One-stop shop for Machine Intelligence, where you may find the best AI services for your task, evaluate them on your data and access any provider via the Single API.
Current Status
The product is operational, the first pilots are running, the first checks are received from clients.
Market
1. SME - to find the best price/quality deals on AI services and integrate with no effort
2. Large Enterprises - to match in-house development with 3rd party alternatives
Top-down:
- AI as a Service market $5.6Bn in 2016, to 13.1Bn in 2020
- Machine Translation market $0.4Bn in 2016, to 0.8Bn in 2020
- NLP/NLU market $4Bn in 2017, to $6.4Bn in 2020
- Image Recognition market $16Bn in 2016, to $32.5Bn in 2020
- Voice Analysis market $4.0Bn in 2016, to $12.1 in 2020
Problem or Opportunity
Cloud AI services is a deeply fragmented market where vendor-based discovery prevails. The main reasons are:
- lack of discovery services beyond search engines,
- non-uniform delivery among providers (lots of different APIs),
- expensive performance comparison (lots of contracts to sign and code to write to compare the performance),
- non-transparent pricing (lots of different service packages and pricing models).
All these results in up to 10x difference in pricing for the comparable services.
Solution (product or service)
For each service category, the product provides (1) a uniform endpoint that packages
services into equal units and regularly updated (2) quality score and (3) estimated unit pricing per provider. We buy services in bulk from providers and resell to our clients with 10-15% margin (at the price equal to what providers ask for direct integration). We take a small fee for using our Single API (5-10% of the whole spending). Also, we earn on evaluating AI providers on the clients' data.
Competitors
Big AI silos (IBM Watson, Microsoft Cognitive Services, etc)
Advantages or differentiators
About the same price as from direct integrations with service providers, up-to-date information on the best deals on the market, ability to switch instantly to any provider, just one integration to access all the providers.
Finance
1. Service evaluation - 80% contribution margin
2. Single API - 25% contribution margin
The pricing is currently custom, we aim to have $1500/momth of net revenues per client, $6M net revenue in 2018, $24M net revenue in 2019
Business model
We take a service fee to evaluate all AI providers on our clients' data, then we resell AI Cloud Services with ~24% margin (mostly goes from the discount owe have from service providers)
Money will be spent on
Sales (~1000 potential clients in the pipeline), adding more AI verticals (already requested by the existing pilot clients)