Straumann Group’s Sridhar Iyengar has a bold mission: To transform the nearly 70-year-old company’s data and technology organization into a data-as-a-service provider for the global manufacturer and supplier of dental implants, prosthetics, orthodontics, and digital dentistry — and to provide business stakeholders machine learning (ML) as a service as well.
“My vision is that I can give the keys to my businesses to manage their data and run their data on their own, as opposed to the Data & Tech team being at the center and helping them out,” says Iyengar, director of Data & Tech at Straumann Group North America.
Doing so will be no small feat. The Basel, Switzerland-based company, which operates in more than 100 countries, has petabytes of data, including highly structured customer data, data about treatments and lab requests, operational data, and a massive, growing volume of unstructured data, particularly imaging data. The company’s orthodontics business, for instance, makes heavy use of image processing to the point that unstructured data is growing at a pace of roughly 20% to 25% per month.
Advances in imaging technology present Straumann Group with the opportunity to provide its customers with new capabilities to offer their clients. For example, imaging data can be used to show patients how an aligner will change their appearance over time.
“It gives a lot of power to our providers in selling their services and at the same time gets more NPS [net promoter score] for us from the patient,” says Iyengar, who believes AI will play a critical role in Straumann’s image processing and lab treatments businesses. Hence the drive to provide ML as a service to the Data & Tech team’s internal customers.
“All they would have to do is just build their model and run with it,” he says.
But to augment its various businesses with ML and AI, Iyengar’s team first had to break down data silos within the organization and transform the company’s data operations.
“Digitizing was our first stake at the table in our data journey,” he says.
Selling the value of data transformation
Iyengar and his team are 18 months into a three- to five-year journey that started by building out the data layer — corralling data sources such as ERP, CRM, and legacy databases into data warehouses for structured data and data lakes for unstructured data.
That step, primarily undertaken by developers and data architects, established data governance and data integration. Now, the team’s information architects, in conjunction with business analysts, are working on the semantic layer, which feeds data from data warehouses and data lakes into data marts, including a finance mart, sales mart, supply chain mart, and market mart. The next goal, with the aid of partner Findability Sciences, will be to build out ML and AI pipelines into an information delivery layer that can support predictive and prescriptive analytics.
“As the information layer gets mature, that’s where the ML and the AI will start seeing some green shoots,” he says, adding that although data transformation was a pressing need when he signed on in 2021, he wanted a more compelling vision to sell the board and business leaders on tackling it.
For that, he relied on a defensive and offensive metaphor for his data strategy. The defensive side includes traditional elements of data management, such as data governance and data quality. The offensive side? That is the domain of AI and advanced analytics that serve a role beyond just insight and business optimization.
“The offensive side is how to generate revenue, all of the insights from the historical data that we have collected and, in fact, forecast the trends that are coming,” Iyengar says. “Most of the data that we get on the offensive side are unstructured, and we want to make sure that it makes sense to the business leaders and help them harmonize and enrich it in such a manner that they can serve their customers more efficiently and that the customers get served and leverage Straumann’s services in a much more robust, frictionless manner.”
Not surprisingly, it was this offensive side that got Straumann’s board invested in Iyengar’s plan for transformation.
“When the customer-centricity and the digital transformation piece was proposed — along with data transformation — I think that resonated with them,” Iyengar says.
Skilling up for the future
Iyengar’s team found success by adopting a use-case approach, not unlike that of one of Strauman’s core businesses. “We pretty much took the same principle of the pre-treatment and the post-treatment images that we show to our patients,” Iyengar says.
The team asked company leaders to pick a number of customer-centric vectors to illustrate how data innovations could be used to drive business outcomes. One of the targets was driving down customer churn. The team started by splitting churn propensity into two values: one for retention of existing customers and one for new customer acquisition. It used typical customer lifetime values and analyzed buying patterns to provide the marketing team and sales team with insights they could use to drive their strategies.
Iyengar says adopting this approach to selling digital transformation internally has made the job much easier. “We are seeing a lot of investments being approved from all the businesses in order to support that initiative,” he says.
In the meantime, as the team begins to build out ML and AI capabilities, it is also imperative to transform the Data & Tech team itself.
“The skill set that we have inherently from our traditional school point of view doesn’t suit the ML and AI part of it,” Iyengar says. “What you need there is statisticians and mathematicians, not programmers and coders, right? So, we have been transforming ourselves as well, culturally and from a skill point of view. That takes its own time. We have a learning curve at our end to build the right skill set within us.”
Iyengar is supplementing his team’s skill set with help from enterprise AI specialist Findability Sciences. The company’s Findability.ai platform combines machine learning, computer vision, and natural language processing (NLP) to aid customers in their AI journey.
“I have a lot of traditional ETL skills in my team,” he says. “What I don’t have is the ML/AI skill set right now. Partners are helping us in that space.”
Ultimately, Iyengar says, these changes will transform how the Data & Tech team interfaces with the business. For now, it operates under a centralized “hub and spokes” model. But he says hiring statisticians and mathematicians in his team won’t be scalable. Instead, what he really wants within three to five years is to embed them in teams closer to the lines of business, so the businesses can run models by themselves.
“Right now, we’re driving the bus at 100 miles and hour and changing the tires at the same time, which is not going to be scalable by any means, though I’m proud of my team that we are doing it,” he says.
Artificial Intelligence, Data Management, Predictive Analytics