2 of 27

Investment Highlights

Bossa Nova is targeting a $4 billion market opportunity

The retail shelf is a place where $5.3 trillion of goods passed through in 2016
Bossa Nova helps retailers and suppliers gain visibility into the retail shelf, allowing them to optimize retail merchandising
Bossa Nova’s existing customer engagements represent a $2.7 billion recurring revenue opportunity
By 2018, Bossa Nova will have a suite of solutions addressing a $4 billion market through licensing of data, A.I. modelling, software and autonomous robotic capture systems

Technology to bring retailers and retail workers into the era of Artificial Intelligence and Robotics

Bossa Nova’s solution brings optimization into the retail shelf, providing retailers and manufacturers with a step function improvement for on-shelf data that is 2x more accurate⁽¹⁾ and 10x more frequent, at 1/3 the cost⁽²⁾
In addition to faster and more accurate data, Bossa Nova’s systems also allow retailers to redistribute their on-the-floor labor force, allowing more workers to be available to interface with customers

Highly profitable, recurring revenue and scalable business model

Bossa Nova deploys, manages and maintains robots in the field and sells data and analytics as a service to customers
Bossa Nova replaces slow, inaccurate and expensive processes with a highly scalable A.I solution and fast, accurate robotic data capture
Performance on any one of the Company’s current customer engagement will allow for scale into >$100mm in revenue

An early lead with a sticky solution and high barriers to entry

Contracts are already signed for the roll-out of Bossa Nova’s solution into XX stores, and S.O.W.’s are in place representing a total opportunity of X,X00 stores
Bossa Nova’s system is being integrated into customers’ ERP’s and physical store processes
Bossa Nova has already captured 30 million images of retail shelves, allowing for the most accurate A.I. model for retail store applications. By end of 2017, Bossa Nova will be capturing 4.2 million images per day

World class team including leaders in A.I. and Robotics research and development

CEO brings over 25 years of experience as a founder, public company CEO, board member, investor and technology pioneer
CBO and CTO have worked together for more than 9 years and bring into Bossa Nova cutting edge robotics research from Carnegie Mellon and real world commercial expertise in business and marketing
Research and development is managed by leading A.I. and robotics engineers from Carnegie Mellon’s robotics program, NVIDIA, Amazon, Google and Microsoft

(1) Results from Feb 2017 customer audit.

(2) Labor expense associated with scanning shelves is estimated to be $275k per year per store.

3 of 27

Walmart Shareholders Meeting (June 2, 2017)

“We have tests going on with…

robotics and image analytics to scan aisles for outs”

Doug McMillon, CEO

Bossa Nova 2016 Robot Prototype

Source: Timestamp 2:07:11,http://corporate.walmart.com/shareholders

4 of 27

Bossa Nova’s Market Opportunity

$5.3 trillion of goods passed through retail shelves in 2016

Critical point where consumer demand meets supply
Customer 1’s brick and mortar revenue in 2016 was $470 billion with a GMV multiple times greater
Customer 1 brick & mortar revenues alone exceed total U.S. e-commerce sales

The retail shelf execution is the biggest operational challenge facing big-box retailers
Despite a highly optimized upstream supply chain, products on the shelf are poorly tracked with the store considered a “black hole” of information
Incentives to keep retail shelves optimized are misaligned with retailers wanting to minimize holding costs, but manufacturers wanting excess inventory on-hand to ensure in-stock rates
Stores are large and sell too many SKUs to be effectively managed by limited in-store staff processes
Repercussions are significant

Product position on the retail shelf is the biggest determinant of sales performance and revenue generation
Poor processes require multiple facings of products to ensure stores are in position to meet sales demand
Stores buy more inventory than they can sell in a given period and build percentage buffers into inventory counts to ensure stores are in stock

Improving shelf execution is expensive and labor intensive, which is contrary to the cost-cutting plans of most retailers
Bossa Nova provides visibility in to the retail shelf which enables stores to drive efficiencies in operational processes by using data to drive labor productivity improvements

5 of 27

A Vision for Brick and Mortar Retail

“Bossa Nova makes retail store operations more effective by using robotics and AI to provide high-fidelity data which helps retailers, their employees and suppliers deliver the best shopping experience to their customers.”

6 of 27

The Retail Solution For The Era of A.I. and Robotics

Automated Robotic Capture (ARC)

Fully autonomous mobile robots to physically capture 2D and 3D imagery from the retail environment

Robot Operations Center (ROC)

Fleet management for stores and HQ to manage operational schedules and monitor KPIs.

A.I. Services (AIS)

A.I. model development and training for shelf compliance metrics including product identification, out-of-stocks, location, price and other metrics.

Data Exchange Portal (DEP)

Single point of access to supplier owned shelf data for suppliers and other third parties.

Field Applications (FA)

Mobile applications for store employees to access alerts and auto-generated task lists.

Customer Analytics and Reporting (CAR)

Management dashboard and reporting tools for processed shelf data.

Solution improves organically by evolving algorithms with more data.

7 of 27

World Class Executive Team

CBO

CTO

CFO

Software

Hardware

Perception

CEO

Delivery

8 of 27

How Customers Evaluate ROI on Bossa Nova

Bossa Nova delivers customer value in five ways. Many of these alone pay for the solution entirely

Greater on-shelf availability to increase sales
Every one percent improvement in on-shelf availability has been tied to a one percent gain in sales
In a recent audit, it was found that there were ~220 out-of-stock items at 7 a.m. in a scan of 15 aisles. This was a time when the shelves were supposed to be fully stocked
Bossa Nova estimates that it can help retailers increase on-shelf availability to a point that fully covers the cost of the system
Lower inventory holds to reduce working capital
High inventory accuracy allows retailer to hold less inventory. A 1% reduction in inventory frees $45k in inventory per store
Bossa Nova estimates that it can help reduce inventory in stores to cover the first year cost of the system
Reduce spend on shelf monitoring processes and redistribute labor to customer facing tasks
Approximately 10% of a store’s workers have a responsibility to check shelves for compliance, out-of-stocks and other metrics. For Customer 1, this amounts to $275k per year in labor costs
Bossa Nova alleviates these duties from employees
Real-time access to data needed to ensure products are being stocked in ways to maximize sales
In one form or another, CPG manufacturers pay retailers for the most desirable shelf space. For example, items at eye level on the shelf have a sales velocity that is multiple times greater than foot level
Given retail workers frequently stock products incorrectly, CPG manufacturers hire external firms and also use direct labor to perform in-store checks. The cost for external contractors to monitor shelf compliance averages $170k per year per store for Customer 1’s top 20 suppliers
Near real-time visibility into retail shelves helps direct suppliers’ delivery routes, saving as much as 15 mins at each locations
Retailers will have the ability to sell the data to their suppliers to help off set the cost of Bossa Nova deployments
Sale of high frequency and accurate on-shelf data to manufacturers and supply chain
Customer 1 already offers suppliers an application called D, which is meant to provide store inventory levels
Bossa Nova will allow customers the ability to offer suppliers much desired shelf images that helps them better monitor the way their products are merchandised on the shelf

(1) Results from Feb 2017 customer audit.

(2) Labor expense associated with scanning shelves is estimated to be $275k per year per store.

9 of 27

Bossa Nova Will Create Billions of Dollars of Value

Retailers spend approximately $150-275k per store per year on labor associated with scanning retail shelves
Across Bossa Nova’s customer engagement pipeline of 28,718 retail stores, this represents $4.3 billion annually for tasks associated with scanning retail shelves
Retailers spend ~$16k per store per year utilizing third party contractors to perform weekly audits on in-store shelf inventory⁽¹⁾

Anecdotally, Bossa Nova’s customer believes $650k per year per store will be regained from avoiding lost sales by having accurate shelf data. Management estimates a similar figure
Additionally, Bossa Nova estimates that its data will allow for more accurate and frequent decisions to increase sales velocity, leading to revenue gain of ~$25-75k per store per year

Suppliers to retailers regularly pay field merchandising firms to conduct shelf audits. In 2016, it’s estimated that Customer 1’s top 20 suppliers spent $750mm on these services, a portion of which is for shelf monitoring
The field merchandising industry is ~$4 billion per year. Management estimates ~20-25% of this market is for shelf monitoring services

Labor Efficiency

Maximize Revenue

Data to Suppliers

Value Created per Store⁽²⁾

($000's)	Value
	Created

Labor Efficiency	$275
In-store Audits	16
Avoiding Lost Sales	650
Optimization of Sales Velocity	75
Supplier Shelf Monitoring	50
Value Created Per Store	$1,066

(1) Management estimate based on retailers performing weekly audits on 700 items per store using field merchandising service firms.

(2) Management estimate of value created per large format store.

10 of 27

Customer Pipeline

Focus Customers for 2017
US$ 000’s	Revenue	Applicable	Fees	Traction
	Opportunity	Stores	Per Store / Year

Customer 1	$337,500	4,500	$75	P.O. for XX stores and S.O.W. into XXX stores
Customer 2	$184,000	1,840	$100	Deployment in Pittsburgh. 3-yrs of single store testing
Customer 3	$180,600	1,806	$100	Testing in 2015-16. Identifying internal business owner
Customer 4	$100,000	1,000	$100	Single store deployment in 2016. Integration in 2017
Total	$802,100	9,146

Pipeline
US$ 000’s	Revenue	Applicable	Fees	Traction
	Opportunity	Stores	Per Store / Year

Global Int’l	$225,000	3,000	$75	Deployments in the U.K., Canada, China and Japan
Global #3	$468,900	4,689	$100	Customer diligence targeting 2018 testing in France
U.S. #3	$277,800	2,778	$100	P.O.C. testing in 2016
U.S. #4	$227,400	2,274	$100	NDA stage
U.S. #5	$220,500	2,205	$100	NDA stage
Germany #1	$102,500	1,025	$100	NDA stage. On pause given location is Germany
Global #4	$96,200	962	$100	Subject to deployment with other customers in the territory
Australia #1	$77,600	776	$100	NDA stage. On pause given location is Australia
U.S. #6	$72,700	727	$100	NDA stage
U.K. #1	$69,100	691	$100	RFID opportunity in the U.K. NDA stage
Australia #2	$32,400	324	$100	NDA stage. On pause given location is Australia
U.S. #7	$12,100	121	$100	NDA stage
Total	$1,882,200	19,572

11 of 27

Expected Roll Outs In The Next 24 Months

Cumulative Units	2017		2018				2019
	Q3	Q4	Q1	Q2	Q3	Q4	Q1	Q2
Customer 1	20	40	50	50	150	250	450	750
Customer 2	1	1	1	1	16	21	71	131
Customer 3	1	1	1	1	1	1	31	41
Cumulative Total	22	42	52	52	167	272	552	922

12.5% of current customer store count

Source: BNR Financial Forecast 2017-07-20

12 of 27

Customer Engagement

Sponsored by SVP X, VP Y
2.5 year of progressive testing proving autonomy, data capture, analysis, accuracy and speed of data delivery
$Xm NRE fees agreed for development of a robot tuned for Retailer and AI trained on environment
XX stores to be deployed from July ‘17– February ‘18 with expansion to XXX stores starting FY2018 Q3
First XX stores billed at $100,000 per robot per store per year for core services.
Installation and maintenance covered under separate fee

Description of Engagement

Engagement Timeline

Projected Revenue ($mm)

Key Agreements Signed to Date

(1) Management estimate based on retailers performing weekly audits on 700 items per store using field merchandising service firms.

Date	Event
Aug-15	MSA and SOW #1
Jun-16	SOW #2
Jul-16	P.O. for $X.Xmm
May-17	SOW #3
Jul-17	P.O. for XX store deployment, $X.Xmm/year

Oct-14	Initial demonstration at home office
Jan-15	2-week paid test and exec demo in concept store
Aug-15	MSA signed and first store deployed
Oct-15	Presentation to CEO
Feb-16	Deployment in three stores
May-16	Customer dedicates internal team for systems integration
Jul-16	S.O.W. for XXX store deployment signed and NRE for $X.Xmm is received
Feb-17	Evaluation test passed
Jul-17	P.O. for XX stores and refined S.O.W. for XXX store deployment signed, NRE fees of $X.Xmm

13 of 27

Traction To Date

Signed Engagements Representing >$500mm

in Revenue Potential

>2 Years of In-Store Testing and Development

Customer Validated and Approved Technology

IP Portfolio

Bossa Nova has 4 paid customer engagements totaling $X.Xmm in NRE fees, representing an annual recurring revenue potential of >$500 million by 2020
Customer 1 opened a P.O. for Bossa Nova’s solution to be deployed in XX stores, which leads into a XXX store deployment in 2018 and up to X,X00 store deployment in 2019-21
In addition, the Company’s sales pipeline represents revenue potential of $400mm by 2020

Bossa Nova has a large early lead in the race to enable retailers with A.I. and robotic solutions, building early customer stickiness and the dataset needed to deliver value
The Company has performed 1,750 hours of in-store testing, scanned over 760 km of aisles and captured over 34 million images
The Company is currently deploying its 12^thgeneration robot, which includes many learnings related to operating inside of a retail store with high pedestrian traffic

In February 2017, Bossa Nova passed a formal evaluation and audit with Customer 1 demonstrating financial ROI. The Company passed the same test again in March 2017, consistently exceeding 96% accuracy (versus human scans at 50%).
In April 2017, Bossa Nova tested data capture in Customer 2 and presented data accuracy of 90% for Out-of-Stocks, 80% for product location and 70% for price.
In March 2017, Customer 1 tested and confirmed an accuracy rate of 90% for product location

Bossa Nova is early in filing patents with fundamental claims on using A.I. and autonomous robotic capture systems in retail environments
More than 20 Disclosures, 6 filed provisional patents, 4 provisionals in preparation, and 2 converted to full patents
Multiple patents that are fundamental to delivering an automated data solution

14 of 27

Company Focus Areas for Next 18 Months

Execute on the Customer 1 Opportunity

Continue to Improve Technology

Develop Additional Customer Engagements

Get Ready to Scale Into >XXX Stores

The Company will remain focused on executing on the Customer 1 opportunity for the foreseeable future as this represents a large opportunity with significant traction
In 2017 and early 2018, the Company’s resources will be devoted primarily to ensuring success of Customer 1 initial XX store deployment. Customer 1 has indicated that a XXX store P.O. would come in early 2018 if the initial XX store deployment is successful

The Company’s A.I. system will continue to be developed to increase speed and accuracy, while progressively reducing reliance on a Human-In-The-Loop (HITL) system
The robot’s hardware will undergo two more planned revisions during the XX store deployment and leading up to a XXX store deployment, in order to incorporate learnings from in-store testing

The Company will aim to reduce customer concentration risk by bringing other customers closer to Customer 1 in terms of traction in 2017, without slowing down progress with Customer 1
There are several customers that have requested in-store testing. The limiting factor today is management bandwidth and the number of robots available for testing, which will improve after the current financing

Customer 1 has indicated a desire to scale into a XXX store deployment by mid-2018. There are also opportunities to deploy into more stores with additional customers
In order to successfully execute on a XXX store deployment, management estimates the Company will require 150 full time employees and the creation of departments focused on store operations and fleet management

15 of 27

5 Year Forecast

(US$ 000’s)	2017	2018	2019	2021	2022
Installed Units	42
Revenue	$ 2,762
Gross Profit	$ 2,285
Gross Margin	83%
EBIT	$ (8,879)
EBIT Margin	na

Other Metrics
Customer 1 has identified X,X00 stores for deployment.

Source: BNR Financial Forecast 2017-07-20

16 of 27

Valuation Model

(1) Source: CapitalIQ as of May 18, 2017

(2) Source: BNR Financial Forecast 2017-07-20

Comparable Companies¹

Valuation based on Year-End Exit in 2019 or in 2020²

17 of 27

THANK YOU

18 of 27

EXECUTIVE SUMMARY

19 of 27

Bossa Nova Enables More with Less

Retailers balance between store headcount vs. store functions such as On-Shelf Availability, inventory counting and customer service
To compete with e-commerce vendors, some retailers cut back on headcount as well as store functions, such as manual inventory tracking processes and customer support. (From 2007 to 2017, headcount per U.S. store at Walmart decreased from 406 to 321. At the same time, some store processes such as manual shelf monitoring and door greeters were eliminated)
Bossa Nova enables the retailer to perform many more functions while holding headcount constant. For example, stores will be able to bring back shelf monitoring processes that track data at much higher frequency and accuracy than was possible before, allowing for greater in-store efficiency and better planning

Headcount

Per Store

Balance Between In-Store

Headcount and Functions Today

2007: 406 employees per store

2017: 321 employees per store

2007: Store processes for inventory counting, OSA and cust. exp.

2017: Store processes focused primarily on OSA

Store Functions

A.I. and Robotics Allow More Functions With Same Headcount

Illustrative Productivity Improvement

20 of 27

THE BOSSA NOVA SOLUTION

21 of 27

The Retail Solution For The Era of A.I. and Robotics

Automated Robotic Capture (ARC)

Fully autonomous mobile robots to physically capture 2D and 3D imagery from the retail environment

Robot Operations Center (ROC)

Fleet management for stores and HQ to manage operational schedules and monitor KPIs.

A.I. Services (AIS)

A.I. model development and training for shelf compliance metrics including product identification, out-of-stocks, location, price and other metrics.

Data Exchange Portal (DEP)

Single point of access to supplier owned shelf data for suppliers and other third parties.

Field Applications (FA)

Mobile applications for store employees to access alerts and auto-generated task lists.

Customer Analytics and Reporting (CAR)

Management dashboard and reporting tools for processed shelf data.

Solution improves organically by evolving algorithms with more data.

22 of 27

ARC – Automated Robotic Capture

ARC is a fully autonomous mobile robot designed to mobilize sensors to capture data for specific tasks.

ARC is a modular design to enable interchangeable sensor payloads and custom compute requirements.

The first iteration uses 2D, 3D and Lidar for optical scanning. Version 2 will use RFID.

Description

The company has dedicated significant resource to develop ARC as a core competitive advantage.

Generation #12 is currently deployed in Customer 1, Customer 2 and Customer 3.

Two further revisions are planned prior to mass production.

Development Stage

Today, Bossa Nova’s cost to manufacture an ARC system is $80k. This cost is expected to decline to $50k for a XXX store deployment.

New customer will be billed on a square footage basis. Management estimates a $30-40k⁽¹⁾ annual recurring fee for most future customers at ~60% margin.

Billing and Margin

Power

Drive

Compute

Payload

(1) Fees for ARC will be reduced as other fees are added, resulting in a higher overall solution fee for the customer.

23 of 27

Mobile Applications

Mobile Applications give store and field-based employees access to Bossa Nova data based on job function and responsibility.

Regional managers have access to aggregated data for a set of specific stores while Store managers are able to view in depth detail of a single store and interface with ARC. Replenishment tasks are also communicated to Store Associates through Mobile Applications.

Description

This service will be productized in 2018 when smaller customers are introduced to the Bossa Nova solution.

Bossa Nova’s early customers each develop their own mobile applications to access data captured by ARC.

Development Stage

Management is targeting a per device annual license at $20 - $40 per month generating up to $19,200 per store per year.

Large format retailers typically have 30-40 mobile devices per store.

Billing and Margin

Compliance tasks are presented in graphic format to simplify communication and increase execution success

24 of 27

AIR – A.I. for Retail

AIR (AI for Retail) are AI algorithms customized to customer specific shelf data to identify the status of shelf compliance.

This includes creating A.I. models for customers, developing the training dataset, and providing inference services for label detection, product recognition, out-of-stock recognition and other metrics as well as Human-In-The-Loop (HITL) services where needed.

Description

Scene segmentation using machine

and deep learning to identify section, shelves, labels and gaps

Bossa Nova has developed Training tools to enable Human-in-the-Loop development of classifiers for common retail problems including label and out-of-stock detection.

The development roadmap will engineer humans out of the process over time. Development is focused on automating the most time-consuming tasks first.

Development Stage

The company is moving to a transaction fee per recognition and management estimates $25-50k per store per year at 80% margins. Pricing will adjust based on customer value.

Billing and Margin

25 of 27

CAR - Customer Analytics and Reporting

Bossa Nova provides retailers with advanced analytics and reporting capabilities through API data streams as well as a dedicated software portal.

Large data sets captured over extended periods of time will enable micro-planning of the retail shelf to optimize sales per square foot.

Description

Today Bossa Nova provides product data to help customers optimize task planning and execution.

With expansion to more stores, data science will determine optimal placement of categories, brands and products to generate insights that will improve sales performance.

The company estimates that 50 stores and six months data collection will yield actionable insights.

Development Stage

Reporting services will be priced for market adoption and may include cost per report and/or a seat license for analytics. All services will be online with margins expected to exceed 80%.

Billing and Margin

Bossa Nova’s Customer Analytics and Reporting allow retailers to monitor and optimize business processes to help transform enterprises into the era of A.I. and robotics

26 of 27

ROC – Robot Operations Centre

A version of Bossa Nova’s internal monitoring tools will be made available to customers to provide full visibility into the status of each ARC and access to scheduling for scans and other urgent needs.

For example, a store manager may request a scan of an underperforming zone, or HQ can ask for validation that recalled products have been removed from shelves.

Description

The Company has already developed tools to remotely monitor each ARC system in near real time. A subset of these monitoring tools will be made into customer facing products that provide limited access to ARC’s planning module. Development is slated for 2018.

Store visibility will be incorporated into Mobile Applications while HQ will access this data through a browser.

Development Stage

Access to ROC will require a seat license charged at $1k - $5K per year and management estimates margins to be in the region of 80% - 90%.

Billing and Margin

First design iteration of ROC data showing mission detail including scheduled aisles, path planning, capture quality and robot health.

27 of 27

DEP – Data Exchange Portal

The Data Exchange Portal is a cloud repository of real time consumer product information designed for use by the supplier, merchandising and market research communities.

DEP is a query based interface of aggregated product information.

Description

The Data Exchange Portal is in the planning stage. The Company will seek input from retailers and suppliers during the design research process to ensure the right information is made available for each user.

Development Stage

Retailers are sensitive about sharing data however, management is confident that many will approve sharing aggregated information that does not include identifiable data to link the source. In addition, it is expected that retailers will expect a revenue share from such a service.

Margins are expected to exceed 60% after revenue sharing.

Billing and Margin

Bossa Nova’s Data Exchange Portal provides the entire retail global supply chain visibility into the most important part of the chain – where consumer demand meets supply