Utility Week - authoritative, impartial and essential reading for senior people within utilities, regulators and government
Issue link: https://fhpublishing.uberflip.com/i/1309237
UTILITY WEEK | DECEMBER 2020 | 31 How AI is changing finance and operations in the utilities sector Tech Talk FABRIZIO TOCCHINI, HEAD OF INNOVATION, CCH TAGETIK How is CCH Tagetik helping utilities companies run more efficiently? The concept behind CCH Tagetik was to create a complete CPM platform that's simple and intuitive for finance professionals. And this is what we do. The real game-changer as far as our utilities clients are con- cerned is the way we're connecting all their operational data to feed the financial analysis and regulatory report- ing, through our Analytical Information Hub (AIH). It helps them model their businesses in all directions – not just accounts data but operational KPIs too. It allows you to see all that data in granular detail, so, for exam- ple, you can do an end-to-end analysis of cost unbun- dling across any business combination, in a simple cascade of steps. It's great for running a C1 table. Our clients are asking us more and more for "any- where connectivity" – pulling data from a huge range of sources. This goes beyond typical ETLs. We need less latency, and we need to make sure there's no data redundancy when you ingest all these sources. We've already released APIs for various databases (on top of our full suite of ETL functionality), and there are more coming. I'm particularly excited about the idea of the "virtual dataset" that we have now. It lets you view the data in another dataset and use it to build a report with- out physically transferring it. This avoids data redun- dancy and ensures data security (because you don't have to store it). How is CCH Tagetik using AI to help clients outperform their competition? We thought deeply about this, beyond just applying predictive analytics to time-series data for more accurate forecasting. We said: starting with data, I have a good first version of my forecast, but I need to know why I reached these results. If I know the "why", my scenario planning and what-if analysis will be more valuable, because I know where to act to change things. This explainability is now possible – to explain how a result has been reached. So, for instance, for each business combination (a given product sold in a given channel at a given time), I can understand which variables have the biggest impact, whether that's the promotion or the channel, for example. That means that, for that specific business context, I can act specifically on the key drivers. I can do a conscious scenario analysis. We're moving from what is likely to happen, to why it's likely to happen, to how I can change what's likely to happen. This explainability helps people to trust the data too. Our next release in January is going to be the first with this functionality. We're also looking at how AI can help with more extended modelling. For example, your business is affected by given metrics (regulations, taxes, costs of distribution) but they aren't easy to predict. If you're accurate on these metrics, then you're doing well. But there are exogeneous factors as well, trends in the wider market like employment levels, how behaviour is being affected by the pandemic, the weather, electric vehicle usage and other decarbonisation trends. Today, there are datasets beginning to explain these wider trends in a way that can be analysed. We can take this data into account and include it as a variable in your modelling. It's like "driver hunting" – looking for other drivers in private or public datasets that allow you to be even more accurate in your predictions. Take cashflow as an example. The problem normally is that you can set whichever terms you want for your credit payment but the reality says that things don't always work like that – delays are caused by claims, inefficient processing, etc. This should be taken into account in the DSO/DPO calculations you do up front. AI can help with this, by analysing the trends of past transactions, so you can see what's going on product by product, supplier by supplier, and you can infer the real DPO you can expect from a particular transaction, as well as what kind of remedy you can use to mitigate the risk of delayed payment. On top of this, you can also calculate your provisions for losses more accurately. Are there any other emerging technologies that finance leaders should be keeping an eye on? Blockchain is definitely a very important emerging tech- nology; it could change the way the transaction world works. If there is a chain that connects the sender and the receiver in intercompany matching then there isn't a need for the reconciliation. The effort will be spent on checking the chain itself, not the transaction. A few of our clients are trialling how they could use blockchain technology for a portion of the ledger, and we're looking at how we might need to develop our product to adapt to this, to check the chain not just the transaction. Looking even further ahead, I'm excited to see how quantum computing might help us predict the future price you can apply to an asset even more accurately, taking into account vast quantities of variables and possibilities. That's a problem I would love to solve bril- liantly, but that's working on a longer timeframe. For further information, visit: www.tagetik.com/uk This is an abridged version of Fabrizio Tocchini's Tech Talk. To read the full article, visit: www.utilityweek.co.uk Sponsored content brought to you by "If I know the 'why', my scenario planning and what-if analysis will be more valuable, because I know where to act to change things." CCH Tagetik is a unified CPM platform which Gartner has named a 'Leader' in both Consolidation and FP&A.