Shift to new digital capabilities by adopting a vertical data platform

Organizations using data and artificial intelligence (AI) are outpacing competitors, turning information into a key driver of performance and innovation. They use an average of 400 data sources, with 20% using over 1,000. This growing variety of sources, formats and integrations requires businesses to rethink data management and value extraction.

Data products—powered by vertical data platforms—help overcome these challenges, unlocking new opportunities and enabling data-driven business models.

A data product is a single source of truth for a particular dataset, packaged with insights and functions to deliver clear, measurable value to its consumer. Data products combine domain-specific expertise with trustworthy, accurate data and follow a formal lifecycle.

Data products are multimodal and accessed over a data mesh. This enables applications to consume data directly through structured query language (SQL) queries, application programming interfaces (APIs) or streaming endpoints. AI applications can connect to a vectorized store for retrieval augmented generation (RAG) processing against a foundation model.

With a structured usage contract defining available insights and actions, data products help ensure consistent service quality. Created and published with formal metadata specifications, they accelerate data-driven outcomes for internal use or external monetization.

Data product-led organizations achieve transformative benefits, often combining cost and risk reduction with new revenue or blue-ocean ventures.

Enhanced time to value
 

Decentralized development empowers domain experts to build needed data products, eliminating bottlenecks with centralized data engineering.

Increased revenue opportunities
 

Monetizing data products or augmenting services with data and AI can significantly increase revenue. For example, a US bank generated USD 60 million in incremental revenue and cut USD 40 million in annual losses from a single product.

Cost and risk reduction
 

Treating data as a product can reduce ownership costs (technology, development, maintenance) by up to 30% and streamline governance, enhancing compliance and risk management.

Strong domain-driven expertise in data product creation increases the value of that data’s intelligence, helping to ensure that the right consumers find the right data with measurable value.

Vertical data platforms aim to enable users to work more easily and safely. Today’s digital platforms are preintegrated, offering self-service capabilities for specific domain tasks, such as human resources (HR) or finance. These platforms incorporate domain-specific governance, policies and standards.

For example, internal developer platforms standardize the application development lifecycle for faster delivery. They simplify decisions such as configuration and software selection, and automate building, testing and deploying application code.

A vertical data platform hosts trustworthy, accurate data embodying domain-specific expertise, managing it through a data product lifecycle. It connects to any data source, regardless of processing type or location. Data products are built for consumption as services in multiple ways, and the platform can distribute them to global cloud service providers.

A vertical data platform is a type of industry cloud platform, crucial for successful vertical AI solutions. Vertical AI or industry-specific AI systems, will surpass traditional vertical SaaS offerings including customer relationship management, enterprise resource planning and HR.

It automates costly, repetitive, language-based tasks across industries. Domain-led data products (for example: customer, employee, medical diagnosis, natural disaster data) are key to powering effective, industry-specific AI.

Vertical data platforms enable 4 types of business models: internal data sharing, broker-intermediary, external data monetization and ecosystem creators. Treating data as products introduces a value-focused mindset.

Data is managed like any other product or service. Business lines begin asking, “Do we have a data product for that?” This mindset treats data products like stock keeping units (SKUs), prompting value-based pricing considerations (manufacturing costs, operating costs, revenue, price, profit, and so on).

Internal data sharing

Streamlining internal or cross-entity data access improves cost efficiency and governance, enabling innovations such as augmenting existing services and products with data-driven value.

J.P. Morgan Chase, early adopters of internal data product sharing over a mesh, saw increased reuse, reduced data copying and improved risk posture through domain expert-controlled permissions.

Broker-intermediary

Platforms such as Datarade connect data providers and consumers, facilitating direct financial transactions. They might also provide workspace hosting for data consumption.

Countries with strong sovereign cloud, data and AI positions are seeing new ecosystem and marketplace offerings.

External data monetization

Organizations can monetize data products through developer portals, marketplaces or cloud catalogs. The platform operator owns the financial transaction and can also produce data, curating and assembling data for added value and manages the financial arrangements.

Transport for London exemplifies this, offering real-time and historical travel data, powering 600 apps and contributing an estimated GBP 12–15 million annually to London’s economy.

Ecosystem creators

Multisided platforms enable collaborative ecosystems where partners (providers, producers and consumers) co-create and consume data products, or share them publicly, distributing revenue and benefits. Hugging Face Hub is an example of this, offering an open platform for sharing datasets and machine learning models, fostering collaboration and generating new AI and data offerings.

Shifting to a data-driven business model might involve several approaches and potential data infrastructure modernization.

Data modernization for data product delivery depends on virtualization or source copying for versioning, caching and distribution. Modernization also involves data engineering shifts to implement data meshes, enforce governance and enable decentralized development (for example: “data workspace as a service”). 

The foundation for a vertical data platform is IBM’s data and AI fabric software, which provides core platform engineering capabilities and integrates with the open data lakehouse and AI capabilities of IBM watsonx™.

IBM Data Product Hub enables the data product lifecycle workflow, covering data integration, transformation, machine learning model training, generative AI integration, streaming data pipelines and third-party API brokering.

Running on Red Hat® OpenShift®, the system offers hybrid cloud deployment flexibility. Different components can be federated across on-premises and public cloud deployments, scaling independently while managed as a unified platform.

Many companies are updating their data strategies to support data-driven business models. Data products offer a straightforward evolutionary path to numerous use cases and benefits. Clear business outcome goals guide the technology and operating model choices for implementing a vertical data platform.