Tatau — accelerating AI adoption in the Utilities sector
As part of our ‘Future of’ series, Tatau takes a closer look at how the use of AI technology can transform the Utilities sector.
The Utilities sector has, in recent years, seen the roll-out of AI in customer service with the use of chatbots, automating customers’ calls and routing customers to the appropriate agents. But there is a vast pool of emerging use cases against which to apply AI.
Removing the Adoption Blockers
Even though AI technology is becoming more accessible, still only ~5% of utilities have a clear AI strategy and implementation roadmap in place, leaving the rest grappling with how to embrace this technology, according to a recent Roland Berger report.
In this report, most respondents believe AI will improve efficiency but have yet to take action — with at least 32% having no AI programs underway. According to my old colleagues at McKinsey, this could be as a result of many organisations still lacking the foundational practices to create value from AI at scale, but also facing the challenge of finding skilled people and partners to implement it effectively.
Additionally, risk appetite is a major inhibitor to AI adoption, specifically relating to technical complexity and fear of a longer than expected implementation period, with 39% of Roland Berger respondents citing this, and a further 20% citing lack of suitable of AI solutions.
In contrast, Digitalist Magazine report that “massive computational power is now available at low cost and can be provisioned in the cloud very quickly. Improvements in GPU design have increased the training speed of deep learning algorithms.”
Making the technology accessible is Tatau, which seeks to bring computing power to the masses through its innovative network of distributed GPUs. Tatau sees a future where access to GPU power is no longer a blocker to the volume of use cases that exist in these sectors.
Embracing AI Technology
In their 2018 report, Evolution to an Artificial Intelligence Enabled Network , ATIS distinguish between two focus areas for the implementation of AI technology, specifically focussed on efficiency, performance, availability, reliability and cost outcomes (management and operations use cases), and revenue-impacting uses. We talk through two less obvious use cases of predictive maintenance and load forecasting.
Predictive maintenance of infrastructure
The utilities sector relies heavily on infrastructure to underpin their service offering to their consumers. Keeping this infrastructure operating at peak is crucial to remain competitive and attractive to their consumer base, whilst minimising downtime.
Predictive maintenance enables utility providers to predict the future failure point of an infrastructure component so that the component can be replaced before it fails, minimising equipment downtime. The enabler to this is AI as predictive maintenance requires extensive power to crunch incoming data.
The predictive maintenance approach combines preventive maintenance and corrective maintenance models, measuring historical and real-time data from the network components6, and enables providers to maximise the useful life of their equipment while avoiding unplanned downtime, minimising planned downtime, and saving costs.
Innovationmatrix.com articulates 3 steps to implementing an effective predictive maintenance model using AI6, which are to:
- collect all historical and real-time data from operations and maintenance services and processes;
- transform that data in order to understand important behaviours; and
- build an advanced analytical engine uses statistical models and that can do all the deduction and identification of discriminatory and probabilistic variables.
In New Zealand, distribution visionary Vector is already doing this through its partnership with VIA. The project aims to help electricity distribution companies take a more proactive approach to transformer management by using machine learning to predict faults before they happen.
All of these projects will require extensive compute processing power — and this is where Tatau can provide support. The new Tatau platform opens up the GPU capacity which, up to this point, has been a major inhibitor in AI adoption.
Effective load forecasting faces a range of changeable parameters based on seasonality, weather conditions, geographic differences, and varying usage behaviour between consumers, amongst others. Historically, forecasting has been manual in nature, which isn’t sustainable due to the increasing number of variables and complexity of the forecasting.
Leveraging AI technology for load forecasting enables numerous complex factors affecting electricity demand to be worked into the forecasting models, and allows energy providers to predict required electrical power and manage energy supply to meet the short, medium or long term customer demand. This helps energy providers to plan and make economically viable decisions with regards to future generation and transmission investments, minimising the transmission and distribution infrastructures as well as the associated losses.
Used alongside predictive maintenance, load forecasting helps in determining maintenance of the power systems. By understanding the demand profile, the ideal time to carry out maintenance can be plan with minimal disruption to consumers.
Fortune favours the brave
As the Utitlities sector moves from highly-regulated and low-risk environments to more sophisticated and competitive landscapes, embracing AI and new business models will set the stage for a more technically-savvy environmentally-conscious consumer market.
Roland Berger stresses the importance of implementing quick-win AI projects underpinned by a clear technology and product strategy. This is mirrored by McKinsey research. Both sectors appear to take more of a follower approach rather than leader, with 60% of Roland Berger respondents falling in the follower camp.
With the Tatau platform in place, this provides utilities providers immediate access to affordable compute power, effectively enabling the brave the ability to leap frog competitors. Tatau removes the barriers to entry, and as such, there’s never been a better time to embrace AI technology. Tatau provides a credible and reliable source of computing power to experiment, pivot, embrace — and lead.
Gone are the days where only the big players get to dabble in emerging technologies — those who had vast budgets to spend on R&D. Now with Tatau opening this market and effectively bypassing the more traditional computational powerhouses, the field is open to all to explore, experiment and focus on benefit realisation rather than spending time on trying to navigate how to bring their use cases to fruition.
Get in touch for more information on how Tatau can bring the power of vast GPU capacity to your fingertips.