What Should an Executive Know about Machine Learning? "Unsupervised Learning"

This article is in continuation of my previous articles on Machine Learning & Supervised Learning.

In this post, I am going to share my views on Unsupervised Learning. I have tried to capture the basics here

The basic fact in unsupervised learning is that that the data model performs prediction/actions/inferences by learning from input training data, which in itself does not have any output/results defined. Meaning there is no particular solution/target/output or even error to evaluate an outcome/prediction.

Unsupervised Learning can be further divided into two categories:

Clustering

It means grouping of items into subsets (or cluster) so that the observations & inferences coming from the same clusters are similar. It also implies that the behavior one subset will be different from another subset.

Applications of clustering:

1. You run an e-commerce firm (with large volume of data on customers & their buying patterns) and want to find groups of customers with similar behavior for chronographic watches. Clustering is what you do

2. You are an insurance company and want to segregate group of policyholders with high average claims.

Dimensionality Reduction (DR)

A straightforward method for feature selection and feature extraction, this method reduces the features to process, so that the performance improves and the technique becomes computationally more efficient.

For example, consider a situation where you want to classify buyers of watches from non-buyers of watches based upon their demography. The dimension of this data can be very large (age, education, race, sex etc.). Therefore, if one start applying classification upon all these dimensions, then the system may take very long to process the records. A computationally easier way can be to use DR to find a subset of data that can represent the original data in a non-redundant way; and hence, both cases will lead to the same result.

In addition, it is common experience that projecting higher dimensions data into 2D leads to better visualization of the data set.

Summary:

· In unsupervised learning, we do not know the outcomes

·  It can be of two types: Clustering (grouping) & Dimensionality Reduction (50,000 features become 10)

Hope it helps in your next sales pitch to convey these concepts better!

What Should an Executive Know about Machine Learning? "Supervised Learning"

This post is in continuation of my earlier post on Machine Learning & the three buckets in which it can be understood. You can read it here

In this post, I am going to express my opinion on Supervised Learning.

In supervised learning, the output variable is known, and this output variable is used in the training.

There are three steps for building a supervised model: Building model, Training model & Testing model. Let us understand these three with the help of an example.

Step #1: Building model

Suppose you have joined a coaching class to learn machine learning. Hence, in this case, you become the model.

Step #2: Training model

Your faculty will be teaching you. She will also use various teaching aids during this process. This is the training process. Here, we try to train the model using historical & recent data. The basis of this process is to identify either patterns or dependencies in the data.

Step #3: Testing model

Now is the time when you (model) has to appear for the exam. Obviously, the teacher will not use the same data to test you on which she has trained you, hence, the exam paper will have similar patterns on which you have to respond, but not the same.

Generally, to test the prediction or accuracy of the model, we test it on the untrained data. Usually, the ratio of training data to test data is 70/30.

If your exam score falls below a configured value, then, re-training happens.

Let me correlate it with IT & Business use-cases:

Case #1: IT | Proactive Maintenance of Infrastructure

You take 100,000 tickets from your ITSM tool, build a data model (70,000 tickets) & test it with the rest of the 30,000 tickets. If the accuracy of the data model is >85% (e.g.), then you roll the model for proactive maintenance of your infrastructure (servers, routers etc.).

Case #2: Business | Detect Fraud Transactions of Credit Cards

You gather data on the fraud transactions. Again you split the data into 70:30 ratio. In this case, let’s assume that the model has 75% accuracy (which may be good for rolling it out live). So, this model, when encounters pattern abc in the new transactions, it can predict the probability of fraud in that transaction. Hence, now, you can take necessary actions

Both the cases which I mentioned falls under the category of Classification problems under our initial Supervised Learning. There is another category, called Regression (same thing which you did during your MBA days using SPSS, & hence you all have learnt some machine learning!), which is the second category under Supervised Learning.

Technically speaking, Regression is independent of any framework: machine learning or any classical statistical methods.

Regression refer to a model to predict some numbers, like real values. This is different from Classification, which predicts discrete variables (fraud, mangoes etc.)

Summary: 

● In the supervised learning, we know the outcomes

● The Three steps process: Build, Train, Test

● It can be of two types: Classification (discreet classes) & Regression (real values)

Hope it helps in your next sales pitch to convey these concepts better!

Special thanks to Aditi Aggarwal for helping me with the content & Debapriya for the motivation!

What Should an Executive Know about Machine Learning? "Reinforcement Learning"

The age of advanced data analytics & cognitive computing has enabled the use of complex algorithms across a wide range of business problems, industries, verticals etc.

These advancements have led to usage of innumerable jargon on Machine Learning, data modelling, AI and RPA. Given, we understand that these machine learning models & complex algorithms solve abstract business problems (of course, with probability), improve performance, & create a differentiation positioning for their user. The challenge is that these algorithms, mostly, act like a black box & too much dependency upon them for decision making, without knowing their boundary conditions, can lead to innumerable risks.

So, understanding Machine Learning in simple English can help us all to make a better choice on our dependency upon such algorithms.

Machine Learning basically can be categorized under three buckets:

Reinforcement Learning

This learning enables software or your data model to self-determine the ideal behavior, ceteris paribus (i.e. other things equal), so that the performance of the software or the data model is maximized.

This model works as an interaction between two elements – the learning agent (your software/ data model) and the environment.

The reinforcement signal is sent to the learning agent as an award for every correct action which the learning agent takes. This mechanism repeats and the continuous rewards improves the agents’ learning of its environment. So, if you are thinking it’s more like a trial & error way of learning, then you are right!

Case in point #1

You are devising a data model for a robot, which will help tourists navigate a historical place.

While the robot is in learning stage, every time the robot takes 10 steps on the left, it is hit by a wall. So, now, after multiple iterations, it knows that 10 steps is dangerous. Hence, the next time, instead of taking 10 steps, it will understand that the passage on the left is 4 (for e.g.) steps away.

Case in point #2

Your robot is designed to store/retrieve products for optimized space utilization in a warehouse.

Suppose the warehouse is divided into blocks. Now the robot is able to place 100 items in block A, but for every 101(st) parcels, the entry to that block is jammed. So, the rewards system from the environment (in this case), entry to the blocks, will enhance its learning of the layout of the warehouse.

The above two examples are very crude way to understand how these algorithm works.

Supervised Learning & Unsupervised Learning

These two are very often heard in business meetings, sold in sales pitches of software vendors etc. To make it simpler, I have written a complete article on both these terminologies. You can read it here

In my next post, I shall try to break down these three buckets into simpler sub-buckets, categorizing many more jargon into these three buckets.

Understanding basics of ML, its components & use cases can lead you to be an innovator & a disruptor in your industry. Start small but start learning!

Automation & Artificial Intelligence: Job Encroachment or Job Enrichment (2/2)

In my previous article, I analyzed the level of jobs in an enterprises and proposed the depth of encroachment due to automation on those levels. (You can read it here)

Now, in this post, I shall try to look at what people working on those three levels should do in order to have a steady professional career.

Case 1: Nine to five computer jobs

It’s time for people falling under this category & doing mundane routine tasks to start upskilling themselves. These people can undergo training on new skills (data analytics, digital marketing, or higher formal education etc.). From an ITIL incident management role, an L1 or Service Desk agent may try upskill to a L2 skill level etc.

Case 2: Mid Management Level

Though the knowledge & decision making skills required at this level, keeps people working at this level safe. But since many service providers are toiling hard to make AI available at this level, hence, for the people at this level, the easiest way to stay relevant is to get their senior management buy-in on their skillset (though they too should try to upgrade). Cross-functional thinking, dependency building & taking end-to-end responsibility of execution may be the initial steps to show their skill-sets to buy management trust.

Case 3: Senior Management Level

Though automation & AI has little or no prevalence here, as of today, the people at this level are the ones who make decisions on adoption of automation in their enterprises. Hence, the challenge for the people at this level is to adopt automation responsibly in their enterprises, investing not only in automation, but also in training, educating & re-skilling their employees.

From job encroachment to job enrichment

For the people at the top level, they should inculcate the thought in the minds of their employees that automation is not going to cause job encroachment but job enrichment!

It’s about a thought process in solving problems in which the two approaches (human & machine) work together by complementing each other to use the advantages that each one has, resulting in Jobs Enrichment!

The leaders at the top management should not bogged down by claims of journalists and experts, but rather think logically upon the automation of process where the two approaches can work together, rather than machines substituting humans – This is what I call the enrichment strategy.

In my opinion, we should work to change the mindset, of employees & employers to ensure a different line of thought: automation is not job encroachment, but job augmentation.

This mindset has the power to change the upcoming time!

Automation, RPA & Artificial Intelligence: Job Encroachment or Job Enrichment (1/2)

In one of the sessions where I was explaining automation and the benefits it can bring to enterprises, I was asked an interesting question, “Considering all the hard & soft benefits which you mentioned, any CXO would be trying to get the bots into their environment, is my job SAFE? Or will I shown a pink slip in near future? What should I do?”

This question made me realize that though at the top level - automation, RPA & Artificial Intelligence - is gaining traction & creating buzz for innumerable reasons: cost optimization, improving process efficiencies & enhancing user experience; at the ground level, the level at which probably L0/L1 works, certain basics are missing resulting in fear in the minds of employees.

In this post, I shall share my two cents on effect of automation & job security at different levels in an enterprise & in the next post (the 2/2), what should a common person do.

As mentioned, for the job security part, I shall try to analyze the situation from three perspectives: from a person in a nine to five computer job, a person from a mid-management level & then from a person in a senior management level job. 

Case 1: Nine to five computer jobs

A typical automation (RPA) process works in bottom-to-top manner. It starts with analyzing what normal people do on the ground in a job, and then removes their intervention, step-by-step. It annuls any manual tasks which are either SOP (Standard Operating Procedures) driven or have a standard way of execution. Adding to this notion is that we have many players in the industry which have built bots that can automate standard routine tasks. So, if you feel that you fall under this category, then maybe you should start thinking on adding more knowledge flavors to your job to be safe.

Case 2: Mid Management Level

At mid management level, people perform certain knowledge work which required understanding language, serial decision making, tasks execution based upon that decisions etc.

Though not completely, AI is the disruption at this level. Here, the same players have tried to build data models and search algorithms to enable machines to perform less challenging cognitive work. But, as of today, many tasks which executives perform will get automated.

Case 3: Senior Management Level

The day-to-day tasks defined at this level are highly obscure, which requires analyzing a lot of conditions, contexts, situations, risks & many other unforeseeable parameters. Given the complexity of decision making and impact these decisions have, the amount of jobs encroachment by automation (RPA) & AI, at this level, as of today, is minimal or negligible.

So, for the people falling under case 1 & 2, the situation may seems dire. And that’s logical as well. But from an economy perspective, if we don’t create jobs at the same pace as encroached by automation, the number of jobless people will grow, and the social & psychological repercussions of unemployed population will be dire!

Invest in Innovation, Not in Technology

After creating a lot of noise and making hulk-like promises, innovators & early-adopters are finally beginning to reap the benefits of AI. With the advancements in computing, algorithms and AI data models are becoming more revolutionary. In addition, the world is generating unimaginable quantum of data to power AI.

These advancements have contributed to three times more investment in 2016 – between USD 26 billion to USD 39 billion- than it did three years earlier. So, when CXOs face the daunting decisions on the next wave of investment, a look at where the investment is happening in the market becomes important. Here is an important piece of information should for consideration – Invest in Innovation and Not in Technology.

What does investment in Technology means?

For example, intelligent automation has created a hype in the IT market and companies across verticals are looking for use cases to adopt it. The only challenge, I see, in this approach is that mostly adoption of intelligent automation is seen as an IT initiative, implying that the problem statement it is used to solve is around improving IT efficiency and effectiveness. However, given the current digital disruptions & VUCA environment, almost 25%-30% of overall revenue is going to come from new business lines, & this requires innovation.

Investment in technology is equivalent to investment in solving current point (as opposed to end-to-end) challenges & issues (reducing tickets or automating the resolution process), faced on a day-to-day basis, and these are incremental in nature. For example, you have implemented a RPA solution to automate restarting of failed batch jobs. The RPA solution only does a restart of required services when all its dependencies are met. This RPA solution does not perform a RCA (Root Cause Analysis) and deter the jobs from failing in the first place or looks at the complete process holistically and takes care of end-to-end automation.

Hence, the mandate of CXOs should be to lower the investment in such technological advancements and invest the savings in disruptive business models and innovations

What does investment in Innovation means?

Many people (including CXOs) believe in the power of innovation. The challenge comes when companies, while trying to be innovative, try to define the complete value of innovation. It is evident that the value of innovation is hazy in their minds, leading to scheduling delays, poor investment strategies & non-alignment of leadership.

Investment in innovation means investment in building companies for the next 20-30 years (long-term focus) rather than the next 5-10 years (short-term focus). This involves focus on people, acquisitions, vision etc. For example, PayPal acquired Braintree focusing upon next gen commerce startups. This investment calls for a greater tolerance for short-term risks & failures in order to pursue the longer-term objectives.

To continue our IT processes as example, it is important that investment in intelligent automation is looked upon as a business process improvement. The focus now shifts from automation to elimination, i.e. how can we focus on making robust systems leveraging intelligent automation where the system can itself self-heal & no ticket gets raised.

What I understand is that CXOs are approaching intelligent automation as a source of not just IT productivity but also as a source of innovation in terms of doing things/getting things done & in a faster way. The only challenge to this approach, I feel, is the typical “budget constraint” & investment priorities not reflecting that investment (in innovation) portfolios. Unless, companies take that step, where they focus upon using technology as a fuel for innovation; innovators, early-adopters & new entrants are going to drive them out of business.

4 Best Ways to Educate your Customers about RPA & Intelligent Automation

Nearly any task or process is impacted by RPA & intelligent automation in one way or another across industries today. By leveraging intelligent automation, zero manual intervention of entire processes, routine and workflows - data collection, analysis, understanding context & finally making decisions- can be achieved.

Until recent times, robotics had more applications in the primary sector- automating & eliminating the human involvement from the production value chain. Now, tertiary sector esp. the financial services industry (also High Tech/Telecom, see below image) has majorly started applying use-case driven RPAs & Intelligent Automation to automate and eliminate low (/no) value-adding activities performed by humans. There is a complete article on the state of Machine Learning & AI, 2017 by McKinsey. You can read it here

Apart from the industries mentioned above, the adoption of RPA & intelligent automation is still in nascent stages in many other industries. The chief reason as per my opinion is that there is a lot of variation in the way RPA & intelligent automation is positioned or marketed. Vendors & Start-ups do not talk much about the risk, compliance, governance, metering etc. of intelligent automation. Therefore, the overall eco-system of how/what/by when/disaster recovery of using RPAs or intelligent automation is not visualized coherently by the CXOs & the decision-makers, leading to low adoption rate of these new technologies.

In my opinion, the four best ways in which start-ups & vendors can build Trust & Authenticity are:

1. Analyst Speak:

There are many independent third party analyst in the IT industry who benchmark the various products in the field of RPA & Intelligent Automation. These analysts compare & categorize the various service providers based upon certain parameters and dimensions, and then, normalize the result to create a framework where potential clients can compare one provider with another.

On a side note, if you are a CXO who is exploring the adoption of RPA & intelligent automation in your enterprise, the best way to proceed is to seek advice from this community and get your uncertainties answered, possibilities of risk mitigation & feasible intelligent automation solution finalized for deployment.

So if you are looking to educate your customers & the market, educate the analyst community first.

2. Seminars, Symposium, Events

This point is for the service-providers. Attend all the major events, get a booth and prepare yourself to show the best version of your product to the people walking in. Try to capture their response, reaction and feedback. This, in my opinion, is one of the best strategies for lead generation for product.

3. Get your hands dirty: Perform Pilots & Proof of Concepts

Simple business rule is- New Customer Acquisition is tougher (both monetary & effort wise) than Existing Customer Retention. Therefore, you can focus upon adding the flavor of RPA & intelligent automation into their already existing managed services account.

Then, to get the initial buy-in from customers, communication & formulation of problem statement & relevant solution is very necessary. The easiest way to start is by doing a proof of concept or pilot for a problem statement, which is simple and can be easily scaled, but at the same time, is a major bottleneck to customers. In addition, what needs to be mutually agreed upon is the success criteria for that proposal (pilot or POC).

The advantage of such proposal is that you will be using the success criteria of that proposal to make the customers comfortable.

Once the POC or pilot exercise becomes successful, deploy the RPA or intelligent automation solution to the complete enterprise.

4. Get Customer Testimonial

Word of mouth has always been and shall be the best way to create Trust & Authenticity in a B2B environment. So, the more you can get your customers to share a dais with you, release a video etc. the better stand you have in the market to gain new customers.

Sweating today to get visibility & becoming an industry leader should be the approach to educate the market and customers. A positive market wave favoring your solutions can give you a first mover’s advantage in this field.

Adopting Intelligent Automation: From Stakeholder Management to Change Management

Changes occur in businesses in two distinct ways: External (Merger & Acquisitions, regulatory compliance etc.) & Internal (organizational restructuring, new leadership, adopting new technology etc.). An enterprise level change has always been difficult and innumerable researches show that they continue to fail. In the case of adoption of intelligent automation, the change involves so much stress (jobs uncertainty for employees) & risks (ROIs) that it becomes very important to focus upon stakeholders & the change management processes.

To adopt intelligent automation in their IT estate, business leaders have to agree upon the direction and the end objectives to be achieved. My opinion is that it’s a four way process to adopt intelligent automation:

Building consensus among stakeholders

Digital, intelligent automation, analytics and the changing nature of workforce have created myriad opportunities and challenges for the stakeholders. For the same set of data, various stakeholders may draw parallel, non-concurrent conclusions which can indefinitely pause the implementation plan of intelligent automation into their environment. To illustrate, the business managers, looking at their current inflow of password resets tickets, may go for some intelligent automation solution. But then, the end users, who are going to use the system may not align with the intelligent automation and may continue to use the earlier way of getting password resets done (tapping the shoulders of agents or sending emails etc.). So, even though there is intelligent automation in place, there is no reduction in the volume of tickets and hence no ROI for business to justify investment in further intelligent automation.

So, considering our case of adopting intelligent automation, building consensus among the various stakeholders becomes the first & the foremost step towards ensuring smooth adaptation and transition for intelligent automation.

There are many frameworks & processes on stakeholder management, but I personally go with Segmenting & Positioning (no Targeting, as everyone is a target for IT); segmenting the stakeholders and passing on the customized message which intelligent automation is going to deliver it to them.

To do this, it becomes important that you prioritize the segments which are going to be the early touch points for intelligent automation & narrate them the ease factor (“hard” & “soft” benefits) which it will bring in their way of doing jobs. Probably for the CIO: cost effectiveness, IT managers & business managers: reduced TAT & improved efficiencies & the end users: simpler processes are some e.g. RACI matrix, too, can come handy in this process.

Involvement of IT community

Believe it or not, a lot of risk is mitigated if the IT community is engaged from the beginning. Specifically, the innovators and early adopters. This group can speed-up the adoption rate of intelligent automation in at least two ways:

a.     Help with the beta test of intelligent automation use-cases

b.    Act as word of mouth for intelligent automation benefits

These are the people who require least effort to understand that intelligent automation is really valuable to them and to the enterprise as whole. In my opinion, if this group is convinced, a lot of push-backs from this complete group gets controlled.

Communication is the key: Counter Culture push-backs

Now, you have the go-ahead from the business & IT leaders, innovators & early adapters, but then the major chunk of people who will be using the intelligent automation have still not entered the picture: end-users & agents. They may still be skeptic about intelligent automation and the way in which it can make lives easier for them.

To roll-out intelligent automation into the enterprise effectively, we should communicate the “soft” changes that needs to accompany the “hard” changes in the IT estate to both these group. For e.g. proper communication explaining the rationale to go for intelligent automation in first place can bring a lot of acceptance without spending much efforts. In addition, training, workshops etc. can be used to make them understand the benefits to-be derived from intelligent automation.

Tessa Basford & Bill Schaninger from Mckinsey have captured the complete mindset & behavior change management here.

Support from the governance

Intelligent automation may take some time to show benefits or the people using it may find it a bit challenging to turn over a new leaf. Many processes may be re-engineered for intelligent automation, which can emit a lot of distress signal in the organization; such testing water situations demands patience from management and a sturdy focus on the woods & not the trees.

For many big enterprises, hiring an external consultant may be a good option as stakeholders within these enterprises have been used to doing things in a certain way. A third party angle to look at the current processes can help in re-designing it so that the implementation steps becomes easier & faster  

Involvement, communication, change management and ultimately alignment amongst all the stakeholders to adopt intelligent automation is what is going to keep the health check of enterprises in place in the future. Adopt it today to reap benefits tomorrow.