Evolving to AI for Finance Teams and the people they serve

Filter By Topic

All

Just before we start

I had to write a piece for an upcoming event we were doing and so I did as I usually do.  I sat down under pressure and produced pure bloggable Gold in an instant.  I wish.  I wrote the article that follows that captures some of the conversations that I've been having lately with vendors, visionaries and Finance professionals.  That done, as you would no doubt know, I had to provide an engaging title and suggest some artwork.

I think you'll find the article engaging as it moves along at a reasonable pace, but drawing a single theme from it was challenging - there is a lot on people's minds right now.  I settled for an image of the double helix of DNA because evolution, in terms of information creation and handling, is accelerating logarithmically all around us.  I've come to think of the evolution in reporting and planning capability and the evolution of the "business facing" piece of the Finance function as tightly linked and symbiotic.  One facilitates the other which then creates further demands and the whole thing moves forward as each builds momentum from the other.  Why does this matter?  Well if we do it right we substantially scale up our potential, as Finance Professionals to continue to raise our relevance to corporate performance while the nature of our daily work rapidly evolves.  The long-held (I heard about these 30 years ago when I was at PWC) aspirations of being the trusted business advisor and high impact commercial player can be realised as we get further and further away from the mechanics of record keeping and compliance monitoring.

Let's dive in...

One Version of the Truth

"Data democratisation" they cried back in 2008.  Now there's a concept that only an evil data dictator couldn't love, right?  Who wouldn’t gladly attach themselves to the freedom movement whose protagonists dreamt of a better world of data democracy and the fight against the tyranny of data?  More correctly the tyranny of the absence of actionable information and knowledge.

The answer, we were told, lay in better tooling, in the hands of everybody - "everyone can be an analyst" and the age of the citizen data scientist was born.  Everyone can roll their own, the one version of the truth is unattainable.  I actually heard that presented at a Gartner conference in 2016.  Of late I have been speaking with many finance professionals at conferences, as well as our clients and another trend is emerging.  Among the sea of new tools, there is a growing realisation.  For data to have power, to become actionable, it must be relevant, timely, accurate and distributed.  To achieve these things require more than just casual focus and they remain at the core of what Finance Teams are expected to provide.

That's a little back to the future, isn't it?  It is, in fact, a "re-"emergence.  No matter, it just signifies that the ground rules haven’t changed.  It does have a fairly profound impact on how we think about making the right strategic decisions when considering decision support system design (DSS or whatever acronym du jour you currently run with).  One version of the truth matters more today than it has historically.  This is driven by one mega-trend.  Signal to noise ratio in a post-digital world.  So much data (noise) but what is a truth that is worthy of trust and can/should be actioned (signal).  Sifting these is the very first step before prioritisation turns into action and while data is shattered into silo's, organisations can easily stumble at this first hurdle.

What is more dangerous than not knowing?  Being convinced you know and being wrong.  "Barney: Careful there.  Is that cliff edge even stable?  Fred: Suuuuuuuuuuuuuuurrrrrrrrrrreeeeeee!".  The problem of spreadsheets has always been errors, duplication, data isolation and data manipulation.  Many of the new tools meant to democratise have, by creating small isolated data silos, made the classic spreadsheet problems much more engaging (through visualisation) and convincing.  The Achilles heel, though, wrong data driving bad decisions - remains unaddressed, as dangerous as ever and now simply more obscured and enticing.

Sharing a single model is power, not just semantics

In Finance, when we design reporting systems we develop a semantic model (all the cost centre, project, account, time etc dimensions and hierarchies) that by their very structure explain the design of the business.  The business changes over time so we enhance the reporting model to reflect our expectations about the future without losing our memory for what happened historically.  In this way, we get to measure past and present relative to each other, see what is improving and what isn't.  To this reporting model, we add budgeting and planning.  Here we attempt to predict the future based on the past, the model and our expectations about the future that are mostly exogenous to the past.  We change the model to reflect our expectations of the future so that it remains relevant to both past and present.  Occasionally we seek to make changes that to the model are so "radical" that we can only really keep the past at high levels of granularity.  We also add calculations to the model of our business to allow us to simulate the way we expect it to operate.

Throughout this whole process, we recognise that at any point in time, the model is largely constant and time is the variable.  In other words, the ultimate goal here is a single place from which to manage the key Corporate Performance Management tasks of Predicting, Shaping and Achieving our organisation's best financial future.  That single place is One Version of the Truth and the best way to achieve that is through a single integrated model of the business that reports actuals, as well as compiles budgets plans and forecasts and a single interface - not 8 different tools with data, sets splintered across a range of systems and worksheets.

Having achieved this one version of the truth, we as finance professionals, are in a position to deeply engage with our business, to speak with requisite authority to help change happen without becoming lost in spreadsheets or constantly let down by bad numbers or continuing surprises.  It also means we can effectively engage with new technologies like AI.  Machine learning (ML), a form of AI, relies heavily (almost exclusively) on well structured historical data.  If your data sets are splintered and poorly curated, ML cannot be for you.  From a single version of the truth, however, ML can become a new, and powerful voice in the room, when it comes to envisioning, agreeing on and sharing your company's vision of its best financial future.

"Do or Do not, there is no Try" - Yoda, The Empire Strikes Back

For Finance teams then, all fads (and dashboards..) aside, the One Version of the Truth is more important now than its ever been and is on the critical the path to really leveraging practical AI applications for Finance.  Once achieved, you and your colleagues can step out from behind those spreadsheets and get on with the real work of the "business facing" Finance team.

So here is a challenge for you - does your team ensure your company's ongoing great health by orchestrating that One Version of the Truth?  The great news is this doesn't have to take a long time to Do.  Complex things are complex, but most of these projects are well established within days and delivered in weeks.

You are the vital component in helping your business leaders Predict, Shape and Achieve the best financial future for your organisation - best of luck and have fun making it happen.

Author: Dominic Parsons, CEO CALUMO

Learn more about budgeting, planning and forecasting for your business here.