Obligatory Disclaimer: This my opinion. i.e. it is not the
official Duck Creek company line. It is all my opinion which based on my
experience from my solid history with the product I won’t speak to futures of
the product or timelines for futures. There is a fantastic Duck Creek team that
is better suited to speak to product direction, etc.
What Insights is
Technically it is a robust set of databases or data stores and
code developed to fill those data stores from production (upstream) systems. Additionally,
there are starter reports, dashboards, etc. in Duck Creek Content Exchange. The
technology is built upon the Microsoft SQL Server suite of products. This robust
and mature database technology from Microsoft means that each client can relatively
easily develop dashboards, reports, pivot reports, feeds, etc.
Insights is designed for the insurance industry by insurance
and data professionals with decades of experience in that space. The data in
each of these databases is structured somewhat differently. From any client’s
source data all the way to the end of the line there are transformations that
take place (computer coding). The intent of the transformations is important to
understand as we get further into Implementing Insights, especially when it
comes to testing.
Insurance Data Model
Most subject areas that insurance professionals need are broadly
represented. It is a framework that enables each Company and each business unit
(BU) to store their own business terms, formulae, etc. Note that when I say
most I am really hedging. I don’t recall specific items that didn’t fit within
the existing schema. The ones that didn’t were very client specific. You may
have legitimate needs that aren’t met but rest assured you can add it without
bringing yourself out of the upgrade path. More on that later.
Insights also provides you with the relationships and depth
of data that you need for regulatory and strategic decision making needs. The
relationships between policies, risks, people, organizations, etc. etc. are all
part of the solution giving you the power to get information on your data in
the way you need it.
Configurable & Extensible
The solution is designed so that you can configure and even extend
the model and easily remain within the upgrade path. This is important for any
implementation but especially important if you are implementing in the cloud
via Duck Creek on Demand.
Configurations are largely via name and value pairs. For
things like risks, risk factors, Bus, insured objects, etc. you use your terms
to customize all those relationships. They are stored in the database initially
and then you can easily add additional pairs as your business changes.
For that special item that isn’t covered by any of the
above, Duck Creek has designed the solution with a way to add entirely new tables
and columns that keep you in the upgrad path. There is a reserved custom. Schema
that will never be used by Duck Creek and will never be touched by any
upgrades.
What Insights is not
Insights is not designed as a “real time” solution. It is
designed as a strategic solution. Depending on a number of factors a client
could set up their Insights jobs to run somewhat frequently and become closer
to real time but it isn’t what the current design is. Designing for near real
time should be thoroughly planned with an application architect and a data
architect who are knowledgeable and experienced with Insights.
Insights is not a “big data” store in the sense
that it isn’t designed as a data store for unstructured data. You can store
blobs of data but the overall recommendation is to link to other big data
stores rather than try to replicate and maintain them.
Insights is not a “black box” product. There is a
long story on this that I won’t get into but the short story is one would be
delivery lead for an implementer didn’t want to follow best practices for
testing BI projects and insisted that they couldn’t possibly do so. They were
and are still wrong in insisting Insights is a “black box.” The data,
how to interpret it and how to use it are all well documented. A qualified BI
team can test it without digging into ETL code. If your systems integration partner
is taking a stance that Insights is a black box my personal recommendation is
to explore other options for implementing.
That’s it for now for my idea of a baseline. Feel free to
comment in my blog or contact me if you have questions, comments or
recommendations.
Future Broad Topics
- Implementation Methodology
- Testing
- Single Source of the Truth
- Sources
- Outputs
- What would you like to see covered?
Test words, etc.