SnapLogic in the Data Fabric

SnapLogic’s core capability is codeless app and data integration. It has a visual designer to help design and build data pipelines with over 1,000 pre-built connectors (“Snaps”) and patterns, connecting to and from popular data sources like SAP, Salesforce, ServiceNow, Netsuite, leading databases, and more. A cloud-native application itself, SnapLogic allows integration of data between cloud and on-premises systems, supporting both batch and near real-time streaming. It also orchestrates and automates business processes and workflows. Its visual designer interface is intended to allow end users to build their own data pipelines with minimal assistance from the IT department. While at the same time, providing IT a flexible, visually driven integration productivity enhancer that also can be customised and driven deeper, under the hood, when necessary.

Fig 01 – Modern Enterprise Data Orchestration

In terms of data fabric and data mesh architectures, Snaplogic sees these architecture approaches as overlapping, offering different strengths, but sharing in common a principle for decentralisation. SnapLogic supports these approaches in several ways. With an emphasis on self-service data pipeline and data product creation and management. SnapLogic provides an AI-augmented integration infrastructure that is multi-domain capable, and can be federated. This includes cataloguing and version control of API-driven data products. SnapLogic can ingest, transform and serve up data packages from widely distributed source systems, but federate control and security. The solution has inherent support for project-based data products by business domain, so fitting neatly within the data mesh concept of distributed data ownership, but with federated control and security, simplifying deployment. It can publish API-Led data products through an internal data marketplace or portal. It is not a complete data fabric solution, but then arguably no one vendor provides everything for a data fabric at present. It is certainly the case that no one vendor currently provides a complete data mesh solution.

Customer Quotes

“It takes us minutes to deploy integrations across 1,800 applications with SnapLogic. We have also been able to reduce overhead and support costs by 25%.”
Swati Oza, Director of IT Emerging Technology, Data Integration and Machine Learning at Hewlett Packard Enterprise

“SnapLogic is our secret weapon. We are able to deliver on the company’s digital transformation initiatve by demonstrating that we can deliver and scale without massive product teams.”
Phil Maguire, Application Service Delivery Manager, ITV

Apart from the core integration functionality described, SnapLogic also has substantial capabilities in artificial intelligence. One example of this is that, for some time, it has been shipping AI-driven AutoSuggest (formerly Iris AI), which dramatically simplifies and speeds up integration builds. Another example is with its SnapGPT copilot, a natural language prompt allowing users to build pipelines, generate sample data or analyse and document existing data pipelines. Lastly, SnapLogic’s GenAI Builder tool was released in January 2024 and allows customers to create their own knowledge store in a vector database (currently Pinecone, with OpenSearch as an option soon) and perform retrieval augmented generation (RAG) by easily loading users structured & unstructured data into a vector database. For example, an employee HR handbook may be indexed and then a natural language interface can be used to answer employee questions like “When will I get paid?”. The private vector database avoids the well-known security issues of public AI tools. There is prompt engineering to ensure that answers are only given within the allowed context, so you get far fewer hallucinations than with general-purpose AI tools like ChatGPT. The product supports OpenAI, Azure Open AI, Claude and Titan (on AWS). One early customer is a utility in waste management that has complex billing formulas in its commercial contracts. They are using the AI capability to accurately extract payment formulas from the contracts for comparison with their billing system, highlighting billing discrepancies. SnapLogic enabled this company to find a significant revenue recovery opportunity and SnapLogic’s ease of use allowed the client to complete a proof of concept in less than 3 days.

Fig 02 – Three camps – with similar attributes, creating confusion

The SnapLogic approach allows the creation of custom data integrations to be either accelerated through IT or delegated to and owned by business domain end-users rather than languishing in a work queue within the IT department. In one customer example, Schneider Electric (a $36 billion revenue, French multinational company specialising in digital automation and energy management) – has over a thousand users provisioned on Snaplogic. They have several categories of end users, from advanced ones who need no IT support, to less experienced user categories, and chargeback more to the least experienced users. This gives a direct incentive for users to become more self-reliant and use the tool themselves. Over 30%, and growing, of Schneider Electric integrations have been constructed by citizen integrators.

Fig 03 – More in common than different

The bottom line

SnapLogic has made rapid commercial progress in the data and app-to-app integration market, and has many examples of customers where end users, with some degree of technical savvy, really are building their own data pipelines and API-led data products. This is particularly suited to the data mesh idea of distributed data ownership, and can also be applied to more centralised data fabric architectures. SnapLogic should be carefully considered as one of the main building blocks of a data fabric or mesh solution.