The core of the Syniti Knowledge Platform is its knowledge graph, which contains both business and technical assets. In turn, your graph and its assets are accessed via the platform’s data catalogue. This allows you to view, edit and manage all of your assets, including strategic assets such as initiatives, strategies and goals; governance assets such as rules (including data quality rules), policies and business terms; and, of course, data assets. What’s more, the Syniti Knowledge Platform understands assets within their business context. This consists of both identifying the relationships and connections between your assets and leveraging those connections to create business value, as evident within the product’s DeepGuidance capability, which uses machine learning to automatically and dynamically suggest improvements to your assets as you are viewing them. This could include suggestions for new business rules, acknowledgement and remediation of data quality issues, or detection of suspected – but undocumented – asset relationships, to name only a few examples.
For business terms in particular, natural language processing is employed to detect and highlight words or phrases that may represent additional, as yet undefined business terms. These terms can be referenced within business rules, which are written in natural language using business terminology. A given rule will also contain its written implications (in other words, its business meaning) as metadata, as well as an enforcement profile that shows where and how it is being enforced (for example, a data quality rule might link to its entry in your data quality solution).
Fig 2 - Applying data quality to improve business outcomes
As far as data quality is concerned Syniti’s view is that unless data quality is combined with business impact it is “just providing metrics for IT”. There is more than a grain of truth in this. For example, consider the screenshot in Figure 2, which shows account payable leakage for major customers, vendor payment terms that are non-standard (corporate rules state 90 days) and where payment terms have been overwritten manually. There are many other such use cases. For example, unrealised discounts or rebates due to supplier duplication.
More specifically, as far as data quality is concerned, Syniti provides the sorts of profiling, cleansing and matching capabilities one would expect, though the company is ahead of many of its competitors in implementing machine learning. As mentioned previously, the architecture of Syniti’s data quality offering is currently going through a significant re-development but it is too early for use to evaluate the product on that basis.