Ethereum is designed to be a 'Turing complete' state machine (a so-called general-purpose computer) that's also decentralized. Seen thus, we can think of smart contracts on Ethereum as the software that runs on the shared computer which, by the way, is known as the Ethereum Virtual Machine (EVM).

Developers write smart contracts in a programming language called Solidity, then deploy them to the shared computer. When you deploy a smart contract to the "computer" (the Ethereum network), what you're actually doing is registering a special address on the network. Anyone can interact with a deployed smart contract by sending ETH to the contract address. This triggers the code in the contract to be executed according to its logic. The code is executed automatically along the 'if this, then that' model just like any other programming language.

For example, a simple smart contract could be written such that when you send 1 ETH to it, your 1 ETH is divided up into 12 equal pieces, each of which is sent to a specified address at one-month intervals. This would effectively create something similar to a trust fund.

For use cases that exist entirely in the digital world, smart contracts allow people to bypass intermediaries altogether. For instance, in our above "trust fund" example, we've effectively eliminated the need for lawyers and managed escrow accounts. Instead, we simply send ETH to the contract and rest assured that it will be sent to the trust fund recipient exactly as the contract states and without any need for further input.

Read more: What are ERC-20 tokens?