KMWorld 2024 Is Nov. 18-21 in Washington, DC. Register now for Super Early Bird Savings!

Achieving a unified view for business without coding

Article Featured Image

Implementing a unified view of relevant data across departments for any singular use case has never been as easy—or as necessary as it is today. For example, sales teams can benefit from previous marketing contacts or historical pricing information an organization already has on its prospects. Machine learning deployments are most accurate (and useful) when models are trained on data across business units. Yet, with the best of breed applications getting more popular than ever, unifying these various point applications across domains—including their diverse data models, schema, and terminology—for a unified view has become too resource-intensive to be practical.

This means that IT is constantly reconfiguring the underlying schema for these models, recoding the business logic, and encountering costly delays whenever requirements or data sources change. Agility is compromised and cost overruns frequently occur. Sales and medical affairs teams in the pharmaceutical industry, for example, are either left waiting for the right reports or are far less effective interacting with providers and patients than they should be.

These temporal and resource-draining engineering efforts are eliminated with a purpose built Common Data Model (CDM) across the enterprise that automatically maps source data to it with AI. Business teams can define their own logic on top of the CDM in a declarative, no code manner. With the ability for respective departments to extend that model and its concepts for their individual use while still granting a unified view across the organization, this modern method heightens operational efficiency while consistently producing desired business outcomes.

Codeless configurations and creating a common data model

Implicit to the benefits of this codeless approach to creating a shared view of data throughout the enterprise is the removal of data pipelines, ETL or ELT, and manual scripting that require data to be orchestrated for a common use case. In pharma, for instance, medical affairs teams often talk to physicians and providers that are often treating patients that their drugs may help. Equipping those teams with information about the insights gleaned from past conversations and detailed surveys enables them to focus on the most important information that is needed to get the best results. Unfortunately, and in most cases, these teams are usually juggling between Medical CRM, MIRF and survey tools, which traditionally creates the need to code pipelines for integrating them.

By capitalizing on a common data model approach, these teams can still use their respective tools while easily leveraging an automated system to map the data from all the tools into the common data model. The aforementioned AI capabilities further accelerate the mapping of this source data to the model so they can also tailor that model to meet their own requirements without writing code. For example, sales teams can prioritize top line figures or total revenues, while finance departments can scrutinize margins in different ways. They each can do so by extending the model for their own metrics, data sources, and users—without code while still supporting a unified view across the organization.

The governance framework supporting the common data model enables it to be customized by departments, while extended throughout the enterprise. Semantic models evolve to include new data sources and requirements in their schema. Other methods require significant recalibration time to accommodate new business requirements, but the knowledge graphs underpinning the automation allow the AI to quickly adapt to new sources without much time and effort. This is pivotal for unifying concepts coming from different departments into a simple unified model

Most importantly, the extensibility of these models allows different users to leverage the same concepts or terminology while building on them, in layers, for their own needs. Thus, they don’t have to recreate this business logic from scratch, they dispel the silo culture existent between business units using different tools and do so without code. This layered approach to business logic lets people leverage each other’s work while still allowing each team to create a deep data model for their specific vertical or use case.

Benefits to establishing declarative business logic

The chief advantage of this common data model method for business users is the codeless means by which they can define their own business logic. Instead of coding logic, they simply declare it in a more natural language while referring to concepts that are understood throughout the organization courtesy of the model. The crux to creating a unified view of data throughout organizations is the ability to encode domain logic in a common language instead of in the language each of the sources demand. This allows automatic mapping of disparate source data to the model and the associated logic with minimal effort. The source mapping is effectively abstracted away from the business rules, which run atop the model.

Another benefit is users no longer have to write rules by directly referencing the source data. When sources change or evolve, business rules don’t have to change or be rewritten—which is typically the case when using other approaches, or when this logic is written according to specific tools. This long-standing problem is alleviated by a common data model and enables different departments to easily tailor both the model and the business rules to their own needs without having to manage source systems. It also allows business teams to self-service their needs with less dependency on the teams that are implementing their CRM, ERP, and other source systems. 

The codeless capabilities at the core of the common data model approach are equally important and notable for many reasons. They enable organizations to abandon code-intensive, time consuming data pipelines and data orchestration efforts. They also let business teams self-service the declaration of business logic and make the information shared in that common data model actionable for any variety of use cases.

Still, a unified data model’s greater worth to the enterprise is rooted in the sharing of knowledge it enables, which empowers professionals with the necessary data to optimize their productivity. As is common in most industries, all data—or even the most relevant data—for achieving organizational objectives isn’t always located in the same domain or tool. A unified view of this data is necessary for the business teams to get real insights. 

KMWorld Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues