HOW LYTICA BECAME A UNIQUE ANALYTICS COMPANY: PART 10
I am in verification purgatory! I use the term purgatory not because of any religious affiliation but because it fits, it allows for hope. In Roman Catholic doctrine, it is a place or state of suffering inhabited by the souls of sinners who are expiating their sins before going to heaven. While the terms sinners and sins don’t really fit my reality, the terms suffering and going to heaven do – if you don’t take them too literally. There is a staggering amount of painful verification work for my team to perform that, once finished, will put all of us in a better place.
As the verification of our AI initiatives are underway, I get a glimpse of what the future holds for supply base operations and, in planning for 2018, some time to think about how our work and that of others will impact our profession. There is no doubt about AI changing the way work gets done.
There are many examples where AI is outperforming humans in speed and accuracy. The quality of its output is just better. In his 2016 book, The Fourth Industrial Revolution, Klaus Schwab writes about computers becoming exponentially better at understanding the world.
Some examples he cites:
- IBM Watson gives basic legal advice within seconds with 90% accuracy as compared to 70% accuracy when done by humans
- Watson helps nurses diagnose cancer 4 times more accurately than human diagnose alone
- Facebook has a pattern recognition software that is able to recognize faces better than humans can
Given that these examples are over one year old, they understate the capability of focused AI.
One of the reasons AI has taken so long to deliver on its promise is that people are trying to create a general-purpose system to mimic human intelligence; building something this universal is a really hard assignment. On the other hand, focused AI addresses only a single subject or topic making the task much simpler. The advancements made in the pursuit of general AI have been applied to focused AI, making it a fruitful avenue for significant productivity gain. The work being carried out at our Advanced Technology Center applies focused AI to applications which address cost, risk and compliance at the electronic component and, next year, product level.
On the surface these advancements would indicate that jobs will disappear, but other evidence shows a different outcome. Amazon, with a heavy investment in AI and automation, has had a correspondingly large increase in hiring. Doctors and radiologists using AI are spending less time looking at X-rays and more time with patients, leading to a better health care outcome; what AI has not been able to replace is the human interaction aspect. So, while some jobs disappear, others get created and many transform into something more interesting than before. It is unclear if a net increase or decrease occurs. Proponents of the glass is half empty view swear that there is a net decrease, but they have been saying this forever while the history to date points to job growth. From my experience with the ATC, I see growth in verification staffing outpacing all other hiring in the field of technology.
It’s interesting to speculate about what will happen to supply chain jobs. To do this, we must understand what the purpose of the jobs are rather than considering the job to be a sum of tasks. Most tasks can be automated or replaced by AI. If there is a distinct purpose of the job rather than it being a task repository, then AI can free up time for the job holder to achieve greater things. In the case of the doctor, their purpose can consist of providing quality health care or eliminating disease. The X-ray diagnosis is a task that becomes automated, the provision of quality health care and the elimination of disease remains the doctor’s challenge.
Let’s examine some supply chain jobs based on definitions taken from Wikipedia:
Purpose: Ensure the availability of suitable components required to manufacture a product.
- Analyze and qualify interchangeable parts from vendors & manufacturers
- Maintain close association with design and manufacturer
- Product lifecycle management
- Finding alternate components for discontinued components
Tasks 1, 3 and 4 can be automated but the second task requires relationship management which is very difficult (maybe impossible) to do with current machine technology. There are strong relationship elements in this position that will be hard to replace. Some component engineers will always be needed although maybe not as many as today.
Purpose: Develop a systematic approach to the entire usage cycle for a group of components through strategic positioning of that commodity with respect to an organisation and its suppliers.
- Work within a multi-step operational supply chain management process
- Supplier Relationship Management
- The tasks of a Component Engineer
Much like the first example, only some of these tasks can be automated or made better with AI. There are strong relationship elements in this position.
Purpose: Create reliable forecasts.
- Work within a multi-step operational supply chain management process
- Improve the accuracy of revenue forecasts
- Align inventory levels with peaks and troughs in demand
- Enhance profitability for a given channel or product
Once again there are strong relationship elements in this position preventing complete elimination through AI.
Logistics and Distribution Manager
Purpose: Organise the storage and distribution of goods.
- Ensure that the right products are delivered to the right location on time and at a good cost
- Coordinate the storage, transportation and delivery of goods
- Oversee and liaise with colleagues to ensure stock is maintained and moved efficiently
- Oversee warehouse, inventory control, material handling, customer service, transportation and planning workers
- Hire, train and evaluate employees
- Prepare worker schedules and ensure workers follow safety rules
The same conclusion can be drawn about the importance of relationships in this position.
All four of the jobs outlined have elements which can be automated yet all require the strong relationship elements that AI struggles with. This strongly suggests that these jobs will morph into something more interesting rather than completely disappear.
While most of the examples above suggest job transitions, one sure cause of job loss is bankruptcy. Companies that fail to adopt new methods enabled by technology like AI have a higher probability of failure than those who adopt and adapt. One recent example involves the chapter 11 filing of a well-known communications company that benchmarked with Lytica about 3 years ago. When compared to two or three other companies in the same market vertical, they benchmarked 20 percentage points lower in competitiveness than their peers. When their competitive position was translated into cost, they were at a significant disadvantage. Despite being told by Lytica exactly which components were overpriced and the price that they should be paying, they failed to act effectively on this information. While their competitors acted and improved margins they remained disadvantaged. This Freebenchmarking.com example, and new insights that come from our prototype AI applications, strongly indicate that people in key positions must act quickly to gain advantage. Action engagement based on informed advantage is cultural, those companies with a change ready culture will prosper.
In my view job loss will result from inaction in the face of compelling evidence that the status quo has shifted. Job transition will occur when action is taken and mundane tasks get replaced with knowledge based assignments leading to new accomplishments. Job creation will result from new opportunities enabled by business growth in companies that have a culture of early adoption and appropriate change management. No one should expect to see what they are doing today being done in the same way for much longer.