Andrei Tara 

 

Embracing.Tech - Case Study #001

Leveraging an Agency to Build an AI Product

Table of Contents

1. Translyvania Labs - Software Development Agency

1.1. How did it all begin?

1.2. What services do you offer?

1.3. What is Fuzzy Logic?

1.4. Why use Fuzzy Logic?

 

2. Openfabric - Product made by Translyvania Labs

2.1. The current state of centralized AI platforms

2.2. What is Openfabric?

2.3. What do stakeholders gain from the platform?

2.4. How to attack service consumer stakeholders?

2.5. Why not take VC Funding?

 

3. Buypal - Product made by Translyvania Labs

3.1. What is Buypal?

3.2. Why did the product fail?

 

4. General Advice

4.1. How to get started with AI?

4.2. How to Embrace AI faster?

In order to have access to even more interesting projects, it was necessary to go to the location of the clients. That's why they opened...

Once Upon a Time in Romania...

Research and Development

The journey with Translyvania Labs began in Romania by doing development. With time, Andrei started taking on harder and complex projects.

New Companies

Research and Development

Business Development

Sales and Consultancy

What services Translyavnia Lab offers?

Enterprise Architecture

Blockchain

Artificial Intelligence

Fuzzy Logic Systems

Enterprise Architecture

Blockchain

Artificial Intelligence

Fuzzy Logic Systems

Fuzzy?

 

Type of logic in which the truth value may be any real number between 0 and 1 both inclusive. It handles situations where the truth may range between completely true and completely false." [1]

GOLD NUGGET

Fuzzy Logic

What did you do with Fuzzy Logic?

Client

Industrial Factory

Problem

Scheduling different types of machines

Solution

Fuzzy logic for the planning and allocating resources in the machines

Everyone is pushing for Deep Learning, why use fuzzy logic? 😉

"Sometimes, using simple tools brings you better results than the complex and very complicated ones."

Everyone is pushing for Deep Learning, why use fuzzy logic? 😉

Occam's Razor

"If you have two equally likely solutions to a problem, choose the simplest." [1]

GOLD NUGGET

"Sometimes, using simple tools brings you better results than the complex and very complicated ones."

Everyone is pushing for Deep Learning, why use fuzzy logic? 😉

End-user doesn't care what's inside, they just want the solution to work. So, Andrei reconsidered the...

Occam's Razor

"If you have two equally likely solutions to a problem, choose the simplest." [1]

GOLD NUGGET

"Sometimes, using simple tools brings you better results than the complex and very complicated ones."

Current Approach

Centralized AI platforms are dominating the landscape.

 

 

 

 

 

 

By combining smart decision making algorithms with data, they enable developers to create advanced solutions.

Current Approach

Centralized AI platforms are dominating the landscape.

 

 

 

 

 

 

By combining smart decision making algorithms with data, they enable developers to create advanced solutions.

Service Consumer

AI Innovator

Data Provider

Infrastructure Provider

Payment

Algorithms

Datasets

Environment

Per-Execution Payment

Per-Execution Payment

Per-Download Payment

Solution

Stakeholder

Offers

Receives

Service Consumer - Hardest to find

Data Provider - Very interested 

Infrastructure Provider - Very interested

Stakeholder Research

AI Innovator - Interested

Service Consumer - Hardest to find

Data Provider - Very interested 

Infrastructure Provider - Very interested

Stakeholder Research

AI Innovator - Interested

"Marketplaces are so hard to get rolling that you should expect to take heroic measures at first." [1]

GOLD NUGGET

How to attract them?

1) Create a community

2) Bring awareness about the solution

3) Make the platform beneficial for non-technical people

Service Consumers

Who are they?

Current Hypothesis - Small and medium-sized enterprises who need AI services but they can't afford to hire internally

Service Consumer

Payment

Solution

Stakeholder

Offers

Receives

Double-edged sword.

Openfabric and VC funding?

It propels you at the beginning, but VCs
have big expectations at the end. You have to justify that with exponential growth and that can in the end ruin the product. Also, potentially do decisions that are not the best decision for your product, but which bring more revenue in the short-term.
Besides Openfabric, there is another startup attempt Andrei had?

BuyPal -  Conversational Chatbot Targeted at the Retail Industry

Problem

Invested a lot of time in the technical solution. Failing to focus on marketing and sales. Only sold a few licenses for a recommendation system.

Goal

Create an NLP engine that is capable to extract multiple intents from the same phrase.   

  • People wanted more to discuss with a human.  We concluded that it's a bit too early for the end-user to adopt this way of interacting.
  • It will take a bit of time for the end-user to accommodate with the idea that they might interact with the technology as with a human being.

"About three-quarters of people (73%) say they are unlikely to trust an AI-powered voice assistant to make simple calls for them correctly. " [1]

GOLD NUGGET

Why Andrei dropped BuyPal?

  • People wanted more to discuss with a human.  We concluded that it's a bit too early for the end-user to adopt this way of interacting.
  • It will take a bit of time for the end-user to accommodate with the idea that they might interact with the technology as with a human being.

"About three-quarters of people (73%) say they are unlikely to trust an AI-powered voice assistant to make simple calls for them correctly. " [1]

GOLD NUGGET

Why they dropped BuyPal?

How to get started with AI?

Build a product.

  • You get to take a look at everything from the user's perspective rather than just learning technology.
  • It's important to understand all the technical parts, but it's crucial to understand why and when to use them.
  • That can happen only by interacting with potential users.
In the end, you can go back to working with customers with all the experience that you've gained in building the product.

How to embrace AI faster?

  • We're talking right now through the Zoom application that is connecting to internet protocols that are under the hood.
  • We don't need to know about it and we can just use it. The same should happen for AI.


Make using AI technology much easier to use for non-technical people, hide the technical details.

You've reached the end! 👏