With over 1 trillion of food wasted annually, codet aims to tackle this by focusing on understanding and predicting fruit maturity to determine the shelf life and quality of the fruit.
2. The agricultural sector faces significant losses before and after harvest due to inaccurate evaluations of fruit maturity.
1.3 BN 13.2%
2 BN
(WFP, 2020)
Globally, one-third of food
(1.3 billion tones) worth $1
trillion is either lost or
wasted each year.
13.2% of losses in agriculture
are caused by pre-harvest
and post-harvest losses
Global food waste could feed two
billion people, over double the
world's undernourished
population.
GLOBAL FOOD LOSS DUE TO HARVEST LOSSES
(FAO, 2021)
(WFP, 2020)
ZERO HUNGER (SDG 2)
Around 40% of postharvest loss in
vegetables (especially fruits like
mango, banana, papaya, avocado,
etc) is due to premature and
overmature harvesting, etc.
40%
HARVEST LOSSES DUE TO
FRUIT IMMATURITY
(ETEFA ET AL., 2022)
Understanding the Global Challenge
Problem Solution
3. CODET: An Improved Approach to Optimize Fruit Harvest Timings
Our solution, CODET, addresses the critical issue of food wastage, specifically targeting the 40% postharvest loss in
fruits caused by premature and overmature harvesting. By accurately determining the optimal time for fruit harvest, our
system aims to tackle the 1.3 billion tons of global food waste, which has profound implications for feeding the
undernourished population (SDG 2 - Zero Hunger).
SOFTWARE
HARDWARE SUPPORT
Problem Solution Use of Ai
4. SOFTWARE
SUPPORT
Solution Use of Ai Competitors
CODET: Handheld Hardware Component
Traditional methods of determining fruit ripeness rely on our
senses, but these can be subjective especially considering subtle
changes within the fruit, like chlorophyll breakdown or
temperature changes by internal activities as the fruit matures.
CODET offers a more reliable solution with its multi-sensory
setup that captures these hidden features with high accuracy.
This system, along with its environmental sensors and on-board
GPS, provides farmers (our main target market) with a clear
picture of fruit maturity and surrounding conditions, all presented
in an easy-to-use interface.
HARDWARE
6. Reasons for using these sensors are rooted in their ability to non-invasively capture critical data reflecting the
biochemical and physical transformations fruits undergo as they ripen (select links to read more).
Sensor
Spectral sensors
Ultrasonic sensors
Infrared Sensors
Reason for use
● Employed to measure the density of fruits, which changes during ripening as a result of starch
conversion to sugar and the breakdown of cell walls. This sensor provides insights into the fruit's
tissue composition and water content, key parameters for assessing ripeness (Hor et al., 2020).
● Through NIR spectroscopy, they enable the rapid, non-invasive assessment of key quality
parameters such as soluble solids content (SSC), moisture content, sugar content, and acidity of
fruits, indicative of sweetness and ripeness, by detecting changes in chemical composition (Pratiwi
et al., 2023).
● Used to measure the colour patterns of fruits, which are crucial indicators of ripeness due to the
changes in pigments like chlorophyll, carotenoids, and anthocyanins as fruits mature (Jayanth et al.,
2023; Kapoor et al., 2022).
Camera
● Used to detect size and damages potentially caused by pests allowing the system to detect
changes in the appearance.
CODET: Sensor Components
Solution Use of Ai Competitors
7. Solution Use of Ai Competitors
CODET: System Overview
CODET Hardware
Data Storage
Azure Blob Storage
Azure IoT hub
Model Deployment
AKS
Front-End Application
Users
Retailors
Farmers
Customers
API
Calls
Sensor
Input
Data
Sensor Data Stored
Azure ML Studio
Category 1
Category 2
Researchers
Category 3
Users
CODET utilizes hardware sensors to send data to Azure Blob Storage through Azure IoT Hub, which is then used to train neural
networks in Azure ML Studio. The trained models are deployed via Azure Kubernetes Service, enabling seamless integration with
frontend applications.
*Note: Category 1 (Farmers), Category 2 (Retailers, Shoppers, Aggregators), Category 3 (Researchers)
8. SUPPORT
SOFTWARE
We were excited to test and manipulate Azure IoT’s new
ultraconvenient framework platform which was perfect for
sending telemetry data to and from our module and the cloud.
Considering our system is based on ESPRESSIF’s ESP32 S3
architecture, which happens to be well and seamlessly supported
by the platform, our system becomes highly scalable,
manageable and rapid.
CODET: Support Component
HARDWARE
Solution Use of Ai Competitors
9. SOFTWARE
SUPPORT
CODET: Software Component
Our machine learning model is designed and optimized for speed
and accuracy. It is based on a direct data pipeline from the
module where each sensor’s value is trained individually to make
predictions about ripeness and then another model, called a
meta-learning model based on logistic regression is trained on all
these predictions.
The model then delivers these insights directly back to the
module to show with its in-build display or allows for API calls to
display on the frontend, providing a convenient and actionable
status state of fruit ripeness.
HARDWARE
Solution Use of Ai Competitors
10. Drawing Insights from Industry Competitors
Use of Ai Value Proposition
Competitors
USA
Country
Product
Reasons for
Success
• Hazel Technologies offers products such as
Hazel 100 with 1-MCP to enhance produce
shelf life and quality.
• Their user-friendly solutions are backed by
USDA and academic studies, requiring no
additional equipment.
• Catering to various supply chain members,
their products significantly reduce food
waste and maintain produce integrity by
tackling ethylene and spore-related
challenges.
• Hazel employs innovative solutions like
sachets and stickers with natural
compounds for effective ripening control,
backed by scientific research and real-
world success.
• Continuous data collection and analysis
from partnerships enable Hazel to refine
their technology for better shelf-life
predictions and produce-specific solutions.
The Netherlands
• OneThird AI focuses on reducing food waste by
accurately predicting the shelf life of produce at
any supply chain stage.
• Their AI-powered solution combines fresh
produce scanners, smartphone imaging, and a
smart insights platform for real-time quality
assessment and decision-making.
• OneThird's technology aids in dynamic routing,
expiration date setting, and dynamic pricing,
substantially reducing fresh produce waste and
improving supply chain efficiency.
• Utilizing computer vision and machine learning, it
provides accurate spoilage forecasts, enabling
better inventory management and decision-
making.
• OneThird's reliance on extensive real-world data
to refine their algorithms enhances the accuracy
and effectiveness of their predictions, ensuring
continuous improvement and adaptation to
changing market needs.
USA
• Apeel's Apeel B coating, made from safe fatty
acids, extends the shelf life of fruits and
vegetables by replicating natural plant barriers,
reducing spoilage.
• Demonstrating its effectiveness, Apeel's solution
has notably prolonged avocado freshness and
reduced apple decay, aligning with their goal to
lessen food waste and aid environmental efforts.
• Collaborations with supermarkets, notably The
Edeka Group, have proven Apeel's impact.
• Apeel's proprietary plant-based coating
effectively extends the shelf life of fruits and
vegetables, showcasing a unique solution that
slows down spoilage and dehydration.
• Aligning with consumer preferences, Apeel's
sustainable, plant-based coating leverages
renewable resources and is safe for consumption,
differentiating it from synthetic alternatives.
Studying innovative competitors sheds light on unique solutions, guiding our approach to scaling and improving our offerings.
11. CODET outperforms in precision, user experience, and reliability, setting new standards in agricultural ripeness
detection technology.
Value Proposition Customer Segments
Competitors Risk a
Redefining Quality Assurance in Agriculture
USA
Country
Non-Destructive
Testing
The Netherlands
USA
Accuracy of Ripeness
Detection
Real-Time Data and
AI Integration
Reliance on multiple
sensors/assessments
Worldwide Ghana
Stakeholder-Specific
Features
12. CODET enhances the efficiency, quality, and sustainability of food production and distribution across the agriculture
supply chain.
A Step Closer to Zero Hunger
Why Zero Hunger?
By accurately predicting the optimal harvest time for fruits, CODET helps in reducing
post-harvest losses, ensuring more food reaches the market in its best condition. This
not only maximizes the yield from each harvest but also improves the availability and
nutritional value of fresh produce. Furthermore, by aiding farmers/distributors in
reducing waste and optimizing resource use, CODET contributes to more sustainable
food systems.
How? Potential Use Cases
CODET's technology offers real-time crop ripeness data for automated, efficient
harvesting as well as identifying damages and external issues (associated with pests).
This reduces labor costs and improper harvest risks, streamlining processes and
enhancing yield efficiency with reduced waste.
The global adoption of CODET offers immense data acquisition potential, providing
valuable insights for researchers and Agri-tech companies into crop growth, regional
agricultural variations, and environmental impacts potentially driving breakthroughs in
crop science and farming method, etc.
Value Proposition Customer Segments
Competitors Risk a
13. Quotes
Key Features
What they Value
Targeting Diverse Customer Segments
Value Proposition Customer Segments Risk and Mitigation
Farmer Francis
“I am tired of losing money on
harvesting unripe and overipe
fruits”
• Owns large commercial farm
and is supplier to many
businesses internationally
• Has incurred many losses over
the years as his current
equipment is unable to identify
properly fruit maturity and
quality without damaging it.
A way to determine fruit
quality from size, ripenes and
damage by pest.
Shopper Sarah
Retail Rheita
“It would be nice to actually buy
fruit that is neither too ripe nor
too unripe for a change.”
“Is managing the quality of
fruits in the aisle impossible?”
• Purchases fruits for their family
• Is not too bothererd by selecting
ripe fruits or unriped fruits
• She has her own techniques she
uses to check for ripeness which
she learnt from her parents but it
does not always work (probably
coincidence?)
• Is a manager of a retail store in
the city
• Many customers complain about
the how the fruits are
riped/unriped/overiped and
selecting a fruit is like a lottery –
nothing is organized
A way to pick out the exact fruit
ripeness level (based on
personal preference)
A way to determine fruit maturity
level and shelf life thereby managing
their supply more effectively,
reducing waste and improving
product offerings.
Scientist Simon
”It is difficult to perform quality
research on fruits with such
limited data”
• Works at an Agriculture
research institute
• Is looking for new data about
fruit development to make a
breakthrough in their research
Large amounts of data for
studies on crop growth, fruit
development, and the impacts
of various environmental factors.
Addressing specific needs across the agricultural supply chain, from harvest timing to research data, supply management,
and selection preferences. We do this by target markets to provide required solutions that fit their preference.
*Note: Images were created using DALL-E Ai image generator
Category 1 Category 2 Category 3
14. Contact
Status
Insights from stakeholders in the agricultural supply chain
Position
Name
Insight
Our insights revealed that those who engage with large quantities of fruits on a daily basis (retailers and farmers) are
most in need and willing to engage with the project whereas small scale businesses/informal sector do not see a fit yet.
Shoprite Fullmark
Largest Retailer in
Ghana
Redeemer Kumah
(Procurement Officer)
Interested
Is happy to test it
within their operations
Dome Market
Large (informal)
Market in Accra
Aunty Ama
(Small-scale shop
owner)
Contacted
Is excited of its
existence but does
not see a fit
Palace
Large Retailer in
Ghana
Ama
(Branch Manager)
Unconfirmed
Suggested engaging
with management
after some tests.
Agri-Impact
Local Greenhouse
Farm
Dennis & Kwasi
(Farm Supervisors)
Interested
Is happy to test it
within their
operations
Melcom
Large Retailer in
Ghana
Mohammed Taha
(Procurement Officer)
Not Interested
Sees no use with the
project and prefers
current tools
Eve-Lyn Farms
Commercial Fruit
Farmer/Exporter
Bassam George Aoun
(Owner)
Interested
Is happy to test it within
their operations
Value Proposition Customer Segments Risk and Mitigation
15. osition Customer Segments Risk and Mitigation Business Model
Key risks Mitigants
Risk of Misalignment with
Market Expectations
Even if CODET offers innovative features,
there's a risk that these stakeholders might
not adopt the technology if it doesn't meet
their specific needs or if they perceive it as
not being a significant improvement over
existing methods.
3
Technology Adoption
1
Our chat with Melcom's procurement office
revealed their preference for their current, albeit
destructive, produce assessment tools, citing their
effectiveness and accuracy. This decision highlights
a significant hurdle for us: even when faced with
innovative alternatives, major retailers' resistance to
change could limit our product's market
acceptance.
Product Performance and
Reliability
2
Eve-Lyn Farms and Shoprite Fullmark, representing
large-scale operations in agriculture and retail, have
expressed interest in innovative solutions but
emphasize the need for reliability. Their goal is to
achieve near perfection in produce quality,
underscoring that any new technology introduced
must guarantee consistent performance to be
considered viable for their operations.
Customization and
Adaptability
4
Our strategy involves tailoring CODET's
technology to meet the unique needs of
category 1 and 2 through customizable,
modular features.
By engaging in ongoing dialogue with
these stakeholders, we'll refine our
product to match their specific
requirements and quality standards.
Additionally, we plan to run pilot
programs across diverse user groups,
demonstrating CODET's benefits like
reduced waste, increased efficiency, and
improved produce quality.
Risks and Mitigation
In the process of scaling, CODET is likely to face risks which affects its operations
*Note: Category 1 (Farmers), Category 2 (Retailers, Shoppers, Aggregators), Category 3 (Researchers)
16. • Production Materials (3D printer,
Filament, etc)
• Labor and Equipment
• Sales and Marketing
• Administration Costs
• Software Development Cost
Potential Cost and Revenue Model:
Funding Requirement (MVP):
Cost Revenue
• Software Maintenance
• Hardware Sales
• Platform Subscription
• Licensing Technology
• Data Monetization
Business Model
CODET's business model effectively blends product sales with a SaaS framework, ensuring a balance of initial revenue from hardware sales and
consistent earnings from software maintenance, subscriptions, and technology licensing. The model accounts for various operational costs such
as production and marketing while emphasizing long-term financial stability and adherence to circular economy principles, focusing on waste
reduction and resource optimization to enhance overall business value.
Future Plans
Risk and Mitigation Business Model
17. • Laboratory testing with
collected fruit samples.
• Field testing with partner
farmers to fine-tune the
system under real-world
conditions.
Development Testing
• Initial rollout of CODET
systems at local farms for
harvest timing optimization.
• Setup at retail points for
quality assurance of
produce.
Deployment
• Collecting data from
deployments to improve
the AI-CNN model.
• Making necessary
adjustments to hardware
based on user feedback.
Refinement
• Expanding the reach to more
farms and retail points.
• Regular system updates and
maintenance based on
ongoing data analysis and
technological advancements.
Scale up and
maintenance
1 YEAR
0-3 months 4-6 months 7-9 months 9-12 months year 1+
This timeline and agenda are subject to adjustments based on funding availability, stakeholder collaboration,
and real-world testing outcomes. The goal is to have a fully operational system within a year (after August
2024), setting the stage for a broader market introduction and an increase in stakeholder value.
Short-Term Plan
• Assemble the hardware:
sensors, boards, and
components.
• Software development of
the AI-CNN model,
including data collection,
training, and validation.
Future Plans Team
Business Model
18. Future Plans for Market Leadership
CODET aims to reduce waste at the
harvesting stage, directly impacting
farmers' profitability and reducing
retail spoilage.
Short-term
Medium-term
Long-term
Initiatives
$
$$$
Enhanced precision of the AI-CNN
model, enriched with more data, will
not only reduce waste and optimize
supply chains but also improve market
standards for produce quality. This
data will bolster predictive analytics
for better crop pricing, distribution
strategies, and ensure fresher produce
delivery.
Initially targeting farmers in the short term, CODET aims to expand to other supply chain users as the technology
advances and more data is collected.
The aim is for CODET to become the benchmark
for quality assurance in agriculture, with ongoing
improvements and updates to keep pace with
technological advancements and industry trends
(partnering with the World food programme/Food
and drug authorities to set new standards of fruit
quality assessment). Here, we aim to make an
affordable component of the product that
everyday consumers can purchase to use there by
improving the quality of their purchases.
Future Plans Team
Business Model
19. The Team Behind CODET
Kwasi Boamah Tano | Ghanaian | 4th Year Undergrad
• Equity and Credit Risk Research
Analyst at IC Securities Ltd.
• Finance and Operations manager
at Sorted Chale.
• Apprentice at Pave Investment
(VC firm in Africa)
Experiences
Skills/Interests
Projects/Research
Edwin Nsoh Awariyah | Ghanaian | 4th Year Undergrad
• Proficient in Pitchbooks
• Proficient in Valuations and
Financial Modelling
• Proficient Refinitiv, Excel,
Tableau, etc.
• Interests in Community
Involvement
• Interests in Travelling
• Interests in Cooking
• Customer Sentiment Analysis
on Target
• Market Basket Analysis
• Human Resourse
Dashboard/Analytics
• Sales Dashboard/Analytics
• Research into the effect of
sentiment on the accuracy of
stock price predictions in
Nigeria
• Vice President of the
Advanced RISC Machines
(ARM) in Ashesi
• Backend Developer with
ReactJS and React Native as
an intern at El-Parah
• Finalist in the Coins Grand
Challenge, Birmingham, United
Kingdom on linking melamine
to cancer, 2022
• Third place in the ShareVision
Global Student Competition on
blockchains, 2022
• I enjoy programming and building
projects related to artificial
intelligence and problem-solving
especially critical but often
overlooked problems of the world
today
• Proficient in C, C++,
MicroPython, Espressif,-IDF
Arduino
• Full stack developer on
emphasis on BaaS and SaaS
• Car Parking Detection System
• Building an eight-bit CPU for
the University
• IoT Library Management
System
• IoT bus parcel tracking System
• Machine Lot Parking Systems
using Computer Vision
• Building a video game from the
ground up
• Dental Management Webapp
• First Runner-up at the CFA
Research Institute Competition
(Local)
• Student Consultant within the HBS
Immersive Field Course (2024)
Future Plans Team
20. Visiting Agri-Impact
Team visiting Agri-Impact in Brekusu and
interviewing two workers (Dennis and Kwasi).
Building and testing the model
One of those late nights building and testing the model
whilst planning stakeholder interviews
Melcom retail outlet
After interviewing a worker at the procurement office of
melcom, we went on to purchase more fruits for testing
Dome Market
Talking with Aunty Ama on how she determines whether
fruits are mature and asking her to grade our fruit, so we
test and compare
Shoprite Fruit Distribution Warehouse
Talking with Mr. Redeemah, we understood how he finds
farms that meet Shoprite’s quality measures to partner
and how he was hoping for improvement in the agricultural
sector.
Thank
You