Data is gold if you know how to mine it. Find out how companies are leveraging the power of data to the max for greater opportunities. Techsauce Global Podcast talks to Stephanie Sy, CEO and founder of Thinking Machines, a leading data consultancy company that is building AI data platforms across Southeast Asia. Sy shares her experience working in Silicon Valley and her mission of making a positive change in the SEA region with the power of data.
On a soft skill side, among the top and most respected talents that Sy has worked with share one common trait of being pretty much ‘self-taught’ and ‘self-driven’, a trait that she aspires to.
“You could get an undergraduate degree in mathematics, you could get a master's in business, you could get a PhD. But at some point, when you're trying to build something new where nobody has done it, you have to just have the guts and have the drive to keep going”
When operating at the peak of the industry, there’s no path to follow but only to pave forward, so the whole methodology and framework around the industry are set around discovery and improvement, both beginning at self to an organizational level.
On the hard skill side, she learned that data technology has become much more powerful during the last decade but also entirely more complex. It is also especially confusing for talents who looks to start a career in data science, whether they should learn machine learning model, statistic, or data infrastructure. Though with this she advised that 80% of the time simple actionable SQL queries that answer the actionable question will help us more than a complex model.
For Sy, the major challenges are rather mindset issues not the technical ones
Big organizations tend to have a cultural blocker in using data to drive strategy, with their established with executing the course of action based on ‘post-fact’ quarterly report. Meanwhile, the Tech companies have the whole a very different mindset through looking at data as an immediately actionable tool, like looking at the data now to make a decision tomorrow.
These big organizations also want to build innovation teams without an actual goal. In the rush to stay relevant, they tend to invest in something sexy sounding like AI, buying onto the idea of explosive growth, pouring fortunes into new offices and hot talents, through forgetting to give that innovation team actionable goal on what should they harness that data for.
Another challenge is building a data team. Given that one, there’s so much competition for talents and two, one talent cannot do everything independently. Within the category of data science, there’s a whole industry where there’s data engineer, machine learning engineering, data visualization, strategist and so forth. Organizations need to think of it as a full team effort if they want to transform through data.
For Sy, she believed that even the most resistant individual acknowledge that data is important.
Therefore, it makes sense to give the team the training and space to continue to learn and grow, as well as introduce them to concepts like growth mindset vs fixed mindset.
"Helping people first feel comfortable with the idea that they're going to be uncomfortable” –
While it can be difficult and take time, at the end of it, they're going to be better versions of themselves, and they're going to use data and AI to further elevate their creativity and ingenuity.
For corporate, the culture change must start at the top, so company heads must start using data and referencing data when they're making decisions, not treading their opinion into policymaking.
For smaller companies, the great thing about 2022 is that they can buy a lot of great AI capabilities, instead of spending months on building it on your own and even trying low-cost investments. Given that a lot of new startups tend to sell software for “really darn cheap”, as it is practically subsidized by venture capitalists
Sy also advised not to build some massive AI tool that only improves business by half a per cent. But rather think about what's the smallest and most agile thing we could do. So be from improving companies reporting capabilities with software through dashboarding tools, or using Metabase for data ingestion, and ETL tools that are open sources and free.
One of the recent cases Sy has worked with was for a big courier company in the Philippines which have over 1,000 branches across the country. They had an existing but messy data system that was that were trying to match the customers and trying to figure out who their customers were.
With Thinking Machines, Sy has helped develop solutions with AI technology, a single customer view that identifies unique customers across all their stores. And now once they have that they can use the AI model to identify who their unique customers are, then they’re able to analyze the customer's journeys and figure out who is selling, who is selling on Shopee and Lazada, from their house.
With informed data, it is the next step of understanding customer behaviour, keeping people happy and keeping them retained. With a business, we can do that step by step, starting from the single customer view, data warehouse and machine learning modelling onwards, optimizing their overall experience.
Humans are going to evolve, what Sy mentioned is that, to succeed in the data-driven field, we must embrace the mentality of a forever learner.
“Everything that I learned at school, 10 years, 15 years ago, I barely use any of those tools today, maybe the only thing I use is SQL. The fundamental mathematics and statistics have been very helpful to me, as my career has grown. But every year I have to change the tools that I use at my job”
She also added that what she’s seeing now is a very cool process where humans are changing behaviour in response to the availability of AI. With the case three years ago when Google wrote GO ( 围棋) -Playing AI that beat some of the best world’s players. At first, this shook the whole GO community. But what happened next is that these top players started studying the way AI wins in, and adapting that to their winning strategies
So, the game where the winning strategies had not changed for maybe 50 to 100 years, suddenly has seen a huge explosion in creativity and dynamism as humans saw that the AI could win using these totally new techniques. Now people Play GO in a different style.
During the last bit of the sessions, Sy has left us with her views on the future of data in Southeast Asia, seeing both opportunities and challenges.
She sees that there will be a bigger pool of experienced talent to draw from, as more Southeast Asian Tech companies get funded, grows, and exited. She also hopes that in the new talents that will lead the coordination of building machine learning machine models, pumping out high quality information and curate data for machine. She also hopes as well as these talents will lead data initiatives of all kinds in the Southeast Asia region.
Currently, she is also working with UNICEF innovation Fund and governments to build “AI for D” or the AI for Development. As a project with goals to leverage the countries fragmented database on wealth, poverty, and climate through creating the pipeline for people to embrace the mentality for open sources data for self-use, share and contribute back to grow together as data-driven citizens.