Buzzmi, the world’s first platform that help individuals create a stronger, more meaningful relationships in business and daily life. Buzzmi aim to provide instant, location and preference based activities to help the modern generation group form their own decentralised community.
By using AI and neural networks technology, we encourage individuals initiate human interaction with effort making, experience and memory sharing together.
Current Status
pre-product, pre-revenue. We have been working on this for 6 months. We have 1 full time employee, her name is Jane and 5 part time employees. Our prototype will be ready in June.
Market
We are going to start a Vlog on youtube very soon to talk about how Asian female entrepreneurs (as we have two female founding members) who were not born in any western countries's career journey and the reason of ended up as an entrepreneur. Vlog is a new style and is very popular right now. We are currently working on a strategy to build vlogs not only on Youtube but also on Red & Tiktok. The first step could help us save budgets in terms of marketing. We plan to attract enough stable followers before delivering our product so those followers who can also become our potential users.
We've also agreed to a partnership with some youtubers and vlogs (through Angus & John's private connections), a Vlog with 500,000 subscribers, we are discussing with her to provide information & feature us sometimes in her videos in exchange for future awesome content from us. If it goes well, we plan to approach more famous vloggers.
Problem or Opportunity
We understand that in any type of relationships, real life interactions between human beings and the moments that they share together are the ultimate foundation upon which a strong and steady bond can grow. We should give others an opportunity to discover the potential between each other. This potential should be based on some effort making progress instead of an easy come and go way. We also understand users' needs in terms of the "other" person they are looking for. A cross matching with people from similar background, similar culture who can share similar value with them are those who can really touch someone's soul.
Solution (product or service)
Most of the Neural Networks (NN) currently are designed as pre-trained models to be used. The reason those NN used are pre-trained is because most of the time, a pre-trained model would work perfectly fine by itself or with some fine-tuning. Training of NN is an extremely computational intensive process, and could involve a lot of manual works as well, therefore, training fr om scratch is done only if it is absolutely necessary, or the result would give a strategic advantage to the solution.
Most of today’s NNs are standalone, individually perform a very specific functionality, producing very specific output. For example, recognise human speeches, identify different objects, perform language analysis, etc.. Individually they do not generate much business value and has zero to very little capability outside of their trained domain. To solve any real business problem would require more than NN.
The traditional way of realising such a solution is to program some logic to interpret the output of these NNs to obtain the desired output. This has the advantage of straight forward, easy to understand and can be done by any competent development team according to the business requirements. The disadvantage is it is hardcoded, fixed algorithm and prone to error if the input do not fit into the designed criteria.
In BUZZMI, we realise the solution in a novelly different way. BUZZMI employs multiple NNs to make sense of different input generated by the users such as speech, video, movements, etc.. However, we do not hard code the logic that produce the output required by the business solution. Instead, we employ a NN architecture to learn from these different output and give the predictions as the results. More than this, we treat these individual NNs like one giant NN and train it just like one. The approach we chose is similar to team building in a company wh ere individuals with different capabilities learn how to work together to boost productivity. This is exactly what we wanted to achieve, train a group of NN and produce results that are largely immune to stochastic errors and has the ability to continue to learn to adapt to new, unforeseen situations without breaking down catastrophically.
Competitors
We don't see exactly competitors in the market. However, we regard dating apps such as tinder, bumble, coffee meets bagel and meetup app such as Meetup, reservation apps such as booking.com, opentable.com as our competitors in certain degree.
Advantages or differentiators
We understand that in any type of relationships, real life interactions between human beings and the moments that they share together are the ultimate foundation upon which a strong and steady bond can grow. We should give others an opportunity to discover the potential between each other. This potential should be based on some effort making progress instead of an easy come and go way. We also understand users' needs in terms of the "other" person they are looking for. A cross matching with people fr om similar background, similar culture who can share similar value with them are those who can really touch someone's soul.
Most of the Neural Networks (NN) currently are designed as pre-trained models to be used. The reason those NN used are pre-trained is because most of the time, a pre-trained model would work perfectly fine by itself or with some fine-tuning. Training of NN is an extremely computational intensive process, and could involve a lot of manual works as well, therefore, training from scratch is done only if it is absolutely necessary, or the result would give a strategic advantage to the solution.
Most of today’s NNs are standalone, individually perform a very specific functionality, producing very specific output. For example, recognise human speeches, identify different objects, perform language analysis, etc.. Individually they do not generate much business value and has zero to very little capability outside of their trained domain. To solve any real business problem would require more than NN.
The traditional way of realising such a solution is to program some logic to interpret the output of these NNs to obtain the desired output. This has the advantage of straight forward, easy to understand and can be done by any competent development team according to the business requirements. The disadvantage is it is hardcoded, fixed algorithm and prone to error if the input do not fit into the designed criteria.
In BUZZMI, we realise the solution in a novelly different way. BUZZMI employs multiple NNs to make sense of different input generated by the users such as speech, video, movements, etc.. However, we do not hard code the logic that produce the output required by the business solution. Instead, we employ a NN architecture to learn from these different output and give the predictions as the results. More than this, we treat these individual NNs like one giant NN and train it just like one. The approach we chose is similar to team building in a company wh ere individuals with different capabilities learn how to work together to boost productivity. This is exactly what we wanted to achieve, train a group of NN and produce results that are largely immune to stochastic errors and has the ability to continue to learn to adapt to new, unforeseen situations without breaking down catastrophically.
Business model
Our revenue model is based on: B2C: Charge monthly user subscription fee by categorise users as Basic user, Premium user & Elite user with different product features
B2B2C (SaaS model): Direct users to different F&B, activities, membership spots, etc charge partners based on SAAS subscription
For the first 2 years, we plan to attract as many users as we can. We target to achieve 500k users in 2 years across Europe and APAC. These are the two markets we are very familiar with and experienced in.
