From AI strategy to implementation, Endila helps energy industry clients transition to being data driven organisations.
AI STRATEGY
Every successful AI implementation starts with a clear strategy. Our AI strategy development services focus on:
-
Determining which specific technical, engineering or operational challenges should be prioritised for AI projects
-
In a portfolio context, identifying the most valuable AI use cases and aligning them with an overall digital strategy
-
Establishing governance frameworks with clearly defined roles, responsibilities and communication structures to ensure AI projects can succeed and grow
-
Defining the critical pathway to AI implementation while addressing cross-cutting issues such as infrastructure, competencies and organisation to streamline business transition towards a data-centric company
Sharing a clear vision helps organisations understand the urgency of change and rallies everyone around a common goal. This is particularly critical with AI development, as benefits can be difficult to quantify at project outset and implementation requires cross-cutting change.
​
We work with busines leaders to develop a vision which contributes to overall AI strategy in the following areas:
-
Articulating the benefits that AI technology will bring to their organisation
-
Defining the scope, effort and change their organisation will need to commit to
-
Develop tangible and data-driven roadmaps for AI development and deployment
Establish a compelling vision
AI maturity hinges on five critical elements – strategy, data, people, technology and processes. We help our clients to understand, for each element, where they are on the AI readiness spectrum, what steps other companies at your readiness level have taken and recommendations for moving forward.
Understand your readiness
A data driven business is supported by many critical enablers. This may include data governance, data partnering and data ecosystems. To increase the performance of these enablers we can assist you with:
-
Creating options to accelerate data acquisition within your organisation and create collaborations to access external data from suppliers or peers
-
Establishing the appropriate governance structure to ensure AI initiatives are embedded across business units and functions
-
Defining the right organisational structure for your data talent and critical internal capabilities
-
Designing and implementing AI training programs for non-data experts within your company
​Our experienced consultants will ensure the implementation stays on track and our data experts will throughout project execution.
Address critical enablers
Decide your use cases
AI use cases sit at the heart of any AI strategy. Every company has its own AI use cases and development priorities. Our industry understanding put us in an unique position to define, assess and prioritise our clients’ operational challenges and to develop compelling and achievable AI use cases.
​
For our clients in the E&P industry, Endila’s proprietary database AIForesight holds over 700 use cases and provides a natural starting point. We believe AI use cases always need to be assessed and prioritised under a portfolio context. An initial focus on low hanging fruit provides a tangible and improved understanding of how AI can deliver value – crucial to gaining wider organisational support for more challenging initiatives.
Decide the right sourcing route
Not every AI solution has to be developed in-house. We advise our clients on make, buy or collaborate based on our clients’ business strategy using our AI supply chain understanding, sector insights and data science knowledge.
PROOF OF CONCEPT
The journey to becoming a data-centric company should always feature proof of concept projects. These projects have strategic alignment, a strong business case and a clear value proposition. Developing a proof of concept is a step-by-step process which has data and client involvement at its core.
-
Identify quantifiable technical, engineering or operational challenges which have sufficient data for AI
-
Explore the data and existing data management process to understand which technical approaches fit best the problem at hand
-
Prepare the data for modelling
-
Develop and test different AI models and workflows
-
Evaluate and refine towards the best possible solution
01
Business challenge definition
Business challenge definition
Our PoC projects start with identifying and prioritising operational challenges that align with business and digital strategies. This leads to a clear definition of the project objective, which can be further refined by researching similar applications in other industries and by incorporating lessons-learnt in similar projects.
01
Data analysis and modelling
We develop, select and apply the most appropriate modelling techniques to the prepared and transformed data. We use in-house, open-source and 3rd party libraries and tools for this process, ensuring our clients benefit from the best available technology for each phase of the process. The modelling process is transparent and verifiable by client QA/QC teams to provide maximum confidence in the models developed.
04
04
Data analysis and modelling
02
Data acquisition and understanding
03
Data preparation and transformation
Data acquisition and understanding
We securely acquire all available and relevant data. Depending on the data governance requirements, project data can be stored on client infrastructure, in the cloud or in Endila’s secure data centre. The data is explored with statistical techniques to understand data volume, variety, velocity and veracity. This provides clear indicators as to which data science techniques are likely to provide maximum value.
02
Data preparation and transformation
Our data scientists and subject matter experts cleanse the data to remove inconsistencies, manage missing values, remove duplicates and correct attributes. This may include the use of 3rd party tools and services to tag, label and classify data in the most efficient way possible. Our platform-agnostic approach provides a seamless route to the most appropriate workflow for what can be a resource-intensive process.
Data relationships are visualised and hypotheses on PoC outputs are developed, working closely with client operational and project stakeholders.
03
Model evaluation and performance validation
Once the solutions have been developed, they are evaluated and tested. Comparative tests against empirical or legacy physics-based models are performed to statistically verify that the PoC results are correct, safe and any limitations are clearly captured.
05
Result communication and model deployment
The final project stage helps the client project team to create a communications plan to all relevant stakeholders explaining how business value has been created, and to define how the AI solution can be operationalised through further development. In some cases, the PoC will provide immediate value, requiring minor changes prior to its implementation. In other cases, further development will be required to produce plug-ins to existing systems or a standalone interface for deployment to field teams, plant management systems and decision-support systems.