Is Automation automatically the answer?

Once upon a time, my company provided services to an organisation who had decided to implement a new leasing system.  They sacked all the middle aged ladies who’d been working there for the past quarter of century, using pens and paper to calculate customers’ lease liabilities, in favour of bright young things, straight from university to work on this new system, since clearly, old (i.e. mid 40’s) workers, were incapable of learning anything.

Can you guess the rest?  My task was to help fix the complete mess (i.e. royal f**k up) that resulted.   Customers paid off the balance of their lease, which was logged as a prepayment by the bright young things, and in a year’s time, lessors were contacted for the arrears on their ongoing lease!  They didn’t have NPS metrics back then, but you can probably guess the score.

Half the role was monitoring the backlog operation, set up to process the corrections on to this new leasing system.  This entailed understanding how the leasing system worked (which the leasing system provider was still developing) as well as how to capture the results of this monitoring (since Excel at the time only held 64,000 rows and there were 65,000 leases to correct). 

The other half was assessing the impact of all these corrections on the P&L for the organisation’s board – who had all signed off on the restructuring / lease system implementation two years before. Obviously, it was the leasing system provider, and bright young things, who had royally f**ked up… 

So, what can we learn from the experience?

Make people accountable for their decisions.   Implement clear metrics on benefits realisation.  The Board and Executive abrogated their responsibilities (whilst still taking their pay cheque) for the decision to move to an unproved system, and to restructure their team at the same time. 

First mover advantage may be over-rated - particularly for systems still in development.  The leasing provider had only recently developed its software and provided a lowball quote to, in effect, beta test its system in a commercial environment.  Ensure systems you procure are robust and your organisation’s not being used as a sandbox to work through the bugs from initial development.

Demonstrate a compelling reason for change – the former, archaic, way of doing business, using paperwork and a large number of staff, may have seemed inefficient, but it was still effective.  Effectiveness trumps efficiency every time. 

Consider the aspirations of your workforce – the bright young things from university were only going to stay a short time in the organisation, before they moved on to bigger and better things.  Whilst for the middle aged ladies, they knew each other’s quirks and habits, formed a tight knit team, and had no aspirations to move on from what they knew, and did, extremely well. 

Be wary of the current trend for “hyper-automation”. I struggle to understand what that means.  The best I can come up with is 0’s and 1’s binding together to make complex decisions in ambiguous circumstances.  How can you ensure that any biases in the processes are not simply replicated in the automation (given the current scrutiny of AI)?  The backlog prioritised fleet customers over individuals, given the volume of leases they held.  But should we have processed the backlog chronologically, or which leases had the most impact on its customers?  How does an algorithm transparently justify whatever choice is made?

So, do you really need to automate?  Consider what costs and risks you’ll incur, for the perceived benefits. A lack of transparency, a lack of control and a lack of knowledge, versus cost savings and the hassle of dealing with humans. 

Other questions to consider are who will sign off once the automated processes are in place? Where will accountability sit?  With the Board, the Exec?   Leaders can blame their tools, but ultimately, they’re accountable for how they are used. 

In closing, there’s no easy answer (“it depends”).  Executives need to consider their workforce and the benefits (and costs) they provide, before answering the question whether automation is the right solution for the organisation at this time.