Traditional Organisations are ill prepared to compete in the age of AI
For the past decade, we have heard and read so much about digital transformation and have seen as many organisations attempt it one way or another. Some set up innovation centers, which (for unknown reasons) comes with customary bean bags, colourful furniture, sexy touch screens and a host of other gadgets. Sadly, many eventually end up being fancy meeting rooms. Some manage to develop successful Proof of Concepts which work wonderfully but fizzle out as soon as it is rolled out. And over here in Malaysia, I have come across several such examples and sadly it doesn’t discriminate by industries or the size of the investments.
It got me mulling.
I reflected upon some of my past similar experience both as a consultant for clients and when I was leading the transformation myself. What does it take for a traditional organisation to fully embrace and ‘exploit’ the age of AI. Is it about having the commitment from the top brass, having access to latest tech, aligning the organisational culture or having enough patience to see through the investment?
My conclusion is that it’s all of the above plus also, having the right operating model; a realisation that hits me when I was reading the book “Competing in the age of AI”.
The operating model for traditional organisation has been shaped over the decades, imbued with leading practices and differentiating processes in order to win in the marketplace. And the operating model has matured to a state where it becomes very good in what it is supposed to do. It is also the very same operating model (which includes organisational culture and technical architecture) that is inhibiting its adaptation to the age of AI. It is the inertia for change.
This can be best explained by looking back at history.
Prior to the industrial revolution, traditional industries relied on artisans to create an entire product, from start to finish. It is known as “filing and fitting” — making adjustments by hand “filling” each part so that it would fit into the entire assembly. It all changed when the Industrial Revolution transformed the production technique. End products are now componentized and created part by part. After which it will be assembled individually, similar to a production line. Workforce is trained to build and assemble individual parts e.g. a door handle as opposed to building the entire car. This increases efficiency, having the workforce build deep skills on narrow tasks. It creates a more predictable and repeatable outcome while building a specialised workforce. Quality and efficiency improves naturally.
And as a result of that, organisations operate as a collection of groups (aka departments) with distinct capabilities and responsibilities. For example sales, marketing, technology, manufacturing or distribution with specific KPIs but also demarcated silos. Silos of work responsibilities, workforce capabilities and data.
This is the traditional operating model that we are mostly familiar with. But such an operating model does not fit well with digital organisations, such as Alibaba, Grab or AirBnb. Digital businesses operate on different business and operating models.
As a comparison, traditional business competes in its own industry vertical. And it normally competes with strategies such as differentiated products, mastery in supply chain, extensive distribution network, niche positioning, unique branding and/or superb customer excellence
However, Digital businesses compete across multiple industries. It doesn’t discriminate by services or geography. It aims to dominate via the network effect using Data and AI. Look at Alibaba. It has businesses ranging from finance, ecommerce, insurance, lending, services to tech. And who knows what next. Same goes for WeChat or Grab. And because of the network effect, the more users it has , the ‘smarter’ the AI model gets and the more services it provides, the more ingrained its business model gets.
To do that, digital businesses build AI and Data as its core of the business. This is what the book calls the AI factory. Data is aggregated across different data sources to a single common repository (single source of truth). Algorithm development is layered on top of the data repository and agile teams across all lines of business experiments and take actions using the models. Models are then deployed organisation wide with a built in learning mechanism in the form of a feedback loop. All lines of business must be ‘plugged in’ in the AI factory — to provide data, make decisions and feedback the learnings back to the factory.
This is a self perpetuating model, where more data improves the model and hence provides better insights to strengthen the network effect. And digital business can scale with very little incremental cost compared to a traditional business.
Image drawn based on the book “Competing in the Age of AI” by Marco Iansiti and Karim R. Lakhani
Coming back to the traditional operating model, why doesn’t it work?
As mentioned, the traditional operating model focuses on specialisation. This is to build deep vertical skills and deliver that with lowest possible cost (effort). And as mentioned earlier, this has created silos of data. The silos of data are notoriously hard to be amalgamated. There are many technical challenges (data from different subsystems, standards and definitions) as well as organisational challenges (head of departments protective of their data and their own positions).
Organisational KPIs are also not aligned for collaboration across functional units. Sales would be rewarded by sales closed, call centers would be rewarded by the number of cases handled and IT would be rewarded by total system uptime.
Like it or not, the age of AI is here. Traditional organisation with tonnes of legacy and built-in habits must be repositioned to stay relevant. Unlearn and relearn. It’s more than just having brilliant data scientists locked in a state of the art innovation center. The change is here, similar to the change during the Industrial Revolution. Much like how Ford’s production line churned out Model T to dominate the market.
And the change will impact everyone, beyond just the Data team in your organisation. I will end the article with a quote I picked up but can’t remember where — Jobs are not displaced by AI, but displaced by employees who use AI.
Freddy credits the book “Competing in the age of AI” which he has referenced for this article. More importantly, it has resonated with some of his past experiences in AI and Analytics transformation. Reminds him of many bittersweet memories!
