Towards scaling AI across the organization - How EnBW is driving AI in its business units
“appliedAI helps us to develop a structured approach to scaling our AI activities and to exchange ideas as well as challenges in a professional community.“
- Rainer Hoffmann, Lead AI at EnBW
With over 23,000 employees, EnBW Energie Baden-Württemberg AG is one of the largest energy companies in Germany and Europe, supplying around 5.5 million customers with electricity, gas and water as well as providing energy solutions and energy-related services. Since 2012, EnBW has reacted to the fundamental changes of the German "Energiewende" with a far-reaching corporate transformation and aims to play a leading role in reshaping the energy sector. EnBW’s strategy focuses on the massive expansion of renewable energies and grids, accompanied by the development of sustainable mobility and intelligent infrastructure solutions. In order to utilize the value in its data, EnBW decided to use AI. EnBW is actively working on implementing AI in its product portfolio as well as processes and the companies’ effort is already bearing fruit: In August of 2020 EnBW was announced as the ‘AI Champion in Baden-Württemberg" and won the state-wide competition of the Ministry of Economics, Labour and Housing Baden-Württemberg with its forecasting service for virtual power plants and the scaling of this technology.
But how does EnBW manage to implement AI into the organization, what barriers did they have to overcome and what role has appliedAI played within their transformation since they joined our partner network in early 2019?
AI Journey Timeline EnBW
1) Initial Situation and Problem Statement
EnBW launched its first AI initiatives and started initial experiments many years ago. The first projects were in the area of forecasting load, wind, and photovoltaics feed-in. A central AI team was set up in 2018. Many decentral teams have built up AI capabilities thereafter. Currently, EnBW runs more than 20 AI solutions in all business units and most supporting units. More than 100 employees work in various roles in AI-related projects. EnBW’s goal with regard to AI is to reach a higher maturity in order to utilize the power of AI as much as possible to enrich the jobs of its employees and improve decision making.
Looking at the AI maturity framework of appliedAI, EnBW is a good example of a company in the ‘practitioner stage’. There are already several use cases being applied and even more developed with the challenge of scaling the use case identification even further to align the whole organization on driving the internal AI transformation and making sure these cases actually make it into deployment to realize value. Therefore, EnBW is ramping up community-building efforts, shaping internal AI and machine learning developer communities and supporting AI development with a holistic data strategy and by centrally driving infrastructure development.
Deep Dive: What it means to be a practitioner
Practitioners are taking on the challenge to operationalize, develop and implement their AI strategy, focusing on optimizing processes or re-inventing products. This includes a principled approach to identifying opportunities, setting up central AI-Teams and first corporations, introducing the right training programs for the organization and preparing the AI data strategy. Change management becomes heavily involved in reducing barriers. Technical prototypes are taken from a "proof-of-concept" to a running service and processes are adapted to allow for the new type of development AI systems ask for.
Having successfully developed and deployed several AI solutions, EnBW’s main questions in 2019 were how to scale AI throughout the organization and how to enable as many colleagues as possible at EnBW to identify potential AI use cases. In particular, EnBW was facing the following challenges:
At first, multiple internal teams started building AI-based products, but there has never really been an alignment between the initiatives. Thus, EnBW faced a lack of overarching perspective which would allow them to align activities and enable scaling as well as a standardized deployment. In addition, some business units of EnBW were not aware and active in the field of AI yet, because they did not have an expert for AI who would be able to point out potential use cases and value pools inside the respective business units. Finally, potential use cases lacked visibility so that the AI teams could not effectively provide the necessary support and help drive projects further.
2) Approach and Methodology
To reach their goals and to tackle the challenges, appliedAI and EnBW developed a purposeful approach for AI implementation into the organization.
To start, EnBW conducted several AI workshops in many divisions of the organization based on appliedAI’s methodology to create an aligned AI vision which applies to the whole organization and offers an overarching perspective. During the workshops EnBW also assessed the status of their AI-enabling factors.
At the same time, appliedAI conducted an Engineering Training with EnBW to train their software engineers on how to use artificial intelligence. The training addressed 20 software engineers over the span of one week with on-premise workshops to introduce the latest tools, frameworks and approaches to use machine learning for real-world business problems. Within these workshops the attendees gained hands-on experience in Colab and trained with real datasets. In depth-discussions between EnBW’s software engineering team and machine learning engineers of appliedAI helped to clarify questions and give EnBW’s software engineers the right tools to apply AI independently and support internal teams in defining their AI use cases.
A selection of “AI multiplicators” across various business units was then nominated and was provided with special training on AI and use case identification. Multiplicators were identified and took part in an intensive two-day training with appliedAI in Munich covering an overall introduction into machine learning and its requirements as well as train-the-trainer formats on methods for AI use case ideation and running use case ideation workshops. Afterwards the multiplicators were onboarded into an internal community with the goal to ensure an ongoing community building. By that, the multiplicators help transfer AI knowledge to the teams and support as well as develop central visibility of potential use cases. With that, EnBW can now provide all relevant business units with appointed AI experts, who bring their expertise on implementing AI into their teams.
Employees from EnBW at appliedAI workshop
With all these measures already implemented, EnBW conducted the appliedAI maturity assessment to gain a deeper understanding of how their measures impact the organization already and in how far they bear fruits. The results pointed out good progress towards AI adoption at scale, but also helped to identify focus areas, in which there is still potential for improvement.
Deep Dive: AI Maturity Assessment
The AI Maturity Assessment is an interactive tool developed by appliedAI that allows companies to assess their state of AI adoption and figure out where they are on the journey towards applying AI. This aims to help them on their transformational journey from first touchpoints with AI technology until the broad application of the technology across the business. As applying AI can only be reached in a long-term transformative process, appliedAI has mapped out the different stages that companies go through from first experiments with artificial intelligence until the enterprise-wide application of artificial intelligence at scale. With the AI Maturity Assessment companies can determine the degree of AI maturity they have achieved across eight different strategic dimensions. The tool covers various aspects, e.g. visibility of an overall strategy, availability of capabilities, IT infrastructure, data management and execution. Based on how companies perform across all maturity dimensions, the tool helps them to identify areas for improvement and gives concrete recommendations for action that help to mature and build their roadmap towards AI maturity.
Maturity Assessment Tool by appliedAI
Looking at the analysis of the maturity assessment at EnBW, it was possible to identify clear calls for action and measures. A large number of employees of EnBW have participated in the maturity assessment and a lot of them already have a good understanding of how AI will affect their business and industry.
Beside the described challenges, the assessment revealed a number of AI-related strengths such as EnBW's ability to use a structured approach when deciding which use cases to implement. Further, most employees feel encouraged to come up with AI use cases and also consider AI as one way to solve their business problems. Moreover, EnBW already conducted and pushed forward a good exchange about AI in internal and external communities. The top management is driving the topic actively and provides the company with the required resources to drive forward EnBW’s AI maturity. In order to accelerate the adoption of AI, EnBW ensured ownerships for established AI systems the company developed. This means that every AI system established at EnBW has an owner who is responsible for this topic on a long-term basis and who ensures that the system is regularly improved.
Besides those outstanding achievements EnBW has already identified areas where action is required to reach the next level of AI maturity. In order to continue to increase AI maturity, an even more precisely formulated AI vision in alignment with the overall company strategy is required. Additionally, EnBW still needs to improve the standardization of development processes used across different teams at EnBW to ensure an aligned application and implementation of AI. The same applies to the need to further develop a centralized and scalable infrastructure. On the other side, EnBW needs to increase the overall visibility of data assets at EnBW in order to leverage potential at scale.
Given those results of the assessment analysis, EnBW now has a clear understanding of what actions need to be taken to improve its maturity level even further within next steps in the future.
3) Outcome and next steps
Based on the conducted actions, the first results towards the goals of EnBW to reach level 3 of AI maturity can be recognized and the work is bearing first fruits.
Through the conducted AI vision workshops, EnBW was able to set up a clear and central AI vision across all of its different business units and align it with the overall corporate vision. EnBW is also now using the results of the AI maturity assessment as a basis to further define its next strategic steps towards AI adoption at scale. The developed vision and AI strategy now enable EnBW to ensure an overarching perspective and guidance on how to scale and align their activities.
Moreover, the training of the 30 “multiplicators” across multiple business units and associated companies already made it possible for the central AI team to get closer to the business units and know which AI use cases are currently being discussed and which AI-related needs exist in the organization. This helps the central AI team to extend and improve its service offering and support the multiplicators to bring AI into their business. Multiplicators help drive change activities and create a mindset that is open to AI and advocates that AI has become an integral part of the scaling strategy of AI activities at EnBW, instead of seeing it as a threat. Multiplicators also exchange in communities to facilitate cross-divisional collaboration and learning. EnBW’s employees generally look at AI in a positive way and already ideated AI use cases on their own.
Due to the described initial technical trainings the IT team at EnBW can now support business units with AI use cases, both with the initial evaluation of use cases and with assessing and choosing the right implementation partners for use cases - or even implement simple use cases on their own thanks to the background knowledge about tools and approaches that they have acquired. In addition, the IT team can now communicate more easily with the internal AI teams to jointly drive infrastructure and tool development to facilitate the development and deployment of AI solutions overall.
In the near future, EnBW will face many exciting challenges to increase its AI maturity even more. For example, increasing the speed of AI development and deployment (in collaboration with IT) will become crucial. Further, enhancing its ecosystem by finding the right partners will be another task. Even though AI is already present in EnBW’s company culture, data literacy needs to be driven even further in order to let as many employees as possible be part of the community that drives the adoption of AI. Additionally, EnBW aims to support the development of AI by implementing a data governance framework and a data platform. These are just some exemplary tasks for the near future which illustrate that even though EnBW has already made significant progress, there is still a long journey ahead.