2. MoveBot
A.I. based full solution allowing
autonomous picking and placing for
ecommerce and logistics warehouses
A software layer for industrial
robot arms to pick and place
random objects
Target market Interviews
B2B: Warehouse Automation 114
3. PhD Mech Eng.
MS École Centrale
eboigne@stanford.edu
MBA GSB
BCG/Uber/IIT
ssankhla@stanford.edu
Visiting Prof.
PhD Eng. MIT
whitebl@stanford.edu
MBA GSB
E.Eng.UT Houston
rzhong0@stanford.edu
MSx GSB
MS CMU-RI
puneetp@stanford.edu
Controls Optimal Control RoboticsConsulting Logistics
Emeric
Boigne
Ray
Zhong
Surabhi SankhlaBrian White Puneet Puri
Team
5. First canvas
Key Partners
Robot arm
manufacturers
Cost Structure
Software development cost
Labor cost
Key Activities
Software R&D
Sales & BD
Aftersale Support
Domain knowledge
Key Resources
Software Simulation
Tool
Value proposition
Enable multiple use of
existing robotic arms
Cost saving
Efficiency increasing
Customer segments
E-commerce player:
such as Amazon and
Jet.com
WEEK
INTERVIEWS
0
Our Journey
0
Revenue Streams
Subscription/ licensing revenue / profit sharing
Customer
relationships
With e-commerce
players
Channels
Direct Sales
6. Nuclear power
Manufacturing
Waste management
Mining
Oil and gas
Agriculture
Medical
Assistance
Food processing
Warehouse & Logistics
Start of Journey
Warehouse
& Logistics
Warehouse
Picker Station
Manufacturing
Waste management
Warehouse & Logistics
WEEK
INTERVIEWS
0 1
0
Our Journey
12
Amazon interviews:
Automate picker station in
warehouse (FC)
High pull from the industry
Our in-house consultant:
‘Guys, narrow down industry!’
Criteria: Size vs Feasibility
Business Model Canvas Customer Segments
Customer feedback:
Since main income is
from govt., willingness to
adopt is low
7. Challenges faced by warehouses
1
Revenue loss due to packaging mistakes2
Labor Shortage during peak capacity3
Repetitive actions leads to injuries
Picker Station
WEEK
INTERVIEWS
0 1 2
0 12
Our Journey
Cost: ~$4.5 Million/ warehouse
(based on Amazon internal data)
23
=
8. A.I software for robot arms to pick and place random
objects
WEEK
INTERVIEWS
0 1 2
0 12
Our Journey
23
Warehouse Software
Movebot’s A.ICloud
Robot Control
Value proposition
1 Cost reduction with fewer errors
2 Ability to handle peak capacity
3 Safer work environment
Business Model Canvas Value Proposition
More Deployment → More Data→ Better A.I.
Additional value created : $2m
$2.5m solution for 20 stations pays for itself < 1 yr ROI
based on Amazon internal budget
9. Our first big “aha” moment!!
$3.5B Revenue
Top 5 robot Manf.
10. “The challenge is finding where to pick ..
.. the robot can do the rest”
“You are onto a big industrial problem”
Application Engineer, KUKA
“7 out of 10 sales people say bin picking biggest problem”
Business Dev. Manager, KUKA
“Can we have an exclusive partnership with you?”
Business Dev. Manager, KUKA
WEEK
INTERVIEWS
0 1 2 3
0 12 23
Our Journey
35
11. MVP - Proof of Concept
Given 3D image of any random object the deep neural network predicts where to grasp
trained 6.7m grasps from Dex-net 2.0 dataset
WEEK
INTERVIEWS
0 1 2 3 4
0 12 23 35
Our Journey
51
Grasp Points
Using Dex-net 2.0
12. Automation industry is complex!, how can we
reach?Oh Damn!
Tomorrow is last
day of expo you
wanna go??
Sure..lets book tickets
and fly out tomorrow.
Brian
Puneet
Here is the
brochure I just
designed.
Emeric
Automation expo
in 2 weeks in L.A
you want to go?
Rafi
WEEK
INTERVIEWS
0 1 2 3 4
0 12 23 35
Our Journey
51
2 weeks
later...
14. Strong validation: from 1 to 7 arm manufacturers
“If you build this solution people will line up to buy it ”
Epson Robotics Sales
“Don’t mind my language but it is dope”
Epson Application Eng.
“I have a customer right now who is looking for a solution”
Denso Robotics Regional Sales Head
1
Learnings
WEEK
INTERVIEWS
0 1 2 3 4 5
0 12 23 35 51 61
Our Journey
114
10
15. Learnings
2
WEEK
INTERVIEWS
0 1 2 3 4 5
0 12 23 35 51 61
Our Journey
Realized the complexity of the ecosystem
System
Integrators
Own Warehouse
(Amazon.com/walmart.com)
MOVEBOT
3PL Manual
Warehouses
3PL Automated
Warehouses
Arm/hardware
manufacturers
16. Learnings
2
WEEK
INTERVIEWS
0 1 2 3 4 5
0 12 23 35 51
Our Journey
Realized the complexity of the ecosystemand identified customers
Arm/hardware
manufacturers
System
Integrators
MOVEBOT
Own Warehouse
(Amazon.com/walmart.com)
3PL Manual
Warehouses
3PL Automated
Warehouses
61
17. We had validated the need
1
We thought we had figured it out!
2
We understood the complex ecosystem
3
WEEK
INTERVIEWS
0 1 2 3 4 5
0 12 23 35 51 61
Our Journey
We had a go to market strategy
17 follow ups lined up with industry contacts who were excited about the problem
18. Testing Hypothesis with customers and expo
contacts
“I will pay for a solution, not a software”
Principal Engineer at Procter & Gamble 3PL
Reject from 4 automated facilities
“Customers pay for the HW, and expect
the SW for free” Engineer at Robotic System
Integrators
Reject from 3 system integrators
11 rejections
“Your vision system looks really good, can you
put it on a real arm?” Operations Manager at Amazon
Robotics
Reject from 4 Stakeholders
WEEK
INTERVIEWS
0 1 2 3 4 5 6
0 12 23 35 51 61 77
Our Journey
19. Software
Provider
Time to PIVOT !
Full Solution
Provider
Cloud
Camera Gripper
MoveBot A.I
Arm
WEEK
INTERVIEWS
0 1 2 3 4 5 6
0 12 23 35 51 61 77
Our Journey
Warehouse Software
Movebot’s A.ICloud
Robot Control
20. Pivot: new business model
Key Partners
Robot arm
manufacturers
Cost Structure
Software development cost
Labor cost
Key Activities
Software R&D
Sales & BD
Aftersale Support
Domain knowledge
Key Resources
Software Simulation
Tool
Value proposition
Enable multiple use of
existing robotic arms
Cost saving
Reduce errors
Customer relationships
e-commerce
warehouses
Channels
Direct Sale
Customer segments
E-commerce players:
Amazon, Jet.com
Robot Arm
Manufacturers
Robotics System
Integrators
Automated 3rd Party
Logistics
Robot Arm
Manufacturers
Rack manufacturers
System Integrators
Full automated solution:
relieve safety, health and
peak capacity painsAccess to an arm to
develop our software
Hardware IntegrationRack
manufacturers
System
integrators
Integration cost
One time sale
1 year ROI
WEEK
INTERVIEWS
0 1 2 3 4 5 6
0 12 23 35 51 61 77
Our Journey
Revenue Streams
Subscription/ licensing revenue / profit sharing
22. “Let’s work on a value
proposition together and talk to
the CEO next week.”
System Sales Director, Kardex Remstar
“Yes, we are looking at this
missing automation step and are
considering different solutions.”
CTO, Exotec
Learning: Rack manufacturers
new partner
WEEK
INTERVIEWS
0 1 2 3 4 5 6 7
0 12 23 35 51 61 77 91
Our Journey
00
Learning: Market opportunity
Total market:
Semi-automated warehouses (L3): 3000
At $2.5 m/warehouse: $7.5b
Growing at YoY:
~20%
23. Learning: Sales process
Corporate Committee Automation Technology
VP Engineering VP Operations VP Business Dev.
Decision Maker
Video of prototype
Plant Automation
Manager
Live demo
Corporate
Automation Manager
MVP
Through Trade
Show+ ASRS
Corporate
Automation
Director Internal Champion
WEEK
INTERVIEWS
0 1 2 3 4 5 6 7 8
0 12 23 35 51 61 77 91 101
Our Journey
9
“Let’s develop a prototype
together and take it to trade
shows”
Sales Manager, Automated rack
manufacturer
110
“We will take this to corporate”
Corporate Automation Manager, UPS
24. In just 10 weeks, we...
Mapped Complex EcoSystem and found partners: Rack & Arm Manufacturers
Identified right customer(3PL) and eliminated others (Amazon)
1
WEEK
INTERVIEWS
0 1 2 3 4 5 6 7 8 9
0 12 23 35 51 61 77 91 101 110
Our Journey
114
10
Evolved our our offering from software to full solution
Understood sales and go-to-market strategies4
2
3
What we have
learned in 10 weeks
takes similar
companies 1-2yrs
Based on conversation with
Director of automation at Major
Rack Manufacturer
25. .. and we keep finding new opportunities
SorterScanner (3D)
UPS - 70% Highly Automated
Truck loading
Another Big Opportunity
Truck loading
$9B value for UPS in 5 years
An opportunity we would never discover
without “leaving the building”
WEEK
INTERVIEWS
0 1 2 3 4 5 6 7 8 9
0 12 23 35 51 61 77 91 101 110
Our Journey
114
10
26. Next Steps
We are applying to incubators!
DENSO Robotics are
loaning us an arm
WEEK
INTERVIEWS
0 1 2 3 4 5 6 7 8 9
0 12 23 35 51 61 77 91 101 110
Our Journey
114
10
MoveBot is a logistics automation solution powered by an A.I. engine
that constantly learns from data aggregation
Broader vision
29. Value created for an Amazon Warehouse
Cost/BenefitAnalysis
Cost
~$2.5M
Benefit
~$5M
Upfront Investment
~$2.4M
Saved Labor Cost
~$1.8M / year
Less Injuries and Safety
Liabilities
~0.2M / year
Avoid Human-error
Related Loss
~ $3M / year
Maintenance & Upgrade
~$0.1M / year
Packages delivered
in Peak Month
11M
# of packages robot
can pick per month
540k
# of robot needed 20
Cost/Robot $120k
Upfront Investment $2.4M
(Based on Real Data)
30. Value created for an Amazon Warehouse
Cost/BenefitAnalysis
Cost
~$2.5M
Benefit
~$5M
Upfront Investment
~$2.4M
Saved Labor Cost
~$1.8M / year
Less Injuries and Safety
Liabilities
~0.2M / year
Avoid Human-error
Related Loss
~ $3M / year
Maintenance & Upgrade
~$0.1M / year
# of worker saved 15
# of shifts 2
Annual compensation
& benefits per worker
$60k
Total saved labor cost $1.8M
(Based on Real Data)
31. Value created for an Amazon Warehouse ~2.5m
Cost/BenefitAnalysis
Cost
~$2.5M
Benefit
~$5M
Upfront Investment
~$2.4M
Saved Labor Cost
~$1.8M / year
Less Injuries and Safety
Liabilities
~0.2M / year
Avoid Human-error
Related Loss
~ $3M / year
Maintenance & Upgrade
~$0.1M / year
Reduce human error
rate
0.4%
Less # of mis-delivered
packages
~75k
Average cost of one
mis-delivery
$40
Total cost saved from
mis-delivery
$3M
(Based on Real Data)
32. 3PL Automated
Warehouses
Hardware
providers
Buy
Provide a solution
Sell, integrate, maintain
$40k $120k
Rack Providers System Integrators
New Partner Map & Go to market
Partner
MoveBot
Cost for one robotic arm
Validate need
& solution
Loan robotic arm
and rack system
Build
prototype
First sale and
implementation
Partnership Process
Marketing at
trade shows
Sell a solution
34. Operation Plan
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4
2018 2019 2020 2021 2022 2023 2024 2025 2026
Product/Market Fit
Repeatable/Scalable/Profitable
Growth model
Scaling the Business
Product
Milestone
Business
Milestone
Proof of
concept
Integrated
Industry
Solution
Simulation
Prototype Introduce AI-Cloud
Learning
Data + Platform
Industry Standard
Pilot
First
Installation
4
warehouses
10 warehouses
20 warehouses &
Diversification
50 warehouses
Financial & Operation Resources
35. Fundraising : Financial/Operation
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4
2018 2019 2020 2021 2022 2023 2024 2025 2026
Product/Market Fit Repeatable/Scalable/Profitable Growth model Scaling the Business
$40M
$30M
$20M
$10M
Angel
$150K
$5M
$0M
Seed
$1.5M
Series A
$5M
Series B
$10M
Series C
$20M
Series D
$40M
M&A / IPO
$50M+
CashReserve
Product
Milestone
Business
Milestone
Proof of
concept Integrated
Industry Solution
Simulation
Prototype
Cloud
Data + Platform
Industry Standard
Pilot
First
Installation
4 warehouses
10 warehouses
20 warehouses
50 warehouses
$50M
Financial & Operation Resources
Editor's Notes
https://youtu.be/WcWUfH-NMM0
General purpose robot automation software allowing robot arms to become intelligent by using learning & 3D-perception to pick and place objects not specifically trained for.
General purpose robot automation software allowing robot arms to become intelligent by using learning & 3D-perception to pick and place objects not specifically trained for.
Email?
How we found the solutions
Refer back to market sizing
Taking the Class, not just doing product marketing.
“And if a customer invests in G2P solutions today,” Schmidt says, “there will be no reason they can’t use a robot once they’re ready.”
Rafi Remove boxes
Rafi Remove boxes
brian/slide : puneet editing it
Great quote, great picture. Both visually fail.
Suggestion: big picture, then quotes large on the image + who said that? [ok to make it up]
Great quote, great picture. Both visually fail.
Suggestion: big picture, then quotes large on the image + who said that? [ok to make it up]
Way too busy slides.
Have demo running, and that’s it: “we trained it, and we could pick this up. We proove it to ourself, we were happy” -> put smiley face
Great story “how did we
What did you learn?
“We learned three big things:
strong demand & Market players & Target customers”
“More than 3 learnings, more like 10”
“Tradeshows are awesome for customer discovery: instead of going one at a time, we were able to talk to 2000 companies/suppliers.”
Find out: how many people to this tradeshow?
Great quote, great picture. Both visually fail.
Suggestion: big picture, then quotes large on the image + who said that? [ok to make it up]
Rafi : ‘Explaing 3rd Party Warehouses.
Great quote, great picture. Both visually fail.
Suggestion: big picture, then quotes large on the image + who said that? [ok to make it up]
Great quote, great picture. Both visually fail.
Suggestion: big picture, then quotes large on the image + who said that? [ok to make it up]
ATX: Big warehouses/industries are shopping solutions (Target, Wal-Mart, Amazon)
Concluded offering .. Slide , conclude the validation and the positive.
Great quote, great picture. Both visually fail.
Suggestion: big picture, then quotes large on the image + who said that? [ok to make it up]
Make the Rejects make feel bigger : brian
As Puneet described, we had fallen off a cliff.
It was time to pivot
Instead of a software provider, we needed to be a full solution provider
This would mean integrating the arm
The vision system
The cloud-based AI and learning algorithms
Into one package
As a result, our business model canvas changed completely
Our customer segments would have to change
Our value proposition had to be different
Our partners and sales channels were now big unknowns
We had to re-evaluate all of our assumptions. But we knew we needed a solution tailored to the needs of the customer: the automated warehouse
And going back to basics helped us identify new partners: The automated rack providers
After all, they would be bringing the “goods to our robot”
So we developed a new MVP demonstrating our grasping algorithm in a physics simulator and showed it to rack manufacturers
Their response was immediate and enthusiastic:
They knew they were missing our step and were already looking for solutions
One said “let’s take this to the CEO”
We also learned about the total potential market: there are currently over 1000 semi-automated warehouse for a total serviceable market potential of $7.2 billion and growing
Finally, talking to the rack manufacturers, and with automation managers at Amazon and UPS helped us understand the sales process and develop a go-to market strategy:
We will build a prototype that works together with the automated rack system. Take it to trade shows, and find plant and corporate automation managers willing to champion our product and push it through to the corporate decision makers
Once they signed off we could begin pilot testing in the warehouse
So just to take a step back and summarize:
In just 10 weeks, we came to understand the players in this complex industry, developed a product definition and a go-to-market strategy
In fact, based on feedback from one corporate automation manager, we learned that it has taken similar companies 2-3 years to arrive at similar conclusions
But… there is one more thing
Toward the end of the course we got out of the building one more time, this time to UPS regional sorting center in Oakland
It’s a automated facility, with package sorting and barcode scanners. But the last step – loading the truck that delivers packages to the regional shipping centers – is still manual and inefficient. Workers stack boxes by hand and the trucks are only 65% packed.
We realized that an AI optimization algorithm information about the boxes in the pipeline could drastically improve efficiency
Based on the estimates of the corporate automation manager just a 10% increase in efficiency would amount to a $9 billion opportunity
An opportunity we would never have discovered without “getting out of the building”
We are excited about our next steps:
We are applying to summer accelerators
We just had a visit from Denso robotics, who we connected with at ATX. They are planning to loan us an arm and gripper and are excited about working with us to develop our prototype to take to trade shows.
Finally, since we keep finding opportunities for AI and automation,
We have expanded our vision to one of an AI solution for logistics optimization and manipulation, that uses data aggregation to continuously learn and improve
With that we want to thank everyone involved in Lean Launchpad, in particular our mentor, Rafi Holtzman, the entire instructional staff and each of the teams, all of whom made this course such a rewarding collaborative learning experience
We’ll be happy to take any questions.
This numbers/ Amazon from who already worked this numbers on a long time.
All these numbers result from a discussion with an Amazon manager who did the math in 2014.
Spreadsheet format? Standard budget format.
All these numbers result from a discussion with an Amazon manager who did the math in 2014.
Spreadsheet format? Standard budget format.
All these numbers result from a discussion with an Amazon manager who did the math in 2014.
Spreadsheet format? Standard budget format.
Net postive warehouse , withing 1 year
1/ Prototype in the end 3PL WH
2/ Later on, sell to SI (OH Steve).