Bene Bono's Warehouse Production Line: Web Technologies Meet Industrial Sensors

Summary

このスクリプトは、Bonuというフランスの会社が、食品ロスを減らすために不規則な形や色の果物や野菜を販売するビジネスモデルと、その生産ラインの最適化について説明しています。Bonuは、顧客がオンラインで注文をカスタマイズできるフルカスタマイズ機能を提供し、生産性向上とエラーの減少を目指して、新しいセンサー技術を導入しました。このイノベーションにより、Bonuはより効率的で信頼性の高いサービスを提供し、将来的にはデータ利活用とより多くの機能拡張を計画しています。

Takeaways

• 📈 会社のミッションは、食品ロスから良い製品（果物や野菜）を救うことです。
• 🥦 販売している商品は、通常の店では販売されないが、理由があって販売されない商品も含まれています。
• 📊 3年間の運用により、フランスとスペインの6つの主要都市に存在し、2万以上の顧客を有しています。
• 🚀 週に1万を超える注文を準備し、最大の倉庫でその数は数ヶ月以内に倍増する予定です。
• 🛠️ 現在の生産システムはシンプルで効率的ですが、よりカスタマイズされた注文に対応できないため、改善が必要とされています。
• 🔄 新しい生産ラインには、紙レスで、従業員のトレーニングが不要、果物や野菜のリストが変わると自動的に対応できるシステムが採用されています。
• 🔧 技術的な選択においては、ロックインされていない高度に信頼性のあるシステムを追求し、技術チームのスキルを最大限に活用することを目指しています。
• 🌐 産業用センサーを使用しつつ、オープンのEthernet/IPプロトコルを利用することで、プロペティのハードウェアから離れています。
• 📱 従業員用のスクリーンは、消費者向けのタブレットで、従業員が簡単に注文を準備できるようになっています。
• 🔄 部署には多くのスクリーンがあり、生産ライン上の各ステーションで何が起こっているかをリアルタイムで確認できます。
• 🔍 データの利活用により、マージン計算や納品チームの選択肢に役立つ情報を得ることができます。
• 🚧 今後の課題としては、より高い信頼性を確保し、アプリのバージョンアップや監視システムの改善が挙げられます。

Q & A

Bonuのミッションは何ですか？

Bonuのミッションは、食品ロスを減らすことです。具体的には、通常の店では販売されない理由がある果物や野菜をセーブし、それらを消費者に届けることです。

Bonuが提供する商品はどのようなものですか？

Bonuは果物や野菜を始めとして、一年半後の現在では食品用品や卫生・家事用品、ペットフードまで展開しています。これらの商品は通常の店では販売されないが、品質には問題がありません。

Bonuの配達システムはどのように動作しますか？

Bonuはサブスクリプションモデルを採用しており、顧客は毎週の注文を準備し、直接消費者に商品を販売しています。また、フルカスタマイズ可能な注文システムを導入し、顧客は毎週の提供商品から自分の注文を完全にカスタマイズすることができます。

Bonuの生産システムの課題は何でしたか？

Bonuの生産システムは非常に基本的で、デジタルツールがありませんでした。従業員は注文準備中に紙のリストを読み、注文の準備を行っていました。しかし、注文の量が増加し、よりカスタマイズされた注文に対応するためには、現在のシステムの改善が必要でした。

新しい生産システムの設計原則は何でしたか？

新しい生産システムの設計原則は、紙レスで、新入社員のトレーニングが不要、果物や野菜の販売リストが変わることに対応できること、そして、新しい技術を導入するための大きなCAPEXコストがなくなるように、技術を選択することが重要でした。

Bonuが選んだ技術はどのようなものでしょうか？

Bonuは、Go言語を使用したバックエンド技術、Reactを使用したアプリケーション開発、そしてIndustrial sensorsを使用しました。また、消費者向けのタブレットを使用し、産業用SENSORと接続することで、より柔軟で信頼性の高いシステムを実現しました。

新しい生産システムの利点は何ですか？

新しい生産システムは、従業員が注文を準備する際に紙のリストを読む必要がなく、タブレット上での直感的なインターフェースを使用できるため、より迅速で正確な作業が可能になります。また、システムのカスタマイズ性や柔軟性が高まり、生産ラインの改善や新しいサービスの導入が容易になります。

新しい生産システムの導入に直面した課題は何でしたか？

新しい生産システムの導入に直面した課題には、適切なセンサーの選定、APIの遅延、常にオンのスクリーンの管理などが含まれます。特に、センサーの選定では、高品質で安価な製品を見つけることが難しく、APIの遅延は画面の即時反映に影響を与え、スクリーンの管理ではMDMソフトウェアの複雑さとHTTPS接続の課題が必要に対応する必要がありました。

Bonuは今後どのような方向性を進める予定ですか？

Bonuは、生産ラインで収集されたデータを活用し、マージンの計算や購入チームの選択肢の最適化などに役立てることを目指しています。また、システムの信頼性を向上させるために、センサーの交換を迅速に行えるようにし、アプリのバージョンアップを自動化することも計画されています。

Bonuのフルカスタマイズ可能な注文システムはどのようなものですか？

Bonuのフルカスタマイズ可能な注文システムは、顧客が毎週提供される商品から自分の注文を完全にカスタマイズできるシステムです。顧客は木曜日に通知を受け取り、その中で注文をカスタマイズすることができます。そして、月曜日から注文が準備され、顧客に発送されます。

Bonuのサブスクリプションモデルの具体的な動作方法是どのようなですか？

Bonuのサブスクリプションモデルでは、顧客は毎週の注文を準備し、選択した商品が毎週配送されます。顧客はサブスクリプションにコミットする必要はありません。毎週、顧客は受け取るバスケットを選択し、スキップすることも可能です。また、いつでもサブスクリプションを一時停止することができます。

Outlines

00:00

📣 導入と現状の説明

MarkがCTOとして、Bonuの最後のトークでピッキング構造とプロセスの最適化について話す。Bonuは、食品ロスから良い製品を救うというミッションを持つフランスの会社。彼らは新しい生産ラインを設計し、デプロイする際に直面した課題について説明する。

05:01

🚀 新しい生産ラインの課題と改善

Markは、新しい生産ラインの課題について説明し、よりカスタマイズされた注文に対応する必要性とそれに伴う生産性の向上、製品のトレース性、エラーの低減が求められることについて話す。また、システムの設計原則として、紙レス、トレーニング不要、果物や野菜のリストの変更への適応性、低コストのテクノロジー選択を重視した理由も説明する。

10:04

🔄 技術的な取り組みと生産ラインのアーキテクチャ

Markは、産業用センサーから見かけないプロペティを除く方法を探求し、オープンプロトコルを使用するセンサーを見つけたことを説明する。彼は、生産ラインのシンプルなアーキテクチャと、GoとReactを使用して開発されたバックエンドエンジン、およびプロダクトラインに接続されたコマースプラットフォームについても話す。

15:05

🎥 生産ラインの実際の運用と改善点

Markは、生産ラインの実際の運用方法と、リアルタイムでのセンサーイベント記録による分析、そしてUXの改善について説明する。また、センサーの選択、APIの遅延、デプロイメントの複雑さなどの課題と、それらに対処する方法についても話す。

20:06

📈 次のステップとデータの活用

Markは、生産ラインのデータを活用してマージンを計算し、購入チームの選択を支援する計画を説明する。また、システムの信頼性の向上、アプリのバージョン管理、およびシステム全体の監視の強化が今後の課題となると話す。

25:14

🗓 プロジェクトの開発期間と費用

Markは、プロジェクトの設計開始からセンサーの発見まで1年かかりました。開発期間はQ1からQ3までで、センサー1個あたりの費用は約600ユーロと述べる。また、生産ライン上の従業員数と、その効率についても話す。

30:15

🚚 出荷プロセスと今後の展開

Markは、入荷プロセスの効率性と、出荷時に使用される特殊なアプリケーションについて説明する。彼は、日々の注文数と、アプリケーションのストレステストについても話す。さらに、Bonuが販売する商品の特徴と、将来的にアプリやバックエンドに追加したい機能についても触れる。

35:17

🙌 終わりの致し方

トークの最後に、Markが参加者に対して感謝の言葉を述べ、質問があれば呼びかけることを忘れずに述べる。

Keywords

💡production line

生産ラインは、このビデオの主題であり、製品を効率的に製造するためのプロセスと設備を指します。ビデオでは、新しい生産ラインの設計と導入に直面した課題について説明し、 Sensors、コンベア、タブレットなどを使用して、注文を迅速かつ正確に準備する方法についても触れています。

💡sensors

センサーは、物体の存在や位置を検出するために使用されるデバイスです。ビデオでは、生産ラインでボックスが移動するたびにセンサーをスキャンし、その情報をリアルタイムで表示することで、生産効率を向上させる方法が説明されています。

💡customization

カスタマイズは、顧客が自分のニーズに合わせて商品やサービスを調整するプロセスです。ビデオでは、フルカスタマイズされた注文を可能にするために導入された新しい生産ラインシステムと、顧客が毎週受け取る果物や野菜の選択をカスタマイズできる方法について説明されています。

💡barcode scanning

バーコードスキャンは、バーコードから情報を読み取るプロセスで、在庫管理や販売追踪などに広く使用されます。ビデオでは、生産ラインでのバーコードスキャンを使用して、ボックスの内容を追跡し、正確な注文を確保する方法が説明されています。

💡e-commerce platform

電子商取引プラットフォームは、オンラインでの商品の販売や購入を可能にするシステムです。ビデオでは、生産ラインと電子商取引プラットフォームとの連携が重要であり、注文の準備や顧客とのコミュニケーションに役立っています。

💡real-time data

リアルタイムデータは、発生する瞬間からすぐに利用可能なデータです。ビデオでは、生産ラインでのリアルタイムデータの使用が重要であり、生産効率の向上や注文エラーの減少に役立っています。

💡traceability

トラセビリティは、商品や素材の起源や流通ルートを追跡する能力です。ビデオでは、生産ラインでのトラセビリティが重要であり、どの顧客にどの商品が販売されたかを追跡することが必要です。

💡digital tools

デジタルツールは、コンピューターやネットワークを使用してタスクを自動化または効率化するためのソフトウェアアプリケーションです。ビデオでは、デジタルツールが生産ラインの効率化にどのように役立つかを説明しており、従業員がタブレットを使用して注文を準備する方法が示されています。

💡subscription model

サブスクリプションモデルは、定期的にサービスや商品を受け取るために顧客が料金を支払うビジネスモデルです。ビデオでは、サブスクリプションモデルが使用されており、顧客は毎週果物や野菜の配送を受け取ることができます。

💡customer experience

顧客体験は、顧客が製品やサービスを利用する過程で得られる感情的または知覚的経験です。ビデオでは、新しい生産ラインシステムが顧客体験を向上させる方法が重要であり、カスタマイズ可能で正確な注文を提供することで顧客満足度を高めることが目的されています。

Highlights

Introduction to the new production line and its challenges in designing and deploying the system.

The average time spent on grocery shopping is 60 to 90 minutes, while warehouses have to prepare an order every 10 seconds.

Bonu's mission is to save good products from waste, focusing on fruits and vegetables that are considered imperfect but still perfectly good.

Bonu started selling grocery products and expanded to hygiene, housekeeping products, and pet food.

The company operates on a subscription model, selling directly to consumers in France and Spain.

Bonu's largest warehouse near Paris prepares over 10,000 weekly orders, a number expected to double in the next few months.

The old production system was basic and efficient but lacked digital tools, leading to the need for a new system to handle increased order volumes.

The new system aims for higher productivity, better traceability, and less error-prone order fulfillment.

The design principles for the new system include being paperless, requiring no training, and adapting to changing lists of fruits and vegetables.

Technology choices for the new system prioritize reliability, adaptability, and leveraging existing skills within the tech team.

The new production line uses an engine based on Go, the same technology used for Bonu's e-commerce platform.

The production line includes sensors, a conveyor belt, and tablets displaying what products to put in each order.

The system allows for full customization of orders, a shift from a standardized selection of fruits and vegetables to a completely personalized order.

Challenges faced during the development included the choice of sensors, network and API latency, and managing always-on screens.

The new system records every sensor event, providing data on handling times for different products, which can help calculate margins and inform purchasing decisions.

The development of the new production line started a year ago, with a focus on launching the full customization offer in Q4 of the current year.

The cost of one sensor is around 600, with additional expenses for wiring and electrical work.

The production line can handle more than 2,000 orders per day, with the main operational limit being the pace of refilling the line and the number of employees.

Bonu sells products that cannot be sold in regular shops due to appearance or size, not because of regulations.

The company operates on a subscription model with no commitment, allowing customers to choose their orders weekly and add extra items as needed.

Transcripts

hi everyone um so welcome to our next

and final talk um I'm have to welcome

Mark you CTO at B AO um and you us about

something about um basically um

optimizing your picking structure your

picking process exactly um yeah and uh

this is then I would like to hand over

to you and uh yeah us tell us a bit

about what you do with sensors and stuff

thank you very much hello everyone so

I'm Mark I'm been you I'm super happy to

be here today to talk to you about our

new production line and explain a little

bit the challenges we we face when

designing and deploying that new system

and I thought with a very simple

question how much time do you guys spend

winkly on grocery shopping and the

answer to that question is roughly 60 to

90 minutes just to fill your baskets on

average

uh depending on the like Europe

countries and unfortunately we have like

uh much less time to uh fully order at

the warehouse and we have to uh prepare

an order roughly every 10 seconds uh at

the warehouse and this number is going

uh has to go down like quite a lot in

the next few uh months because our

volume of order is like going up thanks

to

that uh and so we had a we have a lot of

improvements to do to our own current

production system but let me give you a

little bit of context about uh bonu so

bonu is so it's a French established

company and our mission is to save good

products uh from waste and by good

products I mean fruits and vegetables

which are the the main uh products are

the main product that we we sell we

started with them uh then after a year

and a half we started selling grocery

products and now we also have like

hygiene and housekeeping products and

also pet food and all these products

that we sell uh are s because they

wouldn't have been sold in a regular

shops for many different reasons for

example uh concerning uh foods and

vegetables you see shaping carrot here

yeah most of our uh TR vegetables are

sold on V because they have the wrong

caliber or wrong shape or wrong color

you name it but they are perfectly uh

good products and same goes with like

Grocery and other kind of prod save they

are like some C with reason so that they

cannot be sold in regular shops so we

buy them directly from producers and

sell them to a subscription uh uh plan

to our customer so we sell directly to

consumers in France and Spain so so we

launched three years ago we are present

in six major French cities and think

five sorry five major Spanish cities we

have over 20,000

customers um and we prepare more than

10,000 weekly orders in our largest

Warehouse which is nor near near Paris

and uh that number that number is like

super to double in the next few months

the number of orders we prepare in our

largest warehouse and that's why we have

to add to our production system uh so

that's our current like it's the old

production system that we had before the

current project so it's a very basic one

but quite efficient I mean is no uh

digit digital tool at all uh basically

you have both the customers all in the

conveyor belt here uh all the fruit and

vegetables are just behind on the orange

shelves and here you got a a sheet of

paper close that's the sheet of paper

that indicates what's going to be put in

the in the baskets that's the order

content and I took an example of well

the highest complexity that we may have

with the current system so customer has

like some potato carrots didn't want

cucumber so they remove this one replace

by sweet potatoes and you have two kinds

of grocery products pasta on honey um

and that's the most complex order we can

do now like you can replace one fruit or

or one vegetable and that's it uh more

customization would be like too complex

with that system because the employees

have to read uh the paper during the

order preparation and if the orders are

too different from one another it's like

too complex and we would we have a lot

of uh errors in the order

fulfillment

um so we have uh two new uh constraints

uh that will put pressure on that system

first one is we have like higher volumes

we we have a lot of like new customers

that we gain every week so we have a few

more thousand orders per week in that

that warehouse before the end of the

year um and we hope also have a new

offer that we would like to launch uh

that we will launch in October uh

actually we launch today the offer for

the the employees so that's really the

launch of that project uh and it's going

to be like public for the like regular

cers starting in two weeks and that's

what we call the full customization that

means we're going to shift from a system

when you have a selection of fruits and

vegetables which are almost the same for

everyone to a system you can completely

customize your order based on the

product we will have both for that

specific week that means on Thursday

you're going to have a notification

these are the products that are

available like for you to order you have

like two to three days three days

actually to customize your order and

then uh we prepare the order starting on

the Monday and ship to customers um but

that means when we will have launch that

offer every order is going to be

different and we cannot rely on the

current prediction system to uh F all

the orders the new uh requirements on

the system are a higher productivity

because we have to go from like $10,000

per week to probably the double in that

warehouse only um we have to have a DAT

tracability of the which product we we

sell to each customers because with

higher volumes we have like several

producers for only apples for example

only carrots you have to know with what

carrots did you have in orders for like

for

reason uh and we have to have a system

which is like less error

prone uh to cope with the complexity of

the orders so the design principles that

we had for that project is uh the system

shouldn't be paper based uh so that uh

the employees don't have to like read

the order to read sheet of paper to by

the order um the system shouldn't

require any training uh because we have

a lot of like new employees is coming

every week and that should be like able

to prepare an order like immediately or

after a few minutes of training not more

than that and the system should adapt

itself to a changing list of fruits and

vegetables we sell different fruits and

vegetables every day and we have to

reconfigure the production line uh

probably every day also um and also we

should be able to deploy that produ

neware houses without huge capex uh cost

and that means we have to be like uh

very careful the choice of technology

that we're going to deploy so we have

three um principles for the tech choices

we make um the first one is we didn't

want any V lock uh because we want to be

able to adapt the system completely to

the way we prepare orders or to uh the

offer we make we make to our customers

or our warehouse trategy or you name it

so we going to be like super uh capable

of adapt the system and change it

um we wanted to have like a super

reliable system because we are talking

about production and all incidence are

like super expensive it could stop the

production light so we're GNA like favor

boring Tech to compared to F Tech um and

the last principle is we

also uh wanted

to um

maximize the the use of the skills will

have in the tech team and not having

having to learn like new languages or

new technologies or new platforms so

that means less Tech better than more

Tech uh and try to keep our Tech stch as

rational as

possible so that being said if you look

at the

classic architecture of a production

line uh in a factory or in a warehouse

usually you have like industrial

senses uh which communicate with an

automation system using a an industrial

bus for example modbus which is the one

of the most

well-known uh the automation system is

like usually proprietary uh and then the

automation system displays informations

on on screens R screens uh using also

some kind of prary protocol and this

system communicates with either ANP or

eCommerce platform to be link to your

with your business and for

example what sorry I click on stop y

yeah so some example of a production

line is like this one that's a big to

light system pretty common in e-commerce

companies for example uh that this has a

lot of proprietary hardware and a lot of

proprietary systems not specifically

what we wanted uh so we thought about

how we could get of the proprietary part

of that

architecture um and the thing is so

industrial

senses there are I mean there's no way

you can avoid them you need reliability

you need to have for example high

frequencies point scannings or you need

to have like super reliable systems so

there's no way you can avoid that but

the others we thought maybe we could

like replace the automation set with a

regular uh server that we could develop

using same Tech that we use for our web

servers our API servers and also for the

uh rug screens the public replace them

with consumer tablets uh and develop uh

like a human interface on this tablet

using classical mobile tech um but the

issue now is how we can connect to the

industrial sensors uh and we had to find

sensors that don't use like proprietary

or industrial protocol but regular cpip

protocols uh more on that later but we

were able to find them and we had a

super well think simple architecture for

our production line basically we use an

engine which is based on goong which is

our backend technology that we use for

our e-commerce platform with same

developers who can work on our

e-commerce platform or uh at the

warehouse on our production line

engine um we the that engine connects so

to the Commerce platform and server that

we made to configure that Prof l so all

are based on the same uh technology

that's goong and SQL and that's it um

and

the applications we developed uh on the

production line is using react GS that's

the same Tech that we use on our

consumer app and also they have those

industrial back

region that's the complete setup that

we' made uh you've got the conveyor

built uh that you've seen uh sooner

that's exactly the same it hasn't

changed we added some sensors that you

can see in green uh you still have the

orange shelf and vegetables on the top

got tablets that indicates what product

you should put in each order uh was like

looks like looks pretty easy uh but

let's show it in live let's slides and

more

videos so that's the first uh that's the

beginning of the prodction line uh

what's new with that prod line is we

need to uh generate a label that we can

to stick on the on

the the baskets and the all the baskets

going be put in large plastic box which

are easy to roll on the conveyor and

which are easy to be scanned by the

sensors so first thing is we're going to

put the on the table print the

label and that's going to be the south

of the jary of that box so we put here

up it's scan and you have like

automatically delay all that that it's

magic that's all the label so in each

box you're going to have three customers

orders and then that's

the that's real work when we are on the

production line

so youve got

the it goes so you slide the box here

and you see it changes immediately on

the screen on top we're going to do it

again uh and you've seen all the red

lights

here sens put on

the so that's the prediction line at

work

and that's what's displayed on the

screen that's a react basic react app

that we deployed and it's what's nice

with that approach is super easy to

change the ux and deploy it remotly uh

if we have like I don't know people that

have issues know with the colors of the

displays uh so that their work is easier

um so here you've got the three orders

which are in the bag uh in in the Box

sorry and here for for example this

order we got four different fruits and

vegetables which are possible so each

employee can manage four different

fruits and and vegetables and these are

the places on the shelves in front of

him but that means here for that order

have to take four bags on the left

bottom shelves two two bags on the right

bottom shelves and two on the top shelf

and nothing from the vbl it's super easy

I mean you you could already come to our

house and pick orders you know how it

works that that's was the the aim and

the color

corresponds to the quantity uh for

example uh the cauliflowers it's red

because it's double there are two pieces

the walnuts they're coming like two bags

together here we've got the double St

spread so really it's super easy to uh

know what you should put in each order

um and that's the employee screen

and that's just to show you what's going

to happen uh under the hood so on the

left you're going to have an employee

screen which is on the second station of

the production L not the first one

that's second just show you and here

you've got the dashboard which is

display in the warehouse shows what's

happening on the production

line so we're going to like simulate

this fact that we put the Box on the on

the position line so it will appear over

here on B1 the first station then move

on B2 it's going to appear appear on the

left screen it's second screen okay now

it starts the Box arrive here on B1 and

you're going to see here we have two

2,000 orders to prepare and it's going

to change to one that's the first order

got the three order numbers and the

content here now we move to be to that's

new screen for the second stage and then

another box and there it goes that's the

life of prodiction line so you can see

uh remotely what's happening on the

prodiction line a very nice tool to uh

to manage also remotely what's happening

what do in the

warehouse this is done in react also

super easy to uh change and to

redeploy

and these are the two last videos I

actually uh colleague of mine sent me

those videos just like an hour ago

because today is the day which we

launched that production line actually

so it's really dday for Ben Bono and I'm

not here I'm like I'm hidden in Berlin

where the team is like launching

everything um so that's the production L

really is the first orders we have

prepared today

that system you see the three orders in

each B in each box uh on the screen on

top uh very short videos and the second

one you see a little bit more what's

happening so we put the baskets in the

box that's what we saw that's a complete

view of the production lines with all

the screens which are being used right

now so there are up to 13

screens that mean certain employees and

so what what did we what what challenges

did we face and what lessons did we

learn during that

joury uh the first challenge was uh the

choice of sensus and it was like quite

difficult to find the right one the

criteria where

where like industrial performance that

means

highest barcode scanning and a high

frequency that system is like super

reliable and uh and that you have like

instant change when you put the Box on

the sensor um we needed an EET

connection on the full cpip

compatibility and an afford a affordable

price and that is like super difficult I

mean industrial sensors usually are not

uh fully ethernet compatible I mean untr

level ones uh so we are talking about

highend senses which may be super

expensive so we were able to find like

sense from Kon the Japanese manufacturer

which were like in our price range but

that was really a major major issue at

the beginning of the project um Second

Challenge was uh Network on API latency

is really really matter if you want to

have like an instant change of the

screen

and while talking about like a

production line where the the

productivity you want is like very high

there is no way you can like wait half

of a second or even a quarter of a

seconds of the display on the screens

but that means all the other contents is

stored locally in the production L

engine so we have to synchronize and

that actually adds quite a lot of

complexity on the project uh because the

the production engine synchronizes

itself with the e-commerce platform that

when you scan a box I mean the auto

content is already there and in less

than 100 100 milliseconds the screen

change that's really key for the of the

project the third challenge is up

deployment and always on screens are

really difficult to manage uh I mean we

we made that tradeoff not to have like

industrial screens which are much more

expensive so so well we are in the

consumer world and it's much more

difficult to control completely what

happened on stream so we use MDM

software to deploy remotely so mobile

device management software to uh control

the

tablets but MDM software by itself adds

a layer of complexity usually difficult

to manage or difficult to understand

because they are packed with a lot of

features uh so I mean deploying a new

version is not that easy even now that's

really an area of improvement for us

still uh this we hadn't forcing at

all uh it's quite difficult to have like

htps Connection in a Clos

network uh and on consumer tablets and

but is mandatory for example we thought

we could like do without htps that would

be nice but the MDM software prevented

us didn't want us to connect to deploy

web apps which are not served over htps

so we are in the clo networks where

there are no threat issue at all but we

had to deploy htps but then when you're

in that closed Network how do you

validate certificate and stuff a little

bit of engineering there and lots of few

days I can say uh to solve that that we

hav't Clen really how how long does it

take you to uh like the HPS uh quite few

days I guess three to four days to like

find a walk around first we we wanted to

find a walk around uh trying to use the

MDM software without using hgps but that

didn't work so finally we had have to

find a way to validate

certificate I mean essal

certificate uh that would be used in the

warehouse but not on the public IP

address quite quite

complex uh so now what's next we

launched today so that's a good a huge

my for us thank you uh and what's next

um we have to leverage the data that

that production line produce and that

actually there a lot of opportunity

there so we we every sensor event is

recorded so we know by the milliseconds

how much time it takes uh for I don't

know uh handling the watermelon or

handing a cauliflower and that would

help us a lot to to U uh calculate the

margin that we have in specific order

and to have the purchasing team in their

choices of fruits of vegetables because

we know the associated cost while

preparing the order um and also we have

a lot of uh Improvement still in the

space of reliability so we should be

able to replace sensor tabl light in

less than five minutes without the

intervention of someone from my team are

not there yet at all so yeah lot lot of

work to do there like uh

producing writing sorry processes having

spare parts preconfig kind of stuff

we're not there yet um deploy singly a

new version of the app as I said it's

quite complex uh to have to deploy that

version and you have to understand that

these screens are always on the app has

to replace itself you shouldn't like

relaunch it it has to be like completely

seamless once again it's that done and

and we have to to add a lot of

monitoring uh on the systems everywhere

to be able to prevent and detect

failures before they happen uh but

that's next

step and I guess that's all I had to say

for today I got I got the last video

that's another Warehouse in Madrid uh

that we deployed also and it's a

different system doesn't use the big

plastic boxes but it's a system with one

other at a time in a in a out

books and that's it for the presentation

if you have any

questions what was the development time

in total like when did you start and so

we started designing theem exactly year

ago okay because we thought that we so

we wanted to launch that offer the full

customization

during Q4 this year uh so we had to

prepare that project before because

without the light there's no way we can

like launch that full customization

project but we started designing it a

year ago I guess we found the

sensors in q1 N of q1 and then was

like mainly development from q1 to the

end of Q3

Withers

prority how much is how expensive is One

sensor so One sensor it's it's around

600

okay but then the wiring also with

proprietary you there's a lot of

electrical work I mean when you enter

that all of

industrial uh surface like I mean all

the s are um you need to have a to

provide 24 volt current so it's like

super

specific uh I mean Super common in

automation world and in the automatic

world for ACT world but not really our

space like web

developers and for the biggest well you

said there are 303 people working in the

production line so we have 13 people on

the production line and in the in that

largest Warehouse we have two prediction

light the two prediction light are

equipped so that's 26 employees that may

work at the same time okay so so that's

that makes like a tool of 70 different

products so that 70 different vegetables

and no because you have same produ on

the two production NES One prodction n

should be able to to to have a complete

order but you've got four products per

employee you can go up to 42% vegetables

okay

yeah with the Box they follow the box to

the end or no no no no no they move the

Box yeah yeah yeah that's I mean one

employee is in charge of like four

different FR vegetables and if it's uh

for example a vegetable which is with a

very high demand I don't know potato for

example that's he had only potato in

front of him or her uh and we do only

potatoes because the other issue that we

didn't see there how do you feel the

production L on vegetables that's under

the roof the other side of the prodction

line you've got people uh you know

scanning boxes for traceability and

putting back like new fruits and

vegetables uh on top of the production

line so the production managers for

example um instructing employees said

okay you you are um responsible for

refilling and and Stu exactly exactly

youing and they refal as fast as they

can that's the instruction basically the

prediction has to preder has to be full

all the

time and the W's basically fridge I saw

everyone was check it yeah yeah that was

one of the challenge when we work on

that project because for one our jacket

a few times a long days very cold yeah

it's five degrees I guess or six degrees

okay for the conservation of

the one of the CH you have is to update

the app is it because the screen is 247

always on or or yeah yeah it's 247 okay

but we don't work uh yet uh during we

have only one shift okay um but right

now we so the the tablets are always on

and we'll find a way to like switch them

off automatically during the night it's

not not done yet unless yeah exactly you

don't want that and instead using a web

in wait we well we use a web app so it

can it can refresh uh itself

uh so that's one way of addressing it

but sometimes uh you need to uh so the

the thing is so we we uh the app is

served by the local engine also and we

deplay both at the same time and uh but

yeah one of the ways to do so exactly to

refresh we without the

MDM it's a PM and we mm for

that

okay

one oh that's because um that's the way

we uh ship them uh so around Paris we

have our own trucks and these box fit

perfectly into the truck and the truck

is filled up to the roof and that's a

way to maximize the load in the track

and the the way the order the ERS are

ordered on the prodution line is the I

mean first we're going to put the orders

that have to be in the back of the truck

so we compute the trips uh before before

the day uh for example we have know

2,000 orders to deliver and they're all

in a specific order to be loaded at the

back of the truck if they are at the end

of the

trip a special application

there oh yeah yeah we use like third by

Soft for that it's like super complex

softwares because it has also to take

into account like traffic jams and I

know road which block on pares and

stuff how many orders um will you be

able to to fulfill per day or per

whatever that's a very good question we

don't know yet okay yeah really we don't

know yet uh so we are uh able to uh to

do more than 2,000 orders

uh on one day on one prediction line so

already

already have more than that because so

the the right now the customers order

one week and receive the order the week

after so the shortest we've got is you

you prepare you you make your order on

on Sunday and we deliver on on

Tuesday then you can prepare them on the

day oh yeah yeah we

yeah and did did you do any stress tests

uh for for your application because you

said okay we don't know until

when but the system can go like

much yeah yeah has no like Tech limit

that the limit is in the number of

employees you can have on production

line and

specifically the the pace at which you

can refill the production lineation isue

rather than and and and the the room it

takes to have all the people moving with

boxes uh and filling the prod up that's

the the real issue technically it's all

sorted but operationally yeah yeah

yeah question you said you you sell

vegetables and product that are not they

cannot be sold in other in other shops

is it because other shop just don't want

to sell them or you need to have some

special permission Oh no just they don't

want so because they don't look yeah

yeah

exactly exactly and that's between five

to 10% of the prediction of Which F it's

pretty hard

yes oh so there regulations for for

example

like yeah Qs which are too small we

don't have the right to sell them so we

don't have the right

either but miss shapen carrots I mean

there are no

regulations but yeah people are in

Supermarket you just don't buy

them a subscription model customer yeah

you pay every month no no so it's it's

per week but I mean it's a subscription

with no commitment that means every

Thursday you receive the fruits and the

list of fruits and

vegetables uh and you can choose either

to uh keep the basket or skip it no and

you can suspend any time really no

commitment but I mean there is no check

out you entered your like credit card

when you subscribe and then each week

choose if you want to have the baset or

not and you can add CL

uh basket so we have got another

production system for the all the all

the grocery

products and the part from monitoring

what features do you want to launch in

the future so are are is do you have a

backl already what features yeah yeah

and for like the app and backend and

stuff oh yeah so we line there things

you would like to add is for example if

you want to add like gifts or incentives

to specific customers for example

loyalty programs that's really the kind

of thing we could add very easily on the

screen so I don't know it's your it's

order and have a specific gifts for that

uh that's super easy to like display on

the screen add that gift for example B

to bag in the order super easy to with

that system and that we couldn't do

before I mean there a lot of marketing

SCH that we can now develop building

that

tool okay cool cool thank

you