Forgot your password?

App Store “for those who need”: how a project that will include thousands of AI-startups is being created

Wednesday, May 6, 2020

The proverb says eat an elephant one bite at a time and it seems that a large and complicated project should be divided into many small parts. And this is the ultimate approach of Orchestra startup, which was launched this February — it represents an application constructor with marketplaces of ready-made solutions for business clients; it will contain thousands of widgets which were created and are functioning with the help of the Artificial Intelligence technology. Sergey Gluhota, the founder and the CEO of the project, told us how the startup idea emerged, how it is realized and what it should look like in a perfect-case-scenario.

Startup Jedi

We talk to startups and investors, you get the value.

Instead of an introduction

In a long-term perspective, Orchestra is a platform with thousands of different widgets where it is also possible to design your own application. Any person being a physical entity or a businessman, will be able to find the application for solving their specific problems and set it up for their needs. These applications and widgets are being designed by Orchestra startup team and in perspective, the business model prerequisites that marketplaces would be able to collaborate with thousands of independent developers, and download their applications to the platform. In such a way, they will save money on marketing, monetization, support as these functions will be provided by the platform. For these services, Orchestra will take the revenue fee, which will be obtained from selling widgets.

Orchestra startup
Click on the image to see Orchestra's profile on the Rocket DAO platform


Orchestra startup


- How did the idea to create the Orchestra startup appear?

- The idea came to my mind on February 16, 2020 as a case of a marketplace with ready-made solutions. It happened during Demo Day Demium Startups in Minsk. To be honest, I didn’t receive any support back then — the idea was named interesting but complicated. In fact, I have a clue why so.

During several weeks I had been developing this idea and when I finally formed the concept, I started teeming activity for its realization.

Now, in the project’s framework, we are doing a simple thing: we find solutions that are created by other developers and make deals of presenting their products on our marketplace. At this moment, there is no such a “box” where you can place and process any data from any field.

- What is the perfect working scenario — describe it step by step.

- My task is to create a maximum simple history of acquiring a result. In our case, a process from registration to a result for a person who knows what he wants, will take nothing more than two minutes.

You register on a website and indicate your credit card details. We don’t charge anything; it is the verification requirement. Then you enter the marketplace and search for the widget you need. After that, you “conditionally” buy it. It is “conditionally” because every widget has a test-fly — a trial period when the user doesn’t pay. The developer of this very widget sets the duration of this period.

Then you click “Constructor”, click “Create a program”, enter its name and add the widget you need into this program, adjust the widget and click “Start”.

If you need to do something once, you don’t pay at all. The platform allows everyone to use the service for 20 minutes per month for free. In such a way, we will have users who never pay anyway.

Every developer chooses the most suitable business model for their solution — we considered this moment in the marketplace. The widget can be used for free, it can be paid or for subscription, or it can be used for the number of inquiries, etc.

If the user needs to accomplish the task on a daily basis, he follows all the required settings and pays for the time when the program is actively used. Not for months between every given task but for the time when the program completes a specific task — it may be a minute, a second etc.

Orchestra startup

- According to your plan, there will be many various widgets on the marketplace. Can you give some examples?

- There are a few common widgets: Excel data import/export, Bitrix24 and SAP.

Other widgets usually refer to business analytics.

Such as dynamic analysis of currency price trends; cryptocurrency widgets; sales forecast for a week/month/year, income forecast; widget for creating an investment portfolio; an app that calculates the optimum price of a product; a widget which can help organise doings, trips, transportations, time-being — call a taxi, book tickets and hotel rooms; the application which helps to choose the most appropriate place for starting an offline business; the app for forecasting income of an online-shop… And these are only a few apps from those that are being developed for the project.

We also can skip the Emergency map — a service that helps foretell intensiveness in any region, country or continent. It is a map where any person can add his point of “intensiveness”: it can be anything that provokes panic or flurries inhabitants of a specific region — the shortage of the medicine at chemists, pharmacies, the lack of essential goods in stores, the increase in the crime incidence, the criminal situation, the spread of fires … This widget, based on different types of data (social, meteorological, data from emergency services), will predict the development of the situation and recommend that necessary measures be taken.

The widget learned based on the open-source 911 data in the US. I offered this project to the European AI Alliance and in general, if we manage to show it full-functioning, we will have a chance to win a €100 000 grant.

- What is the AI-function in the project?

If we look at the platform from the perspective of the final user, 98% existed on the marketplace widgets will be completely AI. If we look from the perspective of the platform itself, AI is everywhere there. Down to the fact that we analyse uploaded widgets so they work properly. Roughly speaking, the developer has uploaded a widget and we have shown all possible errors in the code — this is the job of AI.

One of the ideas we are planning to implement is to create a virtual helper like Jarvis in “Iron Man”. You will be able to communicate with him, discuss a wide range of questions, give tasks to solve and he will come up with possible solution schemes on his own and will show you the final one. Everything that a customer will need to do is to set a task and answer a few questions in case the task is not very clear.

We already have a developed virtual helper. He is able to speak and maintain a dialogue. Right now we are “teaching” him to understand tasks. For instance, he can currently set an alarm, add a note to the calendar and boil water in a kettle. Once we have a specific content of the marketplace, the helper will be published for public use. By that time he will be able to fulfill all the necessary tasks and will create apps — down to that we will train him based on user-cases of the people using the Orchestra app.


Orchestra startup

Project development

- How much time had you spent before the idea stage turned into the realization stage?

- The idea hit me on February 16. In April, we created the MVP. In such a way, it took a little bit more than two months from the idea to the product realization.

- You have your Roadmap planned for a year ahead — do you fall behind the schedule?

- No, we don’t. I have two cutoffs: the first 3 widgets and x10 — the first 30 widgets. Currently, there are 28 widgets, those we have designed and developed, originally our products.

- On the project’s page, it is stated that you need $200 000. At the same time, you have masterminded many widgets and each of them required some investments. Where did you get money for the development?

- Orchestra is still developing at my own expense.

Currently, my team and I are searching not only $200 000 but more money which will help us rapidly reach the Seed round with the valuation not less than $6 million. When people ask me how much money I need to “get off a running start”, I have strict calculations — for the first year, I need $2 million. But if I don’t have them, I know where I need to tighten belts to have a quick start.

To be honest, I’ve never done the company valuation from the point of view of what the price to sell it is. I’m convinced that the cost is higher than the project’s current valuation.

But there is an objective opinion that you need to lower the company price in the first round, so the investor who has invested in the project, would receive X in a few months. Then this person would later say that these guys made me at least x2 but in 2 months. I am pragmatic when it comes to this question.


Orchestra startup


- How are you going to promote your product?

- I once worked in consulting and I have a perfect role-model, it is how Apple was developing AppStore. When they launched it specifically for the iPhone, they did a very usual and simple thing — they advertised the application to the target audience in AppStore. And it was a bullseye: the number of downloads started to rise exponentially.

I’m going to do the same thing. I will look for a specific client for every widget. Until I don’t have at least 1,5 thousand widgets, I will advertise the platform to the target audience that needs these widgets.

There is also a great deal of work with AI-startups, developers with ready-made products and there are more than 7000 of them in our database, working with a partnership network. What is more, we are implementing such things as “scouts” — these are the people who get rewards for finding developers, startups and bringing them to our platform.

Now we are doing a great job of interviewing our potential customers and are gradually developing solutions to their problems.

- Who are your clients? Describe them.

- From my perspective, our client is a person with a business task. We choose several spheres where people will be in need of us and where people will eagerly pay money for widgets. These are financial management, sales, company management in general, trading. Representatives from these fields are the most easy-to-understand audience for us, so they will become our target audience from the very beginning.


Orchestra startup


- You have designed the project with three business models: revenue from selling widgets on the platform, revenue from the time spent for using specific widgets and revenue from specific requests. Which model do you consider as the most profitable one?

- From selling widgets.

In terms of the widgets selling, how much money the platform will make is a hard-to-answer question and nobody can foretell what will happen in the end. It’s because nobody knows how many applications will be and of what kind. In my scheme, I put in a pawn a minimum widget price — $10. However, each of our original widgets will cost not less than $79,99. It is better to make calculations pessimistically so in the result, there won’t be less than planned.

- How will you value the cost of each widget? Will it be set by the developer or the platform?

- By the developer himself. He chooses the business model for his widget. We “sew-up” three most used business models:

  1. A sale (one-time payment), when the person pays once and uses the widget during lifetime;

  2. A monthly subscription;

  3. Pay-per-request — when the person pays for launching an application or for requesting something important or, I would even say unique for open-market, data.

- Pay-per-Request — what is it? What are the requests that clients are going to pay for?

- Pay-per-Request is a purchase of data. Let’s assume that two companies are doing completely the same business. For example, those are networks of supermarkets. One of them is interested in how things are going on the market. They can hire a marketing company but their research will be a shot in the dark. It is possible to foretell sales (we have such a widget) but sometimes it is impossible to foretell everything. And then you can “drop” the anonymized data to the place where all market players drop their data. In return, you get access to the anonymized data of other players. However, if the company doesn’t want to share its data but wants to get the market data, it has to pay for the request and for the access. Of course, these data suppliers give permission to sell their data and receive 50% of the funds received.

Now, we have data about food products in the USA and Great Britain.


Orchestra startup


- Who are your competitors?

- We have three clear-cut competitors. Acumos AI, Bonseyes AI Marketplace, Wolfram Alpha are our main competitors as in fact they are giving users partially ready-made AI technology. Nevertheless, in most cases, it is impossible to customize it. If the user wants to do so, he needs to do it on his own and he should be a programmer or a wealthy man to do so. What is more important, neuro-web designers are nowhere to be found.

With it, we consider every player in the market as our competitor; we need to consider, research, study and take only the best from everyone who gives AI-product.

For the last year, our main competitors earned $12 billion. I really doubt that we will manage to earn at least 10% of what they have earned. Taking into account the fact that our model is more simple, the market is broader. In fact, we are creating AppStore but not for everyone, only for those “who need” it.

- Before launching Orchestra, what company/project experiences did you explore?

- Apple, their main competitors, experiences of projects which can potentially appear as widgets on our platform.

As soon as we have an idea, we check if it is already realized or not. If there is a good feature, we will analyze it and set it in the backlog. In general, we create a product which will be extremely simple to understand and it will be simple for everyone.

For example, there is a Delegator24 project which is not one of our competitors at the moment but it will become in the future. The concept of Delegator24 is that the user enters a chat and writes what he needs. Some people will do this work for him as soon as he pays them. Those could be completely different tasks — from finding a babysitter for a day to finding a top-manager. With the help of Jarvis, our virtual assistant, we want to do a lot but with the help of AI, not humans. By the way, this is the only project of its kind on the market and, in the future, it is a competitor to Orchestra.

- The market of which country/region you consider as a launching platform of the service? On which markets do you plan to scale afterwards?

- There is a specific audience of potential clients, and only a small part is from Russian-speaking countries. We are incorporated in the USA and Great Britain — these are the two most important markets for Orchestra. In the future, we are going to grow to Europe and Asia.

There is no problem integrating any language into the Orchestra interface. We have already developed a version for the Arabic-speaking countries that write from right to left. This is a world-wide project and we are preparing to capture each region, we are looking for local opinions of leaders, project managers, scouts.

I want the business to be smart. Orchestra makes business as smart as possible.


Orchestra startup


- The project has quite a big team: tell us, what will each of the team members do.

-On the website there is not the full team shown, because there are people that prefer to stay “incognito”. I can say that among my advisors there is a person from Sberbank who develops “neuron constructor” for inter-company usage.

In Orchestra, I have hired people that I know and with whom we have a similar background. Project founders are the people with more than 50 years of experience for three of them and cooperation with such brands as Apple, Shell, VW Group, IBM, Google, Oracle, D&G, etc. Dmitriy Beslutskiy — a company technical director and IT-outsourcer owner, and due to this fact we can use any developers. That is one of the reasons why the project develops so rapidly.

Sergey Glukhota

- The project has quite a big team: tell us, what will each of the team members do.

-On the website there is not the full team shown, because there are people that prefer to stay “incognito”. I can say that among my advisors there is a person from Sberbank who develops “neuron constructor” for inter-company usage.

In Orchestra, I have hired people that I know and with whom we have a similar background. Project founders are the people with more than 50 years of experience for three of them and cooperation with such brands as Apple, Shell, VW Group, IBM, Google, Oracle, D&G, etc. Dmitriy Beslutskiy — a company technical director and IT-outsourcer owner, and due to this fact we can use any developers. That is one of the reasons why the project develops so rapidly.

Egor Kaleynik is from Belarus, all his life he has been working in IT-project marketing. In this case, he is very useful as a person who has developed products and who knows many people who can help with product promotion.

I need people who I don’t need to tell what to do. I try to work only with professionals. I have to admit that I work with people who are smarter than me. I am proud of it, because if they are dumber than me, why do I need them?

For instance, Kate Barrow who deals with my clients’ issues. People more and more are using products already released; and this person communicates with the clients, “lives” with them, in the good sense of this word, of course.

- Which specialists are the most difficult to find?

- The hardest specialists to find are the people who share your passion for temporary bankruptcy.

Closing all gaps with money will never work. There must be people who quit their stable job and come to work on a project for an idea, sometimes for free. These people are the hardest to find. Because even those people who consider your project a “bomb” are not ready to give up comfort and stability for the sake of perspective.

- Have you found the right people?

- I guess, I did. I don’t have any troubles in communication or in taking hard decisions. Now, I don’t even have a strong division in who does what and for how long. Everyone is just doing their things. There is a nominal agreement, there are shares of the company, there are investors who have invested money. At the same time, people completely consciously, absolutely voluntarily delegated the authority to make decisions to me. Therefore, when I say that I have 100% of the company — the way it is. Trust is the most important resource that I and Orchestra have.


Stay tuned and don’t forget to follow us:





More From Startup Jedi

The founder of the Node PR communication agency about how to plan PR strategy.