Join the dots!

Knowledge is inferred using semantic & application metadata, structured by an ontology. Knowledge graphing applies probabilistic, automated reasoning to determine meaning from semantically-described content, through its context & relationship predicates as defined by a domain ontology, linking object & subject.

The ontology is an evolving neural web: mapping & ranking the inferred perspectives to cohere the organisational knowledge the querent is looking for. The growth mindset, which pushes the frontiers of human achievement, is fuelled by such knowledge: where human imagination & capability are augmented & automated by computer power, speed & logic.


Data, metadata & semantics foundations!

Data is the foundation of knowledge. Ensure your data quality is up to the mark, both structured (numeric) & unstructured (documents, free text, scans, rich media).

If you haven’t already achieved this, then take advantage of our data audit, data discovery & data quality services.

Second, does your data have gaps? Is the data is valid & canonical? Is any aggregation or consolidation needed? Again, if not, benefit from our experience in data transformation, data validation, & general data engineering services.
For the same reason, your application-generated (or object) metadata has to be pervasive, consistent, relevant & disciplined. Missing, opaque or disorderly metadata will impede your search for knowledge and also any AI / ML initiatives you may have planned. If your applications don’t ‘share’ their metadata, this too will impede your pursuit of knowledge.

To help you, we offer metadata audit, metadata discovery, & the ability to significantly reduce the cost & effort of extracting metadata from opaque packaged applications (ERP & CRM).

Remedying metadata gaps still requires manual effort. Missing metadata will upset any Azure & Sharepoint deployments so its better to get into good habits early to benefit your overall digital capability.

Metadata in dark data, archives or legacy systems also requires consideration because it may still, even if unused, be within scope of data privacy & protection regulation & therefore an important part of your corporate knowledge- base.

Taxonomies also provide important metadata, but these are hierarchical, non-semantic & 2D, lacking the multi-dimensionality of perception (meaning, relationship & context).
Structured semantics are how we make sense of ‘The Semantic Web‘ – logically, interconnected data, which humans & machines understand.

So your organisation will need semantic metadata as well as the existing application descriptors (metadata).

We will help you source & organise the terms you actually use to refer to your organisational data & knowledge; to provide domain consistency & coherence.

The components of inferred, automated, reasoning i.e. knowledge system.

We will help you choose which GraphDB to deploy, on premise or in the cloud, to act as the repository of your knowledge schema: the ontology.

The choice of graph database depends on the individual use case plus other considerations such as memory efficiency, scale, end user devices & complexity.

So choosing the right GraphDB is best done with the help of experts.
Your shiny new repository now needs to be populated with the ‘dots’ (the subject & object nodes) & the ‘joins’ (predicate relationships known as edges).

The nodes are the key entities (objects & subjects – tuples or triples) of your information needs & the probabilistic comparative ranking (matrices or lists) per attributes based on relevancy.

As with BI, discovery interviews by an independent consultant can help to achieve consensus, embrace change & ensure the right approach is taken.
By ‘meaningful domain‘ we mean an organisation where, different perspectives on entities & concepts have been consensually resolved into singularities: into a single ontology where human and machine logic is consistent across the ecosystem.

This involves surfacing & modelling information priorities; organisational semantics; the nature of those relationships & probable vocabulary (semantic tags).

The act of ‘joining the dots’ brings implied context & salience. As subjects & objects develop multiple relationships, the magic lies in graphing those as matrices, ranking the relationships as they pertain to search parameters.
The relevance of items in the schema now needs weighting.

Those ‘dots’ & their ‘joins’ indicate the indicate relationships & their degree of salience, just as they do in a search engine.

What you want your systems to do is enable meaningful, automated reasoning, so that searches provide the most relevant results (without significant omissions).

From a systems point of view, that reasoning has to be
machine-interpretable & probabilistic so as to provide inferred knowledge i.e. the whole being greater than the sum of the parts (which is currently how things are).

The database type which enables this is the knowledge graph.

Choose the optimal realisation path towards a future you can’t yet see.

A SaaS approach gives you a fast route forward, effectively subscribing to existing ontologies that are an industry fit, thus benefitting from the investment in time & effort already made by others.

This will also save you having to implement & fill your own Knowledge Graph database if you are not ready for this.
Starter kit
The starter kit approach also means you benefit from the work of others, but this time in terms of content.

If you have a Knowledge Graph project already underway, why not save a lot of time & effort by using triples that others have already written. No need to reinvent the wheel for standard items right?
Strategic planning
For those with more specialised, complex, or simply larger needs, the third option is a full service consultancy offering based on data-centric principles offering strategic planning & delivery: a technique pioneered by Pierre Wack (the man who saw the future) which proved so successful for Shell’s decision-making & portent analysis.

Here you benefit from help with:
the knowledge graph (deployable on cloud or premise);
pre-populated content (industry building blocks relating to internal & external forces of strategic impact);
to this, for instance, can be added typical internal KPI data harvested from your existing systems to form the basis of comparative analysis against external measures.
At times of exceptional business uncertainty, the attribute that Wack foresaw as the differentiator between success & failure, was the ability to look beyond the existing knowledge horizon: to see the future.

With this strategic planning offering, companies, like Odin’s ravens, will be well placed to gain insight from both historical performance (the corporate memory) & having the capacity to think and analyse the real-time world around them, to see the future.