Thank you so much for joining us in this interview series. Before we dive into our discussion our readers would love to “get to know you” a bit better. Can you share with us the backstory about what brought you to your specific career path?
Thank you for having me, and hello to the Authority Magazine community!
In 1997, I got the life-changing opportunity to become a product manager for a Montreal-based tech company. The company had partnered with Netscape Communications to develop standards-based software for businesses all over the world — competing with IBM and Microsoft. It was a time of great innovation, working to establish interoperability between competitive solutions, enable mobile access through synchronization and over-the-air wireless protocols, as well as develop security standards. After the company was acquired by Oracle in 2002, I moved to the Bay Area and expanded my responsibilities to lead product organizations. I have never looked back, and still love building great products and teams for big tech companies, small startups, and every size in between.
What do you think makes your company stand out? Can you share a story?
Lirio’s story is about purpose, innovation, and resilience.
We have set out to tackle a really hard problem that affects everyone: health behavior change. Almost half of the risk associated with preventable and early deaths can be traced back to human behavior, yet interventions designed to help people change their behavior are often not effective. Today, the U.S. spends nearly 20% of its Gross Domestic Product (GDP) on healthcare. According to the Centers for Disease Control and Prevention (CDC), 90% of all healthcare costs in the U.S. go toward treating chronic disease and mental health — most of which can be prevented through behavior modification. Lirio’s purpose is meaningful to each of our employees, which allows us to attract and retain incredible talent. Out of the 56 colleagues, we have 15 with a Ph.D., and many more with master’s degrees; something we could not afford if we were competing for salary alone.
Research shows that new ways to effectively design and implement behavioral interventions are needed to address the problem. Humans are not good at implementing sudden radical changes, instead we do better with small, daily improvements to our behaviors that support us in forming better habits over time. What motivates us and gets in the way of improving our health is unique for each of us, and changes over time. The innovative approach Lirio is taking to tackle behavior change is through highly personalized engagement across the spectrum of health behaviors.
We call our approach Precision Nudging, and it’s born out of the belief that the combination of multidisciplinary science, digital technology, and the abundance of feedback data enables us to meet each person where they are, every day, to nudge them along their personal journey to better health. Finally, the macro environment for HealthTech startups has been challenging for the past few years marked by a global pandemic, a struggling healthcare sector, a talent gap, and the shortage of risk capital. We have been resilient thus far in being able to keep our team together, find investors who share our sense of purpose, and partner with organizations who recognize the need to take a personalized, data-driven approach to tackling behavior change and achieving sustainable health outcomes. In more ways than one, Lirio is a small miracle.
You are a successful business leader. Which three character traits do you think were most instrumental to your success? Can you please share a story or example for each?
As a tech entrepreneur, I’m acutely aware that technology is merely a means to an end — helping people be more productive, healthier, happier, safer, etc. Through the years, my leadership philosophy has become more intentional about putting people first, always. As a result, there are a few “abilities” that have helped me be more successful as a leader, perhaps more so than specific “character traits”:
- Bringing clarity. As leaders we must bring clarity to the people we work with. In the intersection of technology and healthcare, a big part of this challenge is to simplify the complex. Why are we here? What does success look like? How does what I do impact the company? What conduct is acceptable, and what is not? Why is this a priority? What changed? Clarity will help motivate people, and it will help us hold each other accountable. This is not just about the big picture, but about filling in the small gaps every day.
- Building a collaborative culture. Steve Jobs famously said: “It makes no sense to hire smart people and tell them what to do.” If you want to make things happen, you certainly cannot micromanage others, but you also must create a collaborative environment where smart people can be creative, take risks, and grow together. A workplace that looks beyond mere individual contributions and values collaboration; empowering people to get stuff done efficiently without concerns about organizational hierarchy or personal recognition (i.e. “that’s not my job, that’s above my paygrade”).
- Leading with compassion. A mentor once told me that leadership is not for the faint of heart. The idea of leading people to places and coping with things that you yourself have not always experienced takes both self-awareness and selflessness. Every leader is dealt a unique set of circumstances and must make many tough decisions in the face of uncertainty. By actively listening and understanding people’s needs, you can better grasp the impact of your decisions. If you are compassionate, meaning you can do hard things in a human way, you will build trust and strengthen the social fabric throughout the organization. I am grateful that this trait is shared by my executive leadership team and our board of directors. It is difficult to imagine having navigated the many crises of the past four years without compassionate leadership at multiple levels.
Let’s now move to the main point of our discussion about AI. Can you explain how AI is disrupting your industry? Is this disruption hurting or helping your bottom line?
For years, the healthcare sector struggled to recover from IBM Watson’s hype around how AI-powered clinical decision support would reduce diagnosis errors, optimize treatments, and even counteract workforce shortages. The fact that Watson became a multibillion-dollar flop led to widespread cynicism about the applications of AI in healthcare.
Fast forward to today, AI solutions are having a real impact on healthcare from patient risk identification to disease diagnosis, to drug discovery. While still in the early stages of adoption, generative AI is promising to tackle complex use cases that span the healthcare ecosystem, such as streamlining billing and claims processes, optimizing the coordination of care resources, and improving the quality of patient care.
For Lirio, the growing demand for AI solutions in healthcare is helping our bottom line. Most healthcare organizations will now adopt AI applications that show a clear return on investment (ROI), and plan to do so through partnerships with smaller technology organizations with specialized solutions — such as Lirio — rather than with big technology companies with general-purpose AI platforms or developing in-house solutions.
Which specific AI technology has had the most significant impact on your industry?
Rather than naming a specific technology, let me point to the recent global pandemic as an event that had the most significant impact on healthcare adoption of AI. COVID-19 vaccines were developed and produced at unprecedented speed. Of note, mRNA vaccines had not previously been used to treat or prevent any other disease. It is well documented how AI helped accelerate the research, development, testing, and distribution of the COVID-19 vaccines as part of the pandemic response. The COVID-19 pandemic also served as a catalyst for the adoption of digital solutions that improve patient access, both in the hospital and at home.
A perfect storm is brewing, brought on by a change in weather patterns, population growth, globalization, and the deterioration of the public health infrastructure. These factors will have direct and indirect consequences on many aspects of patient health, but particularly on infections, especially among water- and vector-borne diseases, but also fungal infections, antimicrobial resistance, and others. The ability to rapidly develop and distribute new drugs and vaccines will become a necessary human superpower this decade.
Can you share a pivotal moment when you recognized the profound impact AI would have on your sector?
Several years ago, I was part of a team that partnered with HCA in Nashville to accelerate patient navigation services for cancer patients by developing an AI tool that intelligently refers patients diagnosed with cancer to a nurse navigator.
At HCA’s Sarah Cannon Research Institute (SCRI), this tool was initially launched to nurse navigators across 65 hospitals in seven states. In 2018, the nurse navigators saw nearly a 60% year-over-year increase in new patient volume because of the tool. The boost in patient loyalty was significant, especially for breast cancer care, where patients had historically only stayed in network about 50% of the time; because of the navigation program, more than 90% of patients who started their treatment early remained in network. It proved the thesis that AI solutions in healthcare can have a double bottom line: a win for patients and good for business.
How are you preparing your workforce for the integration of AI, and what skills do you believe will be most valuable in an AI-enhanced future?
AI adoption is not an option, it is imperative for any business that wishes to remain relevant. Not all organizations and business functions are going to be at the same level of readiness. Here are some initiatives you should consider:
- Build a robust data and technology infrastructure. Data is the fuel for all AI solutions. Invest in a data management layer that is secure and can easily scale to support a wide range of AI-driven workflows and applications across your business.
- Develop a team and culture that embraces AI. Assembling a team with AI expertise will require you to evolve your hiring, management, and operations. It also requires a growth mindset that is willing to experiment with the AI tools, and a leadership commitment to implementing ethical AI practices across the company. One of the most important skills in the age of AI is critical thinking. Leveraging AI alongside human critical thinking can create a powerful synergy. While AI provides data-driven insights, humans can interpret these insights within a broader context, considering ethical, social, and long-term implications.
- Define clear business priorities and performance goals, and measure progress. Resist the hype. Define clear goals for AI adoption and ensure there are measurable business goals aligned with the adoption. Start with quick wins that offer short-term ROI, but be prepared to transition to more complex projects that drive long-term transformation.
What are the biggest challenges in upskilling your workforce for an AI-centric future?
The demand for skilled AI talent is surging, but the talent pool remains stale. Finding talent is even more difficult in healthcare. We need to upskill the workforce for an AI-centric future, but it won’t happen without a few speed bumps. Those who are hoping to train and hire for AI-skilled positions must understand:
- Training and finding talented individuals will take time: Many people will need to learn the fundamentals of AI. As a nation, we must also improve mathematics education and learn to develop human agency. Then, it will take time for those individuals to apply learnings within a specific discipline — like machine learning in healthcare.
- Healthcare must be ready to compete — and pay for — AI skills: Healthcare organizations will struggle to compete with the likes of Google, Facebook, and OpenAI for the talent they need; and must be ready to pay a premium for skilled people.
- Companies like Lirio can fill in the talent gaps: Smaller companies like Lirio are here to step into the void so healthcare organizations can leverage the AI expertise and intellectual property developed to support healthcare use cases.
What ethical considerations does AI introduce into your industry, and how are you tackling these concerns?
For the development of AI for healthcare, there are several ethical principles published by various organizations, including the American Medical Association (AMA), the World Health Organization (WHO), and the Coalition for Health AI (CHAI), to name a few. The AMA and WHO both highlight a human-centered design philosophy protecting human autonomy and explicitly mention the need for inclusiveness and equity in the healthcare use of AI to prevent care disparity. The blueprint for an AI Bill of Rights published by the U.S. Office of Science and Technology Policy (OSTP) has provisions for AI systems to be safe and effective, protected against algorithmic discrimination, protect user data, have accessible documentation, and offer human alternatives.
We shouldn’t think too narrowly about tackling ethical concerns. No set of principles encompasses all ethical concerns that healthcare organizations or patients may have. Some decisions should be made at a governmental level, while others must be made by healthcare organizations or individuals using the solutions. As AI becomes capable of making predictions that are computationally equivalent to physicians, we could even consider establishing a “Hippocratic Oath” or certification for AI tools that align with medical ethics principles, including autonomy, beneficence, nonmaleficence, and justice.
At Lirio, we have built a culture around the responsible and transparent use of AI. We participate in organizations like CHAI and continue to offer advice to various government agencies requesting input and information to improve AI governance. From a practical point of view, we have extended our compliance posture and risk management framework to be inclusive of policies and controls related to the research and development of AI technology.
What are your “Five Things You Need To Do, If AI Is Disrupting Your Industry”?
This question feels like it belongs in The Hitchhiker’s Guide to the Galaxy. Perhaps it is the ultimate question of AI, the Universe, and Everything? Here are my top 5:
- Know the problem you’re trying to solve with AI: You must precisely define what problem you’re trying to solve with AI. There is no universal learner, so you must have intuition about how you’re planning to model the real world through your data and similarly which algorithms are going to be most relevant to learning the patterns you care about to accurately predict the outcomes you’re expecting. There are literally an infinite number of models that can “explain” any dataset. The data elements you should try to collect and the accuracy you require of their measurement are dependent upon the problem. The biggest consequence of that is that if you’re not extremely diligent and careful, then most of the models you might try to create to fit a dataset will perform poorly in the real world, even if they look to be perfect based on your training set, or even based on other measures that might be too myopic.
Consider the well-known example of social media news recommender systems. When they were originally designed, the goal was to build models that learn what someone is interested in. However, they rewarded/optimized these models based on someone reading more articles. But reading more and being interested in are not 100% correlated. What the models learned was that putting more extreme content in front people led those people to read more, regardless of their original interests. That has led to all sorts of negative consequences in our society.
2. Reimagine a better experience after implementing AI: Online banking, e-commerce, streaming services, and even email and voice assistants like Apple’s Siri — these are all examples where technological innovation goes hand in hand with a much-improved user experience. These solutions weren’t just changing out one part for another. When you start to apply AI to your business, start with the impact you want the solution to have and make sure you develop the right technology integrations, workflow optimizations, and yes, user experience.
3. Define clear metrics for measuring AI’s impact: You want continued improvements over time. That’s why you must establish benchmarks to help you see how the solution is performing — and getting better over time. For example, if you want to scale out or improve upon what an expert human can do, then you likely want to define the human performance benchmark to beat. One example of how this is estimated is the “inter-annotator disagreement.” In other words, you get several humans to do an annotation job, like labeling/classifying the data you’re using to test your solution, and you see how much they disagree. If an automated solution can achieve a level of accuracy that meets or exceeds the human-level agreement, then that is a good level of performance.
4. Understand the bias and risks of your AI solution: In machine learning, “bias” is not inherently good or bad, but it is necessary. First, what is the bias in an AI model? Well, even if you only consider the obvious things, your AI solution is biased based on (1) the samples you give it to train it, e.g., if all participants are male or of a certain race, etc. But note that if you are building a solution for prostate cancer, all your data should come from men, so e.g., bias is sometimes critical to get a good solution. (2) the data elements you measure or collect about your samples, e.g., if you’re building a model to identify diabetes risk, should you include someone’s income? Probably not, but should you include someone’s family history of diabetes? Probably. (3) the form of your model, e.g., if your modeling includes A1C over time, then you’d want a form that can capture temporal dependencies. All these elements contribute to the bias, and they can be good or bad, depending on the problem you’re trying to solve. Without any bias (i.e., expertise), you will inevitably fail, since there is no universal learner. And as for the risks of those biases, there are many, with the one most often cited in healthcare being the risk of inequitable health outcomes. However, it’s critical to recognize that such risks have been prevalent in healthcare long before AI, and because understanding the biases in your model and data is such a critical factor in the success of building ML models in general, you are potentially more likely to surface such problems when applying ML than you might be if you were using some other approach.
5. Consider the human and cultural impact on your organization. There will be a human impact when AI tools are used to try to save costs, automate processes, and drive revenue. Some jobs may be eliminated, while other roles may require new skills. There is a change management aspect. People need to understand how the decision to implement AI solutions will impact their people. There is an opportunity for executives to focus on the positive aspects of evolving their company culture to embrace AI, especially when trying to create novel products and services. You must cultivate a growth mindset. Whatever you try first is very unlikely to work, even if you have a lot of expertise. You may need to collect new, real, noisy data as a prerequisite to a large part of any real progress. You will likely need to iterate many times on the design. When you find that your solution “works,” it’s just the beginning, because you will want to improve it and likely can do so dramatically, once you find the right path. AI solutions are learning based, probabilistic, noisier, and more complex than traditional business solutions. Even if your AI team understands this, it is important to set these expectations with others.
What are the most common misconceptions about AI within your industry, and how do you address them?
It is true that if left unchecked, AI might represent an existential threat to humanity. However, AI is also the human superpower that can make our world a better place for humans to survive and thrive in. AI carries enormous potential to benefit patients, doctors, and hospital staff. Healthcare’s current state is untenable. We have a healthcare sector plagued by workforce shortages, burnout, bankruptcy, poor clinical outcomes — the list goes on.
AI’s broader adoption could help doctors and healthcare workers deliver higher-quality and more empathetic care to patients in communities across the country while cutting healthcare costs by hundreds of billions of dollars annually. It could also help patients make more informed health choices by better understanding their health conditions and needs. We have the AI tools to help, and we should use them, responsibly.
Can you please give us your favorite “Life Lesson Quote”? Do you have a story about how that was relevant in your life?
“The hardest problems are people problems,” is a quote from a former colleague who now leads a genetic testing company. I think about this quote a lot. It is also helpful to keep this in mind as we contemplate the intersection of healthcare, AI, and the future. Many of the threats attributed to the development and adoption of AI arise from the deliberate, accidental, or careless misuse of AI by humans. We should focus on those immediate threats as we contemplate the right balance between regulation and innovation.
Off-topic, but I’m curious. As someone steering the ship, what thoughts or concerns often keep you awake at night? How do those thoughts influence your daily decision-making process?
Over the past four years, we have witnessed growing levels of uncertainty at many levels, including political, social, economic, technological, etc. We have seen increasing threats to human health and well-being from disease, changing weather patterns, social media, and so on. Personally, I believe that leaders must evolve and develop core skills in situational analysis, build resilient cultures that can adapt to change, and reward people’s efforts over specific outcomes. We also must be kinder to each other and recognize that compassion and acceptance are great coping mechanisms for fear and anxiety.
You are a person of great influence. If you could start a movement that would bring the most amount of good to the most amount of people, what would that be? You never know what your idea can trigger. 🙂
When I think about this question, it reminds me of John F. Kennedy, Jr.’s famous quote, “Ask not what your country can do for you, but what you can do for your country.”
I would like to harness the reach and power of digital tools to create a global movement for KINDness — a “marketplace” where people can offer their unique skills and resources pro bono to help others. This would have to be governed and monitored more rigorously than any other social network or digital space, but it would enable activities of kindness without any money, property or products being exchanged.
How can our readers further follow you online?
Thank you again for having me! I encourage your readers to reach out and stay in touch. You can connect with me on LinkedIn or X.
Thank you for the time you spent sharing these fantastic insights. We wish you only continued success in your great work!