Sachin Tanwar

Designing Your Chatbot PoC – Keep it real: Focus on a Specific Problem: Don't try anyhow. In a branched self-service knowledge base, where customers are facing problems, Chatbots can answer their questions, direct them to the desired solutions, or offer quick problem-solving methods to improve customer satisfaction. Thus, setting small achievable goals rather than setting big ones creates a platform where one can evaluate one's success more accurately. Define Clear Goals: Why are you doing this? Why do you need it? Would you like to ask them if it has been? Quantify your goals and discover your real purpose (e.g. "reduce average wait time by 20%"). This will let you see one thing to work toward, and you can see the finish point. Choose the Right Tech: Whatever the form, chatbots have been identified as one of the solutions that could be life-saving. Look for factors such as user-friendliness with your previous system (CRM, etc.), AI's ability to be taught, and for sure, it will also be the cost. Do not be lured in by whistles and bells that are of no use to you. Train, Train, Train: A chatbot is equipped with training data that can only be as good as the training it was given. It is advised to feed a chatbot with a great deal of real customer interactions without forgetting all aspects of confidentiality. The closer the data is to the actual word format, the better the bot will be at understanding and answering the questions. Human Handover: Major! The bot will cease executing commands if a customer expresses a need to interact with a live person and will instead switch the conversation to a human agent. The parameters of escalation must be precise in this case. Then the customer will have a smooth transition to the live chat. Being trapped in constant chatbot loops is the thing which no one likes. Measure and Iterate: Close the data. PoC means proof of concept. Did the bot deal with its inquiries with zero complaints from clients? What was the tactical implementation of the bot, and how did it really perform? Where was the point that faced it the most, and how did the data you receive about it help to refine the bot's training as well as improvement? The point of the PoC is to learn, improve and not to get the perfect result. Risks – What Could Go Wrong (and How to Avoid It): Unrealistic Expectations: Chatbots are not magic. They are not a miracle solution that will immediately make all your customer service problems disappear. Ensure that you are managing your expectations properly and that everyone on the team is participating in them being a trial rather than a full solution at this stage. Poor Data Quality: Garbage in, garbage out. If a robot is trained with bad data, it will give wrong answers. Ensure that your data is clean, correct, and a close representation of a real customer interaction. Integration Issues: Negotiating the chatbot's relationship with your existing systems can be a bit difficult. Provide a detailed plan and initiate your IT department early. Unexpected integration issues may cause failure in your PoC. Customer Frustration: A chatbot with insufficient functionality can indeed cause customers to get even more frustrated. You are in real trouble if the bot is slow, not helpful, or hard to get to a human. The user experience is the number one aspect that needs to be taken care of. Lack of Buy-In: The PoC is likely to have no chance if your customer service team is not supporting it. Tell them about the pluses and cons of them and include them in the process. They are the ones who would interact with the chatbot so their say is valuable. Scope Creep: It's quite simple to go overboard and include more features in the PoC. Don't do it! Don't change your mind. Just complete the tasks according to the plan, then expand if necessary after the first step is approved. Through narrowing down a certain problem, culminating a specific goal, and keeping an eye on these risks, the chatbot PoC can be more productive that it would otherwise.

The evaporating cloud technique enables us to see the unseen. In the case of deployment of AI chatbots, we should cautiously keep in mind the sensitive stability among price performance and most fulfilling consumer service. To use the evaporating cloud technique, we can split down the problem into its fundamental assumptions: Assumption 1: Quick deployment is required to cut costs immediately. Challenge: Can a phased deployment plan strike a compromise between cost reduction and quality assurance? For example, launch a limited version of the chatbot for specific, low-risk requests while training it for more complicated conversations. Assumption 2: To provide high-quality responses, substantial training is required. Challenge: Can machine learning techniques be utilised to speed up the training process? The chatbot could be taught more effectively using massive datasets and new algorithms, thereby lowering the time required to reach high-quality performance. Assumption 3: Cost-cutting and quality are mutually exclusive. Challenge: Can creative solutions be developed to do both; cost cutting and maintain quality? possibly outsourcing certain chatbot duties to a third-party supplier or leveraging open-source tools could save money while also investing in worldly-wise AI techniques to increase quality. By challenging above mentioned assumptions, we are able to open up to new options and doubtlessly obtain a win-win outcome. In this scenario, a balanced method that mixes speedy deployment and non-stop development can be the great choice. This could permit the corporation to make short price discounts even as keeping long-time period purchaser happiness. The evaporating cloud approach is more than a problem-fixing tool; it is an attitude that believes that disputes are regularly because of tunnel vision, and that with the aid of testing our assumptions, we are able to find out solutions we in no way considered ever before.

Self-serving bias is this sneaky neural trick where we usually tend to steal credit for our successes and blame something or someone else when things go haywire. In LSS projects, this can be a critical issue in the area where we're trying to figure out what is or who is actually causing issues in a process or during RCA. Let's say you are working with a team of 15-20 and something went wrong, you want to protect your ego now, right? That's natural and basic instinct, that's where you might unconsciously: Ignore your own role in the process Point fingers Minimize your contribution to actually create solutions, just hover How does it impact during Root Cause Analysis? Disregard your personal contributions where you may tend to downplay their involvement in the creation of the issue. However, overemphasize personal contributions to solutions can also occur; you might overestimate your significance in resolving the problem which is basically stealing credit of your teammate who actually came up with a solution. The natural instinct to blame others for failures is another usual behavior; where you will shift responsibilities to other factors or other people, although others may play a role in the situation as well. Strategies to Mitigate the bias and ensure objective decision making: Establish a blame-free environment: Create a blame-free workplace by emphasising that the goal is problem solving rather than attributing blame to people. When people feel comfortable admitting their mistakes, they are more likely to be transparent. Data Driven Analysis: Data-driven analysis prioritises objective data over subjective feelings. Numbers are objective and do not lie; consequently, using analytical tools such as fishbone diagrams and statistical studies forces one to evaluate tangible data rather than relying on instinct or personal experiences. Rotate Team Perspective: Rotating team perspectives can help find ignored features by analysing the same issue from various angles, equivalent to engaging a variety observer. Bring In Outside Perspective: Bringing in an outside opinion is also worthwhile. Occasionally, a new perspective—perhaps from someone who is not emotionally committed in the project—can reveal features that the inside team may ignore. Finally, practicing radical honesty is vital. When potential biases occur, team members should be encouraged to identify them; it is okay to express concerns, such as saying, "Hey, I think we might be avoiding looking at our own roles here." This method promotes an environment of transparency and responsibility, which leads to more effective issue solving. What's the bottom line? Stay curious, humble, and remember that in LSS, the process reigns supreme. Your main responsibility is to improve the system, not to protect your ego.

What is Quantile-Based Analysis? Imagine you have a pile of rocks. You want to understand how big the rocks are but just looking at the biggest and smallest ones won't tell you the whole story. Quantile-based analysis is like sorting the rocks into five equal groups based on their size. It helps you understand the distribution of rock sizes, not just the extremes. Let review a real-world example. Let's say we're looking at salaries in a medium-sized BPO. Instead of just saying "the average salary is $75,000," quantile analysis helps us see the full picture. Here's a simple salary quantile breakdown: Quantile Salary Range What It Tells Us 20th Percentile $55,000 20% of employees earn at or below this 40th Percentile $65,000 40% of employees earn at or below this 60th Percentile $80,000 60% of employees earn at or below this 80th Percentile $110,000 80% of employees earn at or below this Now post evaluating the above spread, we can see that average salary might be $75,000, but most people are not exactly earning that, some are way high, and some are way too low. This is a classic example. Quantile analysis is a sophisticated technique that provides many benefits such as: It doesn't fall apart in the face of extreme values in the same way a mean does It reveals the actual real-world distribution, not just a single number It detects inequalities or patterns that may be concealed by the averages Even though it's inevitable. Disabilities include: Needs a satisfactory amount of data to be meaningful Might be hard to communicate to those who only like simple numbers Has to be done quite frequently by software generating statistical data to make exact calculations Suppose for example that you're a city planner who wants to know where people can buy homes at various prices. The mean may tell you "$300,000," whereas the quantiles can show you that: 20% of them are the only ones who can afford houses that are under $200,000 80% are the ones who will not buy any property beyond the price of $450,000 This allows you to learn about housing inequality in a different, perhaps more comprehensive way compared with the usual method using entire datasets. The analysis of quantiles is like X-ray vision in the fields of data science and research. It allows you to look past the superficial numbers and get a handle on the information that is really being communicated. Always remember: Number will tell the stories, but quantiles will help you read between the lines.

When businesses are measuring the (ROI) from AI solutions, they should adopt a balanced approach that involve short-term gains along with the long-term value. Short-term benefits are like quick wins, what i mean is short-term gains are cost savings, faster production, or a quick jump in efficiency overall. There were times when AI implementation shows that, most- AI solutions are easy to measure and often display the quick results after implementation. On the other hand, long-term value creation has broader scope. The benefits included in this are the better customer satisfaction, informed decision-making, and the innovation aspect. However, these benefits can be uncertain in nature, but they can bring the company a long-established growth path and a better market position. Below is the representation of short-term benefits versus long-term value creation on foundational view: | Benefits/Value | Short-Term Benefits | Long-Term Value Creation | |----------------------|---------------------------|--------------------------- | | Time to Realize | Instant to a few months | A few months to a long long| | Examples | Cost savings, efficiency | Customer satisfaction, | | | gains, quick productivity | innovation, competitive | | | boosts | advantage | | Measurement | Easy to measure | Harder to measure | | Impact | Quick wins | Sustained growth | | Investment Focus | Quick returns | Tactical investments | By looking at both the short-run and the long-run outcomes, companies can make better decisions about their AI investments and make sure that they are not just running after temporary achievements but also building for the future.

Non-parametric analysis is one of the statistical methods that do not make any assumptions about the underlying data distribution. Unlike traditional methods, it is not required data to be more of a specific pattern like a normal curve. It is therefore flexible, and it can be used in a lot of cases. Industries Using Non-Parametric Analysis: Healthcare: Medical analysis that deals with problems where the distributions are not normal of the medical data like patient outcomes or drug efficacy is often one of the approaches. Social Sciences: Gathering data about people's behaviors, attitudes, and opinions is traditionally the basis for the process of information mining resulting most of the time in nonstandard structures. Environmental Science: environmental data which contains examples such as pollution levels and climate pattern involves non-normal distributions. Finance: In the case of finance, the stock market is analyzed by means of data that do not conform to a normal distribution along with investigating market trends. Advantages of Non-Parametric Analysis with Examples: Flexibility: Insensitive to Traditional Assumptions, Technically, it can analyze data that do not obey the traditional assumptions, creating the possibility of it being used in different contexts. Example: How about assuming you are exploring whether the efficacy of two new drugs is truly different? When the datum on patient recovery times does not match with the Gaussian model, nonetheless, Mann-Whitney U test which is meant to be a non-parametric test can still be used efficiently between the two groups. Robustness To Outliers: It is more to the extent of not being as susceptible to outliers and extreme values, which sometimes may be found due to sampling error can change the results of parametric tests drastically. Example: In a study on income allocation, some very rich individuals might alter traditional tests in a manner in which they become useless. Yet, a non-parametric test is relatively less disturbed by outliers and thus, conveys a clearer dataset. Ease of Interpretation: Non-parametric tests often lead to more precise interpretations, hence there is no need for complicated calculations and operations. Example: Experiences of non-parametric tests often result in simpler models that are, sometimes, easier to handle and describe, especially in a highly sophisticated data set. In essence, non-parametric analysis is a valuable tool that can help you make sense of data, even when it doesn't fit the traditional mold.

Key Problems Along with Examples in LLMs till today: 1. Real-world Understanding: Problem: LLMs have been trained using massive text datasets, although they haven't really grasped the earthbound reality. They handle information and produce text even though they are deficient in real-world background and common sense. Example: Suppose you tell an LLM, "Which is heavier, a pound of feathers or a pound of bricks?" it might get puzzled. Although it understands the notion of weight, it may find it difficult to relate to the physical fact that each has the same mass. 2. Reasoning and Logic: Problem: LLMs can at times commit logical blunders or be inconsistent when they are giving answers. Now and then they may find confronting complicated reasoning tasks a challenge, especially when such tasks require multiple steps and the use of abstract concepts. Example: If we gave an LLM the task to unscramble a riddle that states "What has an eye but cannot see?", the system might end up robotically churning out verbose explanations without coming up with the solution. 3. Bias and Fairness: Problem: LLMs can get the biases covered in the data they are trained on, which can make training data biased or unfair. This would result in their unfair or discriminatory outcomes in situations where racial, gender, or religious topics are involved. Example: A situation where LLM is trained on a dataset containing biased language can result in the tool emitting texts that bolster or perpetuate stereotypes and prejudices. 4. Creativity and Originality: Problem: LLMs can come up with many different creative text formats but may be suffering from truly uniqueness. They are more inclined to patterns of action and information that they have already gathered during training. Example: Suppose you pose the matter of composing a poem to a LLM, extra-ordinarily clever it might churn out a fitting roughwork of such kind, yet the poetry will be behind the mastery of human imagination. Yet, even knowing this, LLMs are still extremely powerful aids. However, they are fundamentally human as technology is developing. Therefore, even though these issues may be less problematic in future, currently they are still a big hurdle.

RAG alongside fine-tuning are two distinct approaches with different capabilities that a large language model can acquire. RAG (Retrieval-Augmented Generation) What it is: RAG is a combination of language model and retrieval system. This means that it can get a reference from a database or a document in order to provide the most exact and actual information. When to use it: RAG works best when the model has access to the bulk of data that might be time-varying such as company policies and product details. It is a perfect way to do tasks where the reference or data is quickly changing. Fine-Tuning What it is: Fine-tuning is the process of making the language model better at specific tasks by training it on a certain dataset. The parameters are then set up to align the model to the new data. When to use it: The fine-tuning method is the best choice when you require high performance in the model when performing a specific task or set of tasks, for example, customer service for a specific product, and interpreting medical terminology. It helps in such cases when the data does not vary regularly, and you need high precision. Using Both Together When to combine: There are times when you may wish to switch between RAG and fine-tuning, which is also an option. For instance, you may fine-tune a model to fit your organization’s customer service to then exploit RAG for the new product information. Hence, the model is both knowledgeable and current.

Can you imagine being in position to repair an overflowing tap? You could simply turn the knob a bit, but what if the actual issue is that the washer has been deteriorated too much? This is the point where Systems Thinking and Design Thinking can come in to help businesses. Systems Thinking can be compared to puzzle-solving when you look at the whole picture not merely the faucet. It allows you to see the connections between different parts of your business. For instance, diminished sales might be due to bigger ones of your sales which are not merely sales. Possibly, it might be due to the product which does not meet customer needs or the marketing which does not reach the right people. Design Thinking suggests that you should try to see the thing as your customers do. It supports you in coming up with out-of-the-box ideas that are based on reality. For instance, instead of fixing the dripping tap alone, you could design an entirely new bathroom which is not only more efficient but also more stylish. Combination of these two ways of thinking results in the birth of an excellent business tool. It can be illustrated as follows: Wholesome Problem-Solving: You can perceive the bigger picture and deal with issues from the ground rather than the surface. Advancement: You can get the solutions to thinking out of the box, which can be a pleasant surprise for your customers. Sustainable Performance: You can make permanent changes that affect your business positively in the long run. For example, a healthcare company might want to apply the Systems Thinking strategy, which would allow them to see that a long patient wait time, doctor burnout, and insurance reimbursement are all connected. Consequently, they could also use Design Thinking to reconfigure the patient experience, that is, by making it more productive and thus more satisfactory for all parties concerned. Therefore, another classic example of this could be a software company that uses Systems Thinking in order to point out the bottlenecks that exist in its software development process. Then, they might utilize Design Thinking to devise new tools and workflows that would automate the process and bring productivity to the next level. Through Design thinking and System thinking, businesses can become more innovative and ultimately drive long term success !!

Named Entity Recognition systems are used to identify specific entities in the text, such as people, places, or organizations. More often than not, though, these systems are constrained by ambiguity. A word can denote more than one meaning, so that ambiguity can occur when the system is not certain about the proper meaning for a particular context. Strategies for Handling Ambiguity: Contextual Analysis: NER systems take into account the words around a potential ambiguous term to dissect what that term means. Consider the word "Orange", which could refer to either a fruit or a company providing logistics support. If it is surrounded with words like "Warehouse" or "inventory," it is more likely to be identified as the technology company. Gazetteers: They are list of entities along with their types. If a word can be found in a gazetteer, then the system is more likely to identify it as the listed entity. Machine Learning - Advanced NER makes use of machine learning algorithms to learn for large, labeled datasets of text. Machine learning identifies patterns and relationships that will allow the system to make better predictions. Techniques for Improving Accuracy: Quality of Training Data: Quality of the training data is critical. If the noisy and inconsistent data are fed to the system, it will most certainly produce incorrect results. Feature engineering: building informative features can enable the system to have a better appreciation of the context in which a word is being used. As such, it could be essential to include features like whether it is part of speech, whether it has been capitalized, and distance from other entities. Ensemble Methods: The accuracy of a number of multiple NER systems can be enhanced by combining these together. These different systems have their strengths and weaknesses, and by combining them, errors from individual systems are decreased. Domain Knowledge: If the domain is medicine or law, then the addition of domain knowledge helps them to understand the nuances of language. By employing these strategies and techniques, NER systems can become more accurate and reliable in real-world applications.

This is when duplication enters the picture, and we are essentially replicating a project that has proven success in completing similar processes. This is a simple way to identify new projects based on what achieved successful results. There is a catch however with patented process. Challenges and Limitations 1. Not copyable: Patents are protection for certain methods and technologies so you now simply cannot replicate them. 2. The situation only gets more complicated when you also have to account for some degree of customization that may need to be performed, in order use the patented processes with an entirely different context. 3. There is Limited Knowledge of the Patented Process — Replicating patented processes can be more difficult, as detailed knowledge about how exactly it works may not exist. Overcoming Barriers and Limitations 1. Patent Holders: Connect and get close to the patent owners, learn their process inside-out & figure out if you can use that tech; adapt with consent from them. 2. Focus on Principles, not Details: Do retain the principles but never copy and paste process. 3. This entails Innovating within Boundaries: Taking the patented process as a basic guideline to inspire innovation without infringing anyone's patent. Making It Work 1. Develop Relationships: Build solid partnerships with patent-holding companies to make co-operation support easier. 2. Make sure your team knows the science behind our proprietary process and invest in their training. 3. Centralize And Write Down Everything: Maintain a single location of record and write down every change and improvement in the same to avoid legal disputes. When you take into account these approaches, it will allow for replication to be leveraged as a tool for Lean Six Sigma project identification even in the most challenging environments where patented processes are present. If you remain professional and innovative, however….then all of this becomes just a stepping stone.

AI systems are built by combining different components that work together to solve complex problems. Each component has its own specific role, but they often depend on each other to achieve optimal performance. Here's a comparative view i have created for my own learning: Component Role Dependencies Isolation Integration Machine Learning Learns from data to make predictions or decisions other components for data or context Can work independently, but often improved with integration Essential for many AI applications Natural Language Processing Enables computers to understand and process human language ML algorithms, knowledge representation and reasoning Can be used independently, but enhanced with integration Essential for applications like chatbots and language translation Robotics Involves designing and building robots to perform physical tasks Computer vision, ML, planning and scheduling Can function independently, but greatly expanded with integration Essential for physical tasks like manufacturing or exploration Expert Systems Capture and apply human expertise to solve problems Knowledge representation and reasoning, ML Can function independently, but enhanced with integration Useful for tasks requiring domain-specific knowledge Computer Vision Enables computers to understand and interpret visual information ML algorithms Can be used independently, but enhanced with integration Essential for applications like image recognition and object detection Planning and Scheduling Plans sequences of actions to achieve goals Knowledge representation and reasoning, ML Can be used independently, but enhanced with integration Useful for tasks requiring complex planning and coordination Knowledge Representation and Reasoning Represents and manipulates knowledge other components Can be used independently, but essential for many AI applications Essential for storing and manipulating knowledge Highlighted "Can be used independently" to answer the specific question VK asked. A great example of combined AI Solutions integration: A self-driving car uses a combination of computer vision, ML, planning and scheduling, and robotics. Computer vision is used to perceive the environment, ML algorithms are used to make decisions based on the perceived information, planning and scheduling are used to determine the best route, and robotics is used to control the car's movements.

Invention vs. Innovation Invention and innovation are often used interchangeably, but they represent distinct concepts in the business world. Feature Invention Innovation Definition The creation of something entirely new The process of improving or enhancing an existing product, process, or idea Focus Originality and novelty Practical application and value creation Outcome New product, process, or idea Improved product, process, or business model Examples Invention: The light bulb, the telephone, the airplane. These are groundbreaking creations that didn't exist before. Innovation: The smartphone, which transformed the way we communicate, access information, and consume entertainment. It built upon existing technologies like mobile phones and computers. Lean Six Sigma and Invention/Innovation Lean Six Sigma is a methodology focused on eliminating waste and improving efficiency in processes. While it can be a powerful tool for innovation, its application to invention is limited. Innovation: Lean Six Sigma can be effectively used to streamline the development and launch of new products or services. It helps identify and eliminate inefficiencies in the innovation process, leading to faster time-to-market and improved customer satisfaction. Invention: Invention often involves a high degree of experimentation and creativity, which may not align perfectly with the structured approach of Lean Six Sigma. However, the principles of waste reduction and process improvement can be applied to certain aspects of the invention process, such as prototyping and testing. In essence, Lean Six Sigma is better suited for optimizing and improving existing offerings (innovation) rather than creating something entirely new (invention). However, it can provide valuable support in certain stages of the invention process.

The Paradox of Excellence is a curious situation where an organization's relentless pursuit of perfection can actually hinder its progress and overall performance. It's like aiming for the moon but getting lost in the stars. Adverse Effects of the Paradox of Excellence When an organization becomes overly fixated on achieving excellence, it can lead to several negative consequences: Adverse Effect Description Fear of Failure A culture of perfectionism can create a fear of making mistakes, stifling innovation and experimentation. Reduced Risk-Taking To maintain the 'excellent' image, organizations may avoid taking calculated risks, limiting growth opportunities. Burnout The constant pressure to excel can lead to employee burnout and decreased job satisfaction. Loss of Focus An excessive emphasis on perfection can divert attention from core business objectives. Stagnation Over-reliance on past successes can prevent the organization from adapting to changing market conditions. Few Remedial Measures: Embrace a Growth Mindset: Encourage a culture where learning from failures is valued and mistakes are seen as opportunities for improvement. Set Realistic Goals: Establish achievable targets that align with the organization's overall strategy. Prioritize Customer Satisfaction: Focus on delivering value to customers rather than solely pursuing internal metrics of excellence. Empower Employees: Delegate authority and encourage employee ownership of projects to foster innovation. Celebrate Progress: Recognize and reward small wins to maintain motivation and momentum. Continuous Learning: Invest in employee development and training to build a culture of continuous improvement. By implementing these measures, organizations can strike a balance between striving for excellence and maintaining agility and adaptability. Remember, perfection is an unattainable ideal. It's more important to foster a culture of continuous improvement and focus on delivering value to customers.

The illusion of control is a psychological phenomenon where people overestimate their ability to influence or predict outcomes. In simpler terms, it's believing you have more power over events than you actually do. While a certain degree of confidence is essential for leadership, an excessive sense of control can be detrimental to an organization's progress. Negative Impacts on Business Excellence: Impact Area Description Decision Making Overconfidence can lead to hasty decisions without thorough analysis, ignoring potential risks and alternative options. Risk Management Underestimating uncertainties and relying solely on internal capabilities can expose the organization to unforeseen challenges. Innovation Excessive control can stifle creativity and experimentation, hindering the development of new ideas and approaches. Team Dynamics A controlling leadership style can demotivate employees, reduce collaboration, and hinder knowledge sharing. Organizational Learning The tendency to attribute successes to personal efforts and failures to external factors can impede learning from mistakes and adapting to change. Few Remedial Measure a firm can take: Promote a Culture of Questioning: Encourage open dialogue, challenge assumptions, and foster a culture where diverse perspectives are valued. Develop Robust Risk Management Systems: Implement comprehensive processes to identify, assess, and mitigate risks. Foster a Learning Organization: Create an environment that encourages experimentation, failure, and continuous improvement. Empower Employees: Delegate authority, provide autonomy, and build trust to enhance employee engagement and innovation. Enhance Decision-Making Processes: Establish structured decision-making frameworks that involve multiple stakeholders and consider various scenarios. Seek External Perspectives: Bring in external experts or consultants to provide fresh insights and challenge internal assumptions. By recognizing the potential pitfalls of the illusion of control and taking proactive steps to address it, organizations can enhance their decision-making, risk management, and overall performance, ultimately driving business excellence. In essence, while a sense of control is important, it's crucial to balance it with humility, open-mindedness, and a willingness to adapt to changing circumstances.