Digital Perseverance: Data-Driven Innovation, per Redis Executive
In a session at the Data Science Innovators Conference 2023, Taimur Rashid of Redis argued that leaders should accelerate digital innovation while building systems that outlast them. He covered vector databases, customer examples from Capital One and Dick's Sporting Goods, and security, sustainability and inclusion.
By The AI Enterprise Magazine Team4 min read
Key takeaways
- Start with the customer. Capital One began its payment system from the desired real-time experience, then chose technology.
- Applications drive data architecture. LLM applications bring vector embeddings and new retrieval needs.
- Latency is the new outage. Rashid said slow responses are experienced by customers as failures.
- Upskill around data. Analyst, engineer, scientist and prompt engineer roles all need data literacy.
- Build to last. Digital perseverance means security, privacy, sustainability and inclusion alongside speed.
In a session at the Data Science Innovators Conference 2023, Taimur Rashid, chief business development officer at Redis, argued that companies should accelerate digital innovation while also practicing what he called "digital perseverance": building systems, processes and organizations that outlast the people who create them. His 34-minute talk, "The Pivotal Role of Data in Accelerating Digital Innovation & Ensuring Digital Perseverance," drew on customer examples from financial services and retail.
Why does change keep accelerating for data and technology leaders?
Rashid said outside forces such as the pandemic and supply chain shortages test business leaders, but change predates them. Companies moved from waterfall to agile development and DevOps, and absorbed cloud and mobile along the way. What matters, he said, is the nimbleness and open-mindedness to respond.
He pointed to ChatGPT as the current example. He said it reached more than 100 million monthly users in two months, and that adoption cycles are shrinking compared with apps such as TikTok and Instagram. In his framing, the answer is two guiding principles: accelerate digital innovation, and pair it with responsibility.
How do applications shape data architecture?
Rashid said applications and the experiences customers expect dictate the data technology underneath. He traced an evolution from relational and NoSQL systems to data powering internet applications, machine learning and the cloud, and asked whether current data technologies suit the needs of the digital age.
Large language models, he said, change how data is represented. Vector embeddings are numerical representations of structured and unstructured data, and organizations now need stores that optimize their storage, indexing and retrieval. He said Redis was working with pharmaceutical and retail customers on architectures that connect vector databases to large language model APIs such as OpenAI and Cohere.
What do customer examples show about reinventing a business?
Rashid described Capital One, a customer, as taking a first principles approach to a next-generation payment system. Rather than starting with technology, the team started with the customer experience, including real-time mobile banking, and worked backward to the components needed.
He cited Goldman Sachs, an investor in Redis, as an example of reimagining identity: the firm sees itself as a technology company, and he said one in four employees is an engineer. He also described Dick's Sporting Goods, which he said enabled contactless pickup within 24 hours of the pandemic's start and shipped a second version 18 hours later, using Redis for agility.
What do customers expect from digital experiences?
"Latency is the new outage," Rashid said, describing a Redis mantra. Customers expect real-time and personalized experiences, and a slow response counts as a failure from their point of view. He cited Instacart's item availability as a case where latency is experienced as an outage.
That expectation changes jobs. Rashid focused on the data analyst, data engineer and data scientist as roles that treat data as a product, and he said generative AI is creating roles such as prompt engineer. He urged leaders to keep data literacy at the core of upskilling and to be specific about what each role does.
What is digital perseverance?
Rashid defined digital perseverance as being responsible and thoughtful so that what leaders build can "outlast us." He named three pillars. The first is cybersecurity and data privacy, since more digital applications widen the threat surface. The second is sustainability, because generative AI and cloud usage carry financial costs as well as energy and carbon footprint costs. The third is inclusion, in how teams are built and how diverse customer needs are met.
What did Rashid say in the Q&A?
Asked to define the data analyst role, Rashid said it centers on making data visible across the organization through business intelligence, and that it is moving from building reports toward answering the most important business questions in an agile way. He said that in his own organization perhaps 10 of the many reports and dashboards actually get used.
On SQL versus NoSQL, he said relational systems powered predictable front-office and back-office applications, while NoSQL emerged for high-throughput uses such as social and gaming applications. For finding solutions from first principles without an obvious tool, he recommended starting with the customer, learning what open source and commercial tools can do, and using the "five whys."
Where to hear more from data and AI leaders
AI Connect continues to host practitioners from the data and AI world. The 2027 schedule includes 24 free live webinars in four series: the AI x Enterprise Panel Series, AI Demo Days, Agentic AI Showcases and the Plug & Play Series. Details are at /webinars/, and registration for 2027 events is coming soon.
Frequently asked questions
What is digital perseverance?
Rashid defined it as being responsible and thoughtful in accelerating digital innovation so that systems, organizations and processes can outlast their builders.
Why do large language models need vector databases?
He said LLM applications represent data as vector embeddings, so organizations need stores that optimize storage, indexing and retrieval of them.
What does 'latency is the new outage' mean?
It is a Redis mantra that when an application does not respond fast enough, the customer experiences that as an outage.
What is the difference between SQL and NoSQL?
Rashid said SQL suits predictable transactional front-office and back-office applications, while NoSQL emerged for high-throughput applications such as social and gaming.
This article is based on remarks by Taimur Rashid, Chief Business Development Officer, Redis at Data Science Innovators Conference 2023 on April 12, 2023. More coverage: More AI Connect Events.
Filed under: AI Strategy & ROI, Data & AI Infrastructure, People, Culture & Leadership
