Reporting, Automatisation, Supply chain management helper, data integration. All of those of AI use cases, also when applied to the sustainability /ESG field, eventually lead to the ‘so what?’ question.
‘So what’ in financial terms that is.
This is the reason to dedicated this last of our instalments of this blog series to investigate how AI could help us bridge that gap between the vast amounts of quantitative and qualitative sustainability data we already do own – or will will going forward – , and our understanding what those mean in terms of balance sheet impacts.
Sustainability is generally perceived as a cost. The primary reason for that is actually rather simple: it implies having to change the ‘modus operandi’ of a company. And change does imply cost – whether or not said change is connected to anything sustainability/ESG is actually irrelevant. However, due to the (mostly, not always) absence of benefit related calculations, the only cost ever in everyone’s mind.
This is where the difference lies to other major themes that imply change to processes, management, products:, such as precisely AI, but also cyber criminality, or Gen Z’s expectations, …. Here, the potential benefits (or absence of losses) have been quantified. For sustainability/ESG, it is only now that indeed such quantification is starting to happen.
As illustrated (again) in Figure 1: Step 2 and Step 3 are components that no doubt offer some interesting benefits for the context given. Also for today’s focus theme. Simply because, and as discussed in the previous posts of this series: By analysing large volumes of data, AI can uncover trends and correlations.
Applied to the realm of financial management, this hence could help managers understand the financial implications of their sustainability initiatives.
But what are those benefits specifically?
And which areas of application do they emerge from?
In the remainder we’ll be looking at a small number of areas where AI could contribute to sharpen, or indeed to enable, the integration of ESG data with financial information.
- Cost Accounting & Cost-Benefit Analysis
- Performance Metrics → Dashboards
- Predictive Analysis: Financial impact of ESG trends and regulations
- Predictive Asset Management: Maintenance cycles
Cost Accounting & Cost-Benefit Analysis
In a company, every project, initiative, team … has been an assigned an accounting role. Either as a cost centre or as a profit centre; sometimes also e.g. a team may be both (e.g. the sales team). Extremely simplified: The first ‘looses’ money, they second ‘makes’ money. Fixed costs such as office rents, are usually allocated (often: proportionally, but not always) to the different centres as they are an organisation wide ‘common good’.
Projects, and specifically change management projects, are cost centres by definition, as their profits usually turn up not at its own door step, but – if at all – just about anywhere else in the organisation, and often quite delayed.
The benefit of AI in management accounting is of course the power it has to crawl across large data sets, allocate where individual data pieces belong, and then summaries this efficiently. It notably can learn from prior data sets that may have been compiled manually, to build on, and likely improve the work done manually, and replicated it on a new, not yet manually treated data set.
This is particularly useful when it comes to evaluating financial costs and returns for concrete change projects of all types. Be it the building of a new distribution centre, the acquisition of a new, in-house waste water treatment plant, or the optimisation of when to upgrade / adjust the company’s vehicle fleet.
If we define ‘return’ to mean additional quantitative measures in addition to money, AI can help us to figure out the cost / energy / Co2 / write off / … equation of solar panels or heat pumps for example. Or help the CFO to fine tune the fixed cost allocation to different products / teams / projects. And the product managers to understand the contribution each product has for the company’s CO2/Energy/Water/Material/Waste footprint.
Real Life Examples:
- Such analysis lead Unilever to be able to say that more sustainable products, and entire such brands, where growing stronger and performing more successfully in the market.
- Similarly, BP uses predictive analytics to assess the financial impacts of carbon pricing on its operations, which helps them taking more informed decisions on their investment into renewables
Performance Metrics → Dashboards
Once the above accounting and analysis homework has been completed, the subsequent step is to figure out which variables drive such and similar impacts – hence where improved sustainability KPIs cause (drive) lower costs or higher profits – and generate Dashboards for the primary functional areas of a company.
Added brownie points: this can theoretically be updated on a regular basis, all the while ensuring that historical tracking remains possible.
In extension then to these performance metrics, the Dashboards could offer their visualisations and ‘at a glimpse’ view. At the last specific to functional areas, possibly – depending on how clear roles are specified – even role specific.
Let’s say for example: the materials manager will be shown his/her individual impact on Co2 emissions, margin of the product, and the amount of ‘more responsible materials’ across the whole product range, as well as the dependency link where existent.
Real Life Example
- Microsoft uses AI-driven dashboards to track its progress and associated financial implications, emphasizing the ROI of its sustainability investments
Predictive Analysis: Financial impact of ESG trends and regulations
Every new law and trend has an actual, or at least potential, knock on effect on a company’s financial figures. Either through extra spend, higher costs, loss of income.
Many companies, as soon as a trend seems to emerge, or a law pops up on the horizon, (try to) figure out what that will mean for their financial health, their operational processes, as well as forward going market presence.
In fact, some companies – largely associated with company size, albeit not always – run a regular exercise modelling such impacts onto their finances.
AI evidently can help with that on various levels, for example:
- Analyse incoming legislation and its interfaces to either operational processes or product-to-market requirements. A poignant such example would be the CSDDD.
- (Support in) Identification of changes in the processes and product requirements demanded by legislation. And based on a company’s internal costing models (time, materials, ….) elaborate a first financial impact estimate.
- Estimation of externalities that either already are, or are bookmarked to be, internalised (e.g. CO2, water, waste, ….), as well as a first identification and costing of externalities not as of yet to be internalised any time soon.
All of these dimensions have evidently impact on some positions of the balance sheets, at the very least mid-term, and the calculation of such financial impact data aligned with the cost and benefit structures of the company evidently has relevance.
Predictive Asset Management: Maintenance cycles
Assets have write off periods, upgrade and maintenance cycles, as well as ‘replacement’ timelines. They are interdependent for evident reasons:
The write off period is often influence by tax law and a company’s tax scheme, and depends on the size of the original asset investment. Maintenance is dependent on a the usage of an asset, as well as external influences that affect its health. And replacement timelines again depend on tax requirements, insurance premiums, technological advances, usage hours and modalities, as well as the success (or absence thereof) maintenance efforts.
Maintenance cycles are already know often flagged in the production management software in cases of factories. For real state assets its the ‘assumed’ live cycle and market standards that drive maintenance (aka: refurbishments or upgrades).
But the moment all of these variables are not considered in isolation from one another, but interlinked and interdependent, it is challenging to come up with suitable information and insights that apply e.g. to a specific machine, an individual house, or indeed a subunit (engine of a machine, apartment of a housing block).
AI, in principle, is capable of helping us with that by doing this rather complex, very manual, both tedious and highly skilled task. Mapping the different cycles against one another, that against investment cycles, tax requirements, and P&L/ cash flow optimisations.
Real life Example:
- Such efforts are already being put in place my large industrial conglomerates such as GE for their manufacturing.
Conclusion
Adding complexity to reduce complexity – this could be the motto under which the benefits of AI reside in the areas mentioned in this blog post.
Adding complexity by combining a number of data sources that possibly never have been crossed over to me optimised as a whole. And reducing complexity by getting help from technology in order to do what by hand would be extremely hard to do. And all that in real time, possibly ‘on the fly’.
The sheer effort of doing that alone is no doubt already an interesting exercise of ‘return-on-invest’ calculation … the invest being the AI-tool and the return being any cost savings and optimisations resulting from its application in an area so far never tackled at this complexity level.
There will be wins, and losses – no doubt of it.
