. Time to start sharpening your pencils to find room in your technology procurement budget.
Unless your initials are NSA, I suspect your organization doesn’t need any sort of Quantum, yet. So, it’s all theoretical at this point. But, it is still a worthwhile exercise to inventory your current business processes and problems. Which ones are the most complex, today, gobbling up your computing resources? Which ones are the nagging problems, still begging for a solution? Can any of these benefit from a new tool in your chest, namely the (insanely) high-powered problem solver that is Quantum Computing?
Over the coming months, I will dive deep on each of these three areas, exploring how you can empower your organization to create new business value.