Swarandeep Kaur Juneja

In order to pick the best suited solution to fix a problem, generate ideas targeting the root cause of the problem which is identified at the analyze phase.
 
Start by developing potential solutions for the validated root causes by using techniques such as: a) Round Robin Style of Brainstorming b) Critical Thinking via Six Thinking Hats
 
Evaluate solutions using: a) Pugh Matrix b) Delphi Technique c) Multi Voting d) Nominal Group Technique or/and FMEA. These tools will help prioritize one solution over the other considering multiple criteria such as cost and potential benefits. For e.g. Pugh Matrix may be used to draw comparison between two alternative solutions in a software company tackling slow bug fixes, if they should add more coders or enhance their testing tools. Both the solutions can be scored using multiple criteria such as a) Cost b) Data Privacy c) Accessibility d) Training Requirements e) Time to deploy etc.
 
Changes can be first tested using DOE (Design of Experiments to evaluate process inputs generating desired output. Post which the selected solution can be implemented (pilot run) in a smaller sub-section of the process before full scale implementation.

There are certain things that AI excels at like logical reasoning, for e.g. shortlisting candidate profiles based on qualifications, invoice processing, monitoring inventory, IT helpdesk chatbots etc. On the contrary there are certain subjective questions where AI struggles and is prone to decision making mistakes just like human beings, which is also known as the “Linda problem”, for e.g. decision making on HR policies pertaining to employee's behavior, issuing warning letters to employees or even termination letters in worst case scenarios.
 
There is a stark distinction between traditional automation tools and AI, as AI is not dependent on fixed rule-based decision making, rather it learns from previous patterns and relationships to determine the best approach to address a given situation.
 
A detailed example of area best suited for AI deployment in corporates is pertaining to IT helpdesk. Say for e.g. as per company policy all queries should be answered/resolved within 24hrs. AI can be extremely useful to categorize the queries correctly, route it to the appropriate team and even provide resolution to tier one queries without any human intervention.
 
As AI is dependent on previous patterns to fetch the best answer, if previous patterns include bias, that bias will be carried forward by the AI in selecting the best approach. Therefore, subjective decision making like termination or issuing warning letter to an employee based on certain kind of employee behavior cannot be completely left up to AI. For e.g. There is a corporate rule pertaining to usage of social media sites in office. If an employee is found accessing social media site, AI cannot be fully trusted to pass a judgement in this scenario as there might be a lot of other factors to be considered before a warning letter can be issued.

Often times I have heard people saying only repetitive, monotonous tasks can be standardized or only “factory model” can be standardized as that is transactional but “knowledge processing” tasks cannot be standardized especially when it comes to managing the tasks of Team Leaders and Managers.
Lean Standard Work can be used conceptually to avoid being completely dependent on Managers/Team Leaders to build the culture within their teams, their own way, without any pre-defined standard tasks to be performed.
Standardized work is applicable to a greater or lesser degree, across all organizational levels. We need to decide how much of leader’s work can be standardized.
As the name suggests Leader Standard Work is about standardizing some of the tasks performed by leaders and not only that but also bringing in a great shift in mindset that it’s only possible to standardize work at the executive level.
In order to develop common culture across the floor it is important that leaders are driven by common set of values. This means that even culture, behavior and mindset could be standardized to some extent.
Included in standard work for leaders (managers and team leaders) is up- and downwards accountability that is evaluated on a daily basis or several times a day using visual controls. Daily evaluation is critical, because it not only verifies that work is being done properly, but ensures that everybody is held accountable for working to the standard. Most of the leader’s standard work is focused on activities at the place where work is done (gemba) with the remaining time spent supporting incremental improvements. (https://traccsolution.com/blog/leader-standard-work/)
How can this help build continuous improvement culture?
As it is often said that “What gets measured gets managed or what gets measured gets done”. Based on this concept in order to identify opportunities for improvement it is crucial to measure the current performance and in order to measure it, it is also important to draw out a list of standard practices to be followed by the leaders across the organization. This would in turn help build a continuous improvement culture where there are no biases when it comes to measuring the performance of individuals at each level in the organization.
Advantages of Lean Standard Work -
1. As they say right examples are set at the top of the organization and from there culture seeps down to lower levels. If standard set of expectations are set from the leaders and their tasks are evaluated, this would create a culture of “fact based approach” in the organization which would in turn eliminate people dependency.
2. For continuous improvement it is important to have quantitative data and such practices ensures that.
3. This also helps in eliminating the concept of each team or each department having it’s own culture. It is important that the organization is run on some common set of values.