Alexa

30 Innovative Brands of the Year 2021

Tête-à-Tête with Florian Quarré, ExponentialAI Chief Strategy Officer: ‘We’re Devoted to Accelerating Evidence-Based Decision Making and Enterprise Automation Using AI’

thesiliconreview-florian-quarre-cso-exponentialai-21.jpg

“The operational transformation over time with the adoption of AI leads to a symbiotic corrective environment.”

ExponentialAI is a leader in enterprise AI for digital transformation. Enabled by their Platform, Enso, ExponentialAI accelerates the process of building, deploying, and scaling enterprise-grade AI solutions, enabling evidence-based decision making, accelerating automation of complex enterprise processes, and continuously gathering feedback to curate the strategic and operational knowledge of the models it operates. The company’s enterprise-grade platform enables organizations to create value in production in less than three months.

ExponentialAI partners with leading Fortune 50 clients across Healthcare, Life Sciences, Retail, and Financial Services, accelerating their enterprise digital transformations with AI.

The company was established in 2016 and is based in Atlanta, Georgia.

Florian Quarré, ExponentialAI Chief Strategy Officer, spoke exclusively to The Silicon Review. Below is an excerpt.

TSR: Talking about digital transformation and accelerated decision-making process, how do you manage to understand a company’s culture before strategizing a transformation plan?

Florian Quarré: That culture, as you hint upon, is always different from one company to another. How much of the culture influences their digital transformation varies depending on the degree of maturity. We use a mix of art and science to understand companies better.

On the art side, it is our experience of engaging for decades with healthcare organizations, knowing operations as former operators ourselves that gives us a good sense of their readiness for the transformation, and ultimately enabling decision-making with AI. For example, we know that Administrative Audits and payments reconciliation functions have been starving for drastic transformation accelerators such as AI.

Simultaneously, the science aspect is more of looking into the operations, their KPIs, and their level of operational debt. We try to understand and map out in a very traditional manner with a high-level value chain to understand the various operations. For example, which areas could benefit from the immediate transformation? What areas are ready to adopt this decision intelligence in ways where AI or intelligent automation are components that can be leveraged to help accelerate the transformation?

Therefore, strategizing a transformation plan brings together balancing out culture and operational debt, vetting the art and science, that we typically work on at the very onset of collaborating with new clients.

TSR: Internal inefficiencies like the lack of intelligence and unorganized operations can undermine an organization’s overall output. Do you help your clients patch up their internal inefficiencies?

Florian Quarré: We do. That’s a great question. I think the way you phrase it is correct. We look into internal efficiencies or lack of efficiencies.

Let’s talk a little bit about the stages of decision intelligence and patching up those kinds of inefficiencies. The first part is—what is inefficiency? Some organizations may find a manual process that relies on their workforce to be sufficient and satisfactory. In contrast, others may be more interested in automating that as much as possible. So, identifying what qualifies as inefficiencies is the first aspect.

The second part of what we find is that ultimately it’s all about data. The ability to bring in the enterprise AI systems, expert systems, or decision intelligence systems leans on the access of representative information. You have to have information in breadth and depth, a time series that helps you understand enough of your business over time and with enough details so that you can see patterns emerge. With the help of technology, in our case Machine Learning, you should identify the signals that contribute to influencing a specific behavior that you want to curb or change to transform the way an organization operates.

In the third and final stage, our system delivers recommendations for action and pairs up with a continuous curation of all operational metrics. Every step performed in that ecosystem becomes an opportunity to improve upon and continuously teaches the system on the signals we have surfaced. Inefficiencies generating enterprise operational debt are finally addressed with a varying degree of automation and collaboration between the AI agents and the operators of the company, yielding a highly qualitative operational throughput.

TSR: No one is a hundred percent sure of every decision that is made. There is no single fit-for-all methodology. How is ExponentialAI making this better?

Florian Quarré: Everything you said is so right. We believe in three things—transparency, explicability, and trust. You could argue that transparency and explicability lead to trust.

The ultimate hypothesis is that when you look into human error, we learn very fast and get exhausted under pressure. We found that depending on the task, there’s a point of about 70-85 percent accuracy on a hundred paths human operators perform daily. Some people are better than that, some are a little bit worse, but what we find in an operational science manner is that most people fall within that band.

No system will pull you all the way to 100 percent, and perhaps what we’re also suggesting is that no organization should try to go through a transformation with the goal to achieve 100 percent accuracy across all tasks. However, what if you were to adopt a system that gives you the certainty of attaining 85 percent consistent accuracy, or maybe even up to 90 or 95 percent accuracy in a systematic manner?

As I indicated earlier, as you surface the inefficiencies, you showcase the signals of problems and operational debt. This way, you have the working framework to hammer through all of that operational debt. You’re going to find that the automation and decision intelligence that come into play will provide the lift for the organization to reduce their operational debt or inefficiencies dramatically.

Now, back to transparency, explicability, and trust. With our system, we make sure that we transparently communicate the operation and the management of statistical AI models in our platform so that it is traceable back to enterprise operations. We believe that having an ability to look at AI models’ behavior in more of a grey box approach rather than a complete black box, where you understand the consumed signals participate in adoption.

Next, we deliver explicability by providing insights on which of the hyper-parameters are used by the models that provide specific recommendations, identifying what part of the data is used against those parameters, and why the system has recommended a particular action to be performed.

As traceability and explicability are delivered, we train the organization to engage and collaborate with a system they can understand, a system they can relate to. This is when trust appears, and the collaboration between the system and the organization becomes symbiotic, creating a greater adoption towards reducing operational debt and inefficiencies.

That’s how we see that operational transformation over time with the adoption of AI leads to a symbiotic corrective environment.

TSR: Do you have anything else you would like to add?

Florian Quarré: While we are a general-purpose platform, we’ve doubled down in the healthcare industry with a strong belief that leading with a platform pre-wired with a healthcare ontology, taxonomy, and expertise enable us to solve for the majority of use cases we are exposed to when joining our healthcare clients. We believe that solving the industry problems in a very targeted manner adds a layer of expertise that injects itself back into the models themselves, ensures a continuous refinement of the models’ accuracy, and creates greater enterprise value to our clients, far more quickly than if we took a more generic approach.

However, it doesn’t mean that we do not see tremendous value generated in other areas. We work with retail and financial services clients who adopt our technology for a very different purpose, mainly focused on the deep complexity of information, extraction, and harmonization. Enso delivers great technical acceleration for them,  which ultimately benefits all our clients using our ecosystem.

The Leader at the Helm of Exponential AI

Florian Quarre is a technology executive in the health information technology space with over 15 years of expertise leading and advising the transformation of large healthcare organizations within the US and globally, using emerging technologies – such as blockchain and AI – with an accelerated incubation path to swiftly deliver enterprise business value. Florian was the Chief Digital Officer for Ciox Health prior to joining ExponentialAI, before which led Deloitte Consulting’s healthcare blockchain practice, focusing on the adoption of blockchain, challenging the status quo on the realm of possibilities for the industry.

“The purpose of our platform is to bring together development and production of AI models so that we can evolve from a problem that you are tinkering on and have the ability to bring it all the way to production with the scale and strength of a business that runs millions of analyses.”