Product Management

Boring Chat Bots are Coming – #ChatBots #Future

Disclaimer: Before you ask…as always all opinions are mine/personal. This article is my take on how Chat Bots would evolve in the near future

Dear Product  Managers,

Heard this? “bots are the next big thing” and “Chat bots will solve your communication needs” and especially “your apps will become redundant”  and let me not forget “your business is dead if you do not have a Chat bot”

Yes bots are great and you will probably see more of them around and we need to prepare , especially you ..the  next gen Product Manager trying to build the best awesome personalized human like bot .

So here is a list of my complaints/rants  against bots and how many Bot scenarios are sometimes potrayed


Chat Bots to replace everything

God knows what all can Chat bots replace or even God does not know what Chat bots cannot replace.

One of the most common examples I have heard is “Pizza ordering bot” , Like a full scale pizza ordering chat bot

Heaven’s it is a bad idea to even think of.

This is how many scenarios play out in our heads(Would seem normal to some and unnecessarily drawn out to the latter, I am sorry)


Me: Hey I need Pizza

Bot: Awesome, what type? Your regular?

Me: Naah, want to experiment. Suggest me some veg options

Bot : We just launched Awesome Veggie supreme  pizza. It has pineapple, mushrooms and tomatoes

Me: Anything with broccoli

Bot : Sure , How about Broccoli surprise ?

Me: Swell . Broccoli surprise would do

Bot: What size

Me: Medium , oh and please add extra cheese

Bot : Cool! Anything to drink? We have coke, pepsi and Mirinda

Me: Coke 1 ltr please

Bot : Great . Total is 1 gazzilion Rs.

Me: ok

Bot: Your address is same as before

Me : yes

Bot: Great , order received. Will keep you posted !

Me: Thanks


Sounds decent enough, like talking to a normal person at the counter, doesn’t it?

Boring Chat bot


Typing  Typing Typing and Typing

You basically ordered Broccoli pizza and a coke and had to do a long drawn conversation. What if you wanted to replace pineapple in one pizza with broccoli ?

What if you got in a mood for No extra cheese?  And god forbid if you were ordering for a group.

It takes me 30 minutes trying to get a group to agree to a pizza order, I don’t want to have that same conversation with the Bot. This negotiation does not work.

It’s a BOT

What’s all this crap about having a conversation …”Natural language” . Yes giving a personality to a Chat Bot works, but why exactly are people trying to impress a Bot?

What is this need for ” Hey ” “Can you please” .

Not people’s fault, it seems that is how many advertisements around bots expect us to behave.

These are dumb machines, can’t even pass a Turing test (1 did I guess) and we are trying to be Nice to them?

Think about all the human potential wasted trying to be Nice to Bots

Do you go to google and type , “please tell me the nearest pizza places since I am hungry”

Someday, bots will feel bad but that day is not today. Last time I checked, my car does not give better mileage if I talk nicely to it.

You could argue that this can be further simplified , heck you could, and that is exactly the points.


Many Chat Bots will and should become dumb and stop hiding that they are bots. You would either see many new bots that do just 1 job (while the app can do 1000–Eg  Dominos bot–you say Pizza and you are done) or drop the pretense and talk to you like they should…Like an IVRS system. Yes the ones that everyone hates , but are still very prevalent

Eg–something like this

Me: Pizza

Bot: Please select which one

  1. Your previous order (Broccoli blast)
  2. Veggi paneer
  3. Spices Maximus
  4. See Non veg also

Me: 2

Bot: Please select

  1. Complete order
  2. Order more
  3. Edit order

Me: 1

Bot: Great ! The total is 1 gazzilion dollars


This Bot is BORING , this bot is efficient.

It is like an IVRS. Whenever you think about bot scenario, think about how an IVRS system  works, would you do that on an IVRS?

IVRS: Please speak your name after the beep

Me: Hello IVRS, My Name is Madhur

IVRS: You are stupid Madhur


But there is ray of hope.

Bots platforms have started and soon will go gung-ho on giving a rich interface to bots within chat .

So in the pizza example, you could open the entire menu inside the chat and just Tap on the one you want

You would make a selection on the first chat message and add cheese on the next, again by a simple click/tap

This sounds very familiar. Where have I seen it? THIS IS AN APP

Now you see? Bots are just screens of your App, starting with few functionalities and possibly finally your entire app,just another UI surface .

Do not force fit your entire app into  a chat interface, it will just give you PR, not happy customers.

Would you browse amazon using an amazon chat bot ?


Do Bots fit nowhere:

They do, they are awesome, especially when you do not initiate the conversations, but an external event does. These are the best use cases IMO


Finance Bot : Pssst: ICICI just crossed 1200 Rs. Wanna sell? I recommend sell , all of it. Press Y

Me: Y100

Finance Bot : Kewl! Sold for 1210, you made 10,000 rs profit…you rich spoiled brat.


Your apps are not dying, you can survive without BOTS,it is OK.

There are a LOT of advances made using AI and computers can figure out your stated needs : Eg:”Veggi delight pizza enough for 4 people” can directly order aVeggiei pizza large size.

But you need to know what you really want

So yes, dear product managers and users: your next wave of bots will be boring , they will be as close to chatting with a person as IVRS is to talking to a human. Even Siri and Cortana wear off the novelty in a day(OMG, it can tell me a joke)

Things will improve,but you need to go past boring bots first( we are still dealing with IVRS, remember)


A new ERA begins…The Boring Bot Era


Sincerely Yours

Broccoli pizza fan #LeaveMyPizzaAlone

I send infrequent newsletters. You can signup for them below.
* indicates required
What are you interested in
Product Management

How I made my first android app over a weekend- A lazy sloth way

The story of AMG Math App

So I made an app…big deal you….

Though currently I am a Product management professional and am not required to code, I have been a coder for a major part of my professional life and that itch to create something of my own always remained.

There are various reasons people want to build their own apps: Creating their brand portfolio if they are designers/devs , executing a business idea or as in my case to learn and understand the dynamics a bit more.

But somehow, despite my coding background (I was good at it and am not humble enough to deny it) I would not pick up learning the basics for a long time. Let me list down the reasons for not doing it

  1. I am as lazy as a sloth
  2. I procrastinate–basically also means lazy
  3. I thought re-learning coding again would be too much hassle –in simple words: I was lazy
  4. and did i mention that I am a very lazy guy?

So here are the things I had tried to break my procrastination/laziness

  • Bought a Course on ionic app development thinking that my love for money and my middle-class upbringing that teaches me to extract maximum value + more if you have paid money  trumps my laziness…news: it does not. Why Ionic?: Ionic is a hybrid platform that lets you create apps once that can be published to all platforms(android+IOS)…you can see why i picked it. Who will learn 2 platforms
  • Chucked that and decided to just go to udacity free courses and followed the tutorials (Thank you google) …Got damn bored after a few lectures. They were too slow for my taste and I picked up one 2 videos every few days, implying i would forget half of the previous lectures
  • Tried to read through the dev guides…Nop

It all changed when I got talking to a friend who also wanted to learn coding. We decided to make a pact to release 1 app ,even if it was a hello world app by end of the month.We went for android.

Lo behold , suddenly I was back at the udacity course page and ravaging through the basic course like a mad man on a mission. I was watching videos at 2x speeds to save time…(Tip: Youtube feature to speed up videos is google’s greatest gift to humanity)

Life lesson: Challenge motivates me.


I love sloths

But the boredom again set in…i decided to then do something else. I started even before I was ready and started optimizing on everything…whenever I was stuck,I asked

“What would a sloth do..

So this is how it went down

  • I looked around to see what kind of simplistic app I would like to make and decided to go for something close to my heart…Fun +education. How about a simple Arithmetic game. I would give people  set of 4 numbers and a target. They would need to use all the numbers in an arithmetic expression  so that it solves to the target.
  • I decided to forgo perfection . I went for 1 screen app mainly because I had not yet learned how to make multi-screen apps. Even the help section was bundled into the same page
  • I learned the very basic of formatting (mainly linear layout) with 0 designing etc put in
  • I realized that to make a huge question bank, I would need to do a lot of work. Even making 1000 questions means I would need to type a hell lot. In Microsoft I was a Product Manager for health and fitness app ,and I clearly remember that in “exercise details” feature that I had done , there was a csv file with some small typo errors or special characters that needed to be removed for the code to read them properly . The insane amount of manual effort required to clean up a simple csv file still gives me nightmares (Not literally). There was no way in hell I was going to create a question bank for a fun app. So I automated it and here is the algo
    • Every digit is randomly picked from the range specified
    • The operators to be used are also randomly picked but they also have a probability of being picked . EG: * is far less probable than + and –
    • The 4 digits are stitched to the operators and evaluated using the same function that evaluates the answers user inputs, to calculate the target .
    • This entire expression becomes 1 question
    • An array of question is generated and then picked at random and presented to the player
    • The process repeats every time you start a new game. So the probability of seeing the exact same question is rare
  • The system had no scoring, no gamification , no analytics . Just a 1 screen app made by a product manager who wanted to see what the hoopla is all about
  • I broke down the entire process to very very small steps. Baby steps do help if you are a sloth
  • I got stuck , learned specifically that piece and moved on.

The entire coding , most of the learning , and publishing the app was done over a weekend. Yup, just 1 sleepless, sitting in my home office (yeah I am showing off  ) continuously weekend.

Being a Sloth is not bad afterall

And AMG Math App was released.It was initially called “Can you Solve this”

First Version of AMG MATH App
First Version of AMG MATH App

Initially I thought my wife would be upset that I was in front of my mac(Yes I have a mac..gosh I am good at brand placement..appple pay me money) for the entire weekend …surprisingly she was not. She was happy that I was least not being a sloth , that’s progress

Since every article needs to have “learnings” , here are my learnings

  1. Understand what motivates you
  2. It is better to learn whatever minimal you can learn to start and just implement it . Don’t wait to be a guru. Learn on the way
  3. Doing everything on your own gives you great insights into how the process works. I spent un-godly amount of hours to correct simple logos .. There are so many sizes you need to take care of and the android process does not make it easier. Every resolution has to be made
  4. Having something out there can be a great motivator. Since launching the app I now keep on doing simple upgrades
    1. Done 3 App upgrades
    2. Added gamification
    3. Added analytics
    4. Finally added more than 1 screen
    5. Soon planning to have animations etc
  5. Google android app store analytics are super useless for new apps. Always have google analytics and see real time what’s going on
  6. Whenever stuck, think “What would a sloth do” 
  7. Automate automate automate…find shortcuts, it is not a bad thing

Let me be very clear , it is not a great app in itself. Too simplistic, something that most devs would make in a day…But I really really do not care about what the app is, I just wanted to go through the entire process from app submission, making it prod ready , seeing basic analytics, understanding discovery etc , also i did not really care about what people would think about the first version of the app 🙂

Coming up soon: How I optimized the AMG math app ..the sloth way

This is how the AMG math game app gameplay screen looks now.

Would love to receive some comments/ reviews on the app. What do you want to see next..remember I am learning whatever I need 😀

How the AMG math game app looks now
How the AMG math game app looks now


I send infrequent newsletters. You can signup for them below.
* indicates required
What are you interested in
Haiku Product Management

Product Management Haiku #5 : Trump Style : Use case without users

I have a use case,
It is a good use case ,Huge,
Zero users though

There are a lot of interesting use cases you can solve. Stop any person in your company and ask them , do you have a use case to solve? The answer invariably would be Yes(If it is NO often , you need to think hard about how much relevant your own product is for your own company)

The question is not just what use case you are solving, but how many users actually are there? And are they even your users?

I send infrequent newsletters. You can signup for them below.
* indicates required
What are you interested in
Haiku Product Management

Product Management Haiku #3: Tech debt

If unrepaid will haunt
the tech debt has to be paid
circle of PM life

Moral of the story: Pay your tech debts. Features can wait sometimes, you can have released where you have nothing to show. Absolutely 0 visual progress. Your product update would be ” Fixed some bugs” or “better logging”…do it and thank me later

I send infrequent newsletters. You can signup for them below.
* indicates required
What are you interested in
Haiku Product Management

Product Management Haiku #2: Feature Prioritization

Pitchfork in hand I stand
Cut the feature cut it out
ahh that’s my real life

I send infrequent newsletters. You can signup for them below.
* indicates required
What are you interested in
Haiku Product Management

Product Management Haiku #1: The Data is God

The design is sleek
The Product is awesome but
data does not admit


I send infrequent newsletters. You can signup for them below.
* indicates required
What are you interested in
Product Management

Linkedin Recruiter – Listen up , here is why your email is unasnwered

This is for a Linkedin Recruiter 

I feel like ranting a bit 🙂

Let me start by saying that I got both my current and last job via LinkedIn/unsolicited emails and both turned out to be very rewarding , so in no way am I against recruiters reaching out to potential recruits on any platform.You should, it is great for both you and the employee

My issue is with the sheer unprofessionalism of almost 90% of “amazing offer” emails that I receive from Linkedin recruiter s. Here are a few types of “recruitment” emails that kind of bug me.

Linkedin Recruiter Types

1. The Amway :

I have an amazing leadership position opening in the leading Indian e-commerce company in Bangalore. Send me your resume ASAP

Really? Do you think you are hiding the identity of the company  , and why this secrecy? You already gave me enough hints about the company, so why hide the name? Funny part is, every single time I reply “Name?” , I get the answer in a few minutes.

This is like an Amway ad ” dude amazing business opportunity, just come to the seminar and your would know…I won’t tell you what it is”

Dear Linkedin Recruiter :

Unless this is for  FBI or CIA , asking people if they are interested without even telling them which company it is, is kind of disrespectful 

2. I am God –You Mortal

I have amazing openings for you, send me CV, Expected CTC…… of your first born etc etc

I like how you did not even do the pretense of “hey are you looking out”

Linkedin Recruiter : NO

3. Don’t want this apple, how about this banana? Or the 100 other fruits we have

Hi, blah blah blah….we want to interview for position XYZ . Not interested? We can also interview you for “A,B,C,D,D………entire catalog of jobs I have with 20 companies

Send me……CV. CTC, …..DOB of your pet…etc etc

Oh come on, you started off well but then why oh why did you send me pretty much your entire catalog of jobs. I know people use templates , but don’t make it so obvious.

If you are not sure, just say let us find a good fit.

Dear Linkedin Recruiter :

Focus , It helps. Do not tell the menu before someone has even sat down

4. I use a template and you know it

Actual Email from a Linkedin Recruiter:
Hi Madhur, hope all’s well at your end in Bengaluru 🙂
I believe you’re currently working with Directi… Based on your LinkedIn profile, I think you would be an ideal match for this position
I like how you not only used a template but even made sure I know it by Bolding the words you replaced. Loved the smiley at the end so that it feels personal.
Please oh please, I really do not want to know My Name and My Company and My location. If anything, BOLD your company name .
 Dear Linkedin Recruiter,
Use templates but don’t rub it into the candidate’s face.
PS: BCC is also super annoying

5. I am all about research

Actual Line from a Linkedin Recruiter:

I’m all about doing my research before ever contacting for a position I’m tasked with filling

This email is fine once. But when you receive the same email from 4 different people multiple times, it gets on your nerves

Dear Linkedin Recruiter , divide your leads

Divide your leads

6. I never changed my email

From: [email protected]


Yup, your email needs to be professional . Please stop using the first email ID you ever created to recruit

Dear loverboy,

You really are not a recruiter but, just a freelancer ? Too new to even have a domain email ID?

7. I do not read


Would you be interested in joining as fresher Software Engineer at this BPO at a salary that you drew 8 years ago?

I have been a dev for 6 years and am pretty proud of it. I still love code but sadly it is no longer my area of expertise. This is visible not just from my higher education choice (Management not MS) but even my current role.
Just because I mentioned that in 2008 I was a sharepoint expert does not imply I am still interested in the exact same job in 2016. Please READ the profile of the person . It wastes your time and mine

Dear “I do not read recruiter” ,


How to write a good Hiring Email ?

Over the years I have received many hiring emails and based on my personal(only mine) experience I can say the template of a good hiring email should be

Hi<my name>

<a little bit about who you are and what do you do–Important>

<If possible, how you found me and why are you reaching out>

<a little bit about the role and department –The MOST important thing.Eg Do not tell me Product Head role, tell me which product and what does it do, what are the plans if any>

<Ask me if I am interested and suggest next steps>

DO NOT ask me for my updated CV, my notice period, my expected CTC etc etc. Let us first talk before jumping the gun. Most conversations end by simple “sorry I am not looking out”.

Basic points to keep in mind

  1. Drop the pretense: you are a recruiter and the other person is a potential recruit , both need each other. Do not stress on how you know my work location and do not mention how you are impressed with the work i am doing in <My current company> (I got this email 1 month after joining a new company . So someone outside my company  was impressed even before I was? )
  2. Do not assume everyone is looking out and would jump at the mere thought of someone hiring them. So before asking for CV, CTC etc etc atleast initiate a call to see if the person in interested
  3. Context Matters: Always set the context. Think of it like a face to face conversation. The basic info such as your own intro and the role info needs to be upfront.
  4. Be specific: Do a basic research on the kind of job the candidate does before bombarding him/her with random openings
  5. Use templates, just do not shove it into the face: Yes you probably will be and should be using a well-crafted template, but do not make it too obvious
  6. Be like a normal human being talking to another human being

You have awesome jobs to offer to deserving candidates, let your emails shine the way you expect the candidates  to in the interview 🙂

and happy Recruiting



I send infrequent newsletters. You can signup for them below.
* indicates required
What are you interested in
Product Management

If Not Surge pricing , what else could Uber Do?

I had earlier written an article on why I think Uber’s surge pricing is detrimental to Uber’s growth . My major point was that with surge pricing as insane as 7X times normal fare, I as a customer would never want to be in a position where I do not have a counter leverage to Uber.

By having surge pricing Uber is actually making space for cheaper options as counter-leverage in the market.

Let’s be clear, Uber is not making any money(it is burning cash) ,and will possibly not make any money unless it is almost the market leader and consumers do not have much counter leverage , enabling uber to move away from it’s current predatory pricing towards a more sustainable pricing model.

But if not dynamic  surge pricing , what could be a possible pricing model assuming you have to have some kind of surge pricing

  1. Time Slab Based Surge Pricing: Prices can be jacked up based on time slabs. Eg: Night rates , office hours . Something that is predictable and transparent.This would become much similar to generic Taxi model (which actually earns) but may not fit in Uber’s current approach.
  2. Surge Protection Model: The initial assumption of dynamic surge pricing was that surge pricing would enable more drivers to come on the road . But a recent Washington Post article suggests that Surge pricing only leads to redistribution of drivers and not necessarily bringing in more drivers. Ie: People in areas with less demand start seeing much longer wait times, something that beats the whole concept of “Everyone’s private driver”.  If the article is correct , we can assume that the primary objective is not  being met

The other way to make this a success would be to have much more users use uber .Having users prefer Uber over any other mode of transportation.And that can be done Via Surge protection Model. this is how it would work

Surge Protection Model

Now this is like differential pricing , where certain section of users would be prevented from surge pricing while others would still bear the burnt. Think of it as surge insurance. For gaining entrance to the elite group you need the first hook/investment this could be 1 Ride on surge prices . Just 1 ride

So anyone who takes 1 ride in Ubers Surged prices , say 5X times normal would be protected from surge pricing for say next 1 month. This user pays normal fares while everyone else pays the dynamic price

How to keep the user hooked

A simple rule where surge protection would continue only if the user avails 1 uber ride a week (numbers can be worked out) . We could potentially extend surge pricing protection for he user indefinitely until the user keeps using Uber at-least once a week.


  • A user knows that by taking just 1 high priced ride, he/she can potentially be unburdened of thinking of surge pricing for a month(indefinite , based on scheme). The thought of future gains may offset the pain user feels now, improving conversion rate of people taking rides at surge pricing
  • Whenever such a user will have an option between various players, user would tend to use Uber to keep surge protection alive thus increasing the number of returning users
  • Human beings do get a kick out of knowing that they are getting a much better deal than others. When a user sees that they are travelling at say 10$ Km while others are being charged 7X the amount , their ride would feel much more affordable and happier
  • This could even enable Uber to jack up it’s base prices to bring them in-tune with actual costs
  • Higher the surge prices go, more loyal the customers in the surge protection program become . In a weird way , surge pricing may become a marketing tool.

Obviously numbers can be worked out, and the idea can be a little offbeat but no harm in thinking about it.



I send infrequent newsletters. You can signup for them below.
* indicates required
What are you interested in
Product Management

Could Uber’s Surge Pricing become Uber’s failing

Uber , Ola, Taxi for sure and all other taxi Apps are truly disruptive to the taxi business and have made life easier for many people.

Now a taxi is just a touch away and you pretty much know how far away it is and when it will reach . Like scores of customers , I am sold on this idea and am a frequent user of almost all these apps.

So lets talk about what these Apps finally want to be. For the sake of this discussion let me talk about Uber . Uber’s tagline was/is Everyone’s Private Driver.

Sounds good right? Yup it does. Uber wants to be your driver wherever you want to go , whenever you want to go , as if you own the vehicle without actually owning it.

So what does Uber need to do to achieve this objective

  1. Available all the time: This can be solved by having more and more drivers sign up . Uber , Ola etc are aggressively recruiting drivers and this should be a solvable problem. The next step in this should be better demand forecasting and having drivers on standby in not so busy places. This currently is being done but seems to be mostly driver driven rather than company driven.
  2. Affordable: It should be cheaper to use Uber than Own a vehicle. The trick here is that it needs to be perceived as cheaper , not just be cheaper. People do not take out a pen and paper and calculate cost of ownership of a vehicle , they simply would calculate Per Km Cost and would be OK if Uber is slightly expensive in this regard. This would need considerable marketing and customer education
  3. Reliable: There are 2 types of reliability here
    • Reliability of availability : I should know that I would get it , irrespective of where I am , it could be a super busy downtown or a picnic spot at the end of the city. I should be able to rely on it. If I cannot, I will always have the need to buy my own vehicle or maybe use other modes. I cannot plan a trip with uber with a possibility of not getting one on my return journey. This again is a solvable problem by having more and more drivers and using better placement and routing algorithms
    • Reliability of Cost: This is something which people do not really talk about much . Why do you think people in various Indian cities do not like Autos and use them as primary means of commute? Why do people use autos because they cannot afford a car yet ? Apart from all the above mentioned points , there is an added issue of auto drivers literally looting customers by jacking up the price at will . This is something that Uber and Ola have inherited from auto drivers and call it surge pricing(price increases based on demand) .     You never know what price these companies would quote, and you cannot even bargain. Just because it is in a swanky app and has a cool name and is done automatically by an algorithm , does not really change what it really is: Extracting as much money as possible.

Now I am not debating if it is right or wrong, it makes business sense to charge as high as you can especially since all these companies are simply burning cash ,but my point here is about “will it help Uber achieve its ultimate objective or being everyone’s personal driver”?

To demonstrate the point , let me bring in a personal example. I wanted to use uber to go till a bus stand instead of an auto , simply because it was too easy to get it. Uber suggested that since it was peak hour, I would be charged 1.2X times for a distance of 5 Kms. I still decided to use it though. The moment I reached the bus stand , just out of curiosity I rechecked the price and it was 5.0X times normal. This was in the span of 15 minutes and a distance of 5 Kms. If I had been quoted this rate, I would have definitely not taken uber . Looking  at such variation, my instant reaction was , I cannot always depend on Uber and possibly missed the 5.0X window by a few minutes. It is like a game.

Surge pricing is Ok if it is not too high. Eg: when you quote using the “Bata Pricing” of 1.XX times(say 1.98 X) , people still perceive it as 1 ish . It seems Ok and can be tolerated but when it crosses integers and is in the range of 2X 3X and even 5X , I get a feeling that I possibly cannot depend on it. Your own car or an auto running on meter (autos are legally not allowed surge pricing , except for night rates) would not surge prices.

Basically , I cannot depend on uber during the times I might need it the most and that exactly is the problem

This simply means that people will always keep a second option open, an alternate mode of transportation , thus defeating the whole purpose of Uber’s tagline . I have all the apps on my phone in addition to various numbers of local taxi players not signed up with uber.

Uber cannot make a profit until it becomes a dominant player ,and it cannot become a dominant player until there are enough cheaper options available in the market and having things like surge pricing might actually keep the cheaper options lingering in market longer.



I send infrequent newsletters. You can signup for them below.
* indicates required
What are you interested in