AI in the workplace

Daniel Hulme, now CEO of Satalia founded the company in 2008 while he was studying his PhD. He realised that a lot of the cool kind of inventions, algorithms and optimisation techniques that were being developed in academia were never being commercialized or being put to any kind of good use in industry.

Satalia was born to almost bridge the gap between the inventions of academia and industry.

Firstly, they created a platform called The Solve Engine which is like a SaaS product but instead of being just software it was optimisation as a service. You could log on, submit problems, open the Cloud but this was happening in 2009/10, so it was before big data had exploded; in hinesight, they were very early to market.

Since then, the team have focused efforts on delivering artifcial intelligence (AI) / Machine Learning (ML) and data science consultancy for live scale clients. Their expertise lies in vehicle routing and vast scheduling with faily common optimisation solutions for customers in retail, professional consultancies, logistics and transport.

The team grew to 8 in 2015 where they mainly worked from cafe's and from here, there and everywhere. Over the past coulple of years they've had major growth with an army of 70 spread across London, Norwich and Europe. 

On top of delivering comprehensive customer projects in-house, Satalia are doing a lot of internal, organisational development projects; almost like mini audits. This involves carrying out a network analysis using AI, potentially, to analyse data to reveal decision making processes, leadership development and *interestingly* social relationships. And every once in a while, they'll support companies to implement AI into their business with confidence. 

A lot of businesses today have an issue with implementing AI because of all the fear-mongering created by the press.

Angela plays a big role in the cultural navigation of Satalia, challenging and championing internal processes that might hinder or help their company better foster innovation. 

With regards to AI specifically, Satalia states there are 4 major components:

1. Data collection > inputting data from lots of different sources
2. Data generated insights > involving ML and data science to understand what the world is (and this is where it ends for a lot of companies but there are 2 other key components) 
3. Decision making based on insights
4. Optimisation of data > data adapts and it adapts on its own in production and so, it’s extremely difficult to build systems that adapt on their own to changing environments in production

Components 1 and 2 are being rolled out internally at Satalia; for example...

1) Data collection > looking at all different types of communication data across a lot of different channels such as...

• Google Docs
• Slack 
• Email
• GitHub

2) Data generated insights > using the above data to generate peer-to-peer networks to understand how:

• colleagues are connected within the organisation
• colleagues are connected to one another
• strongly colleagues are connected based on the type of communication that they’re having with each other

Interesting, and perhaps somewhat controversially, Satalia are able to use this peer-to-peer network to review paying processes. This is the idea that your colleagues and closest peers can run through this by looking at your communication data and decide your salary, not your manager. Because, potentially, your line manager can be very distant from what you’re actually doing. 

So, that’s some of the ways that we’re using data science internally to improve a pay review process.