our machine learning technology
efficiency + happiness
Rethinking landscape design.
Livee uses computer vision and machine learning to reshape the landscaping design process. On average, landscape designers take one week to produce a proposal. We do it in 4 seconds.
Our technology.
1.SKU image auto-generation (computer vision model)
Our computer vision model compares thousands of SKUs (plant and pot combinations) and matches them within seconds.
2.Proposal recommendation system
Using the principles of matrix factorization, our algorithm analyzes and compares the features of the user and the product to produce the most accurate proposal.
3. Transformer-Based Language Models for Sentiment Analysis
We aim to implement a system to monitor and linking cross-dependencies between placed landscaping and it's effect on productivity, happiness, level of stress and turnover rates.
1.SKU image auto-generation (computer vision model)
Using digital images from cameras, videos and deep learning models, machines can identify and classify objects and then react' to what they see.
Our model analyzes thousands of plant and pot combinations.
This process takes several seconds, whereas for a human it takes one week.
The 'reaction' is the recommendation.
The computer vision model is used to produce the most case-accurate recommendation, comparing the stock and the client's input data i.e. the amount of workers and size of the office.
Our current AI systems achieve 85% accuracy with SKU recommendations.
Just like plants, our technology is growing.
We are training our system to consistently grow in accuracy. By evaluating SKU (stock keeping units) accuracy and validating the best outcomes, we are improving the efficacy of our algorithm exponentially. The more data it processes, the more accurate our computer vision model becomes.
Our system started at 30% in SKU generation.
We trained the system to achieve an 85% accuracy rate.
With the current trend in growth, we forecast our model to achieve 95% accuracy
Smarter proposals = +proposals and +business
2.Our proposal recommendation system
The best way to explain our machine-learning recommendation system is by thinking of Spotify or Netflix's algorithms that recommend the most appropriate music or movies, based on your data. If you tried to find new music without such a tool, it would take you a lot more time. We're applying this same concept to our proposal generator.
We use AI to generate proposals
Datasets validated by real designers.
Real landscape designers approve or disapprove past projects, teaching our system what to recommend.
Machine learning means constant improvement.
The crux of AI technology is improvement. Our system is constantly teaching itself to be more accurate and more efficient.
Saving time and cost.
Our proposal recommendation system not only widens the design options for our clients, but saves a huge amount of time.
Matrix factorization
Our recommendations use what is called matrix factorization — a class of collaborative filtering algorithms used in recommender systems. The system allows lower user interaction, by employing two underlying matrices. This recommendation method was made famous by Simon Funk in the Netflix Challenge.
A flexible recommendation system
Customize for your case.
We want to know about you, to give you the best product, for you. We ask you physical specifications for your future green space as well as the level of impact you want to provide. We generate tailored proposals for large scale spaces to home offices.
Opportunity cost
average time for a proposal (human) = 1 week

average weekly salary (US) for landscape designer = $1,100

average time for a proposal (Livee) = 4 seconds
More time = more proposals
3. Happiness impact + ecosystem (transformer-based language models for sentiment analysis)
We aim to use monitors, censors and data on turnover rates, sick leave and KPI achievement to formulate the tangible effect of landscaping on humans in the workplace. To do this, we will set up monitors 2-3 weeks before installing greenery and then make a comparative analysis.
The psychological benefits of greenery.
Our happiness + indexes happiness as something tangible that we can improve through landscaping. This is why this is so important.
Reduce stress
Lighten mood
Improve concentration
Increase productivity
Boost energy level
Promote social interaction
AI does not replace humans, it improves their capacity.
Machine learning is superseding human labor in areas such as design. But this doesn't mean that designers will no longer have jobs. It means that they can stop spending unnecessary time on routine tasks which our systems can do far more faster and efficiently. In turn, designers can process more leads and commit their time to more projects. Machine learning does not have to replace human labor, rather, it makes it more efficient.
happiness + efficiency
Countless studies have shown the positive psychological outcomes of plants and greenery in the workplace. But due to the logistical complexity of commercial landscaping and the significant design and administration costs, too many workplaces are not facilitating optimal working environments. We aim to smoothen this transition.