Accessibility Links

How to Commercialise Data and AI

16/03/2018 By Barty Isola

You've heard it before, (or maybe saw it on the cover page of The Economist, May 2017): data is the oil of the new economy; a valuable resource that can bring economic rewards to the companies who know how to exploit it. 

Cover page of The Economist, May 2017

 

Thanks to falling storage costs, growing use of AI, and increased digitisation, businesses now have more data on their customers, suppliers, and processes than ever before. Generating and gathering vast volumes of information is no longer a problem for companies: the challenge is now turning that data into usable business insight.


At La Fosse Associates' recent 'Data Advisory Board' roundtable event on the commercialisation of data and AI, C-level execs shared their experiences and aspirations for data science and artificial intelligence.  

What's holding back investment in AI and data science?
For now, one of the biggest blockers for AI and data science is convincing management that spending on the technology today, will pay dividends further down the line, and persuading them of the need to invest in the right tools. 

Confusion around the definition of AI, and which products are genuine artificial intelligence, is adding to reluctance in some sectors to explore the concept more thoroughly. It's a challenge that can be addressed by broadening management's understanding of what's possible with AI, in order to build an appetite for using the technology.

Where should you start?

Begin with a clear need – increasing the personalisation of a product, automating processes, or saving money in a certain part of the business, for example. From there, roll-out an AI project that be completed in a relatively short time, and show a clear return attributed to it. 

By kicking off with a simple, quick-win, execs can then prove the financial value of AI or data projects, and encourage the business to reinvest that value into other, more complex projects. 

Do you need to capture everything?
Capturing and using clean data – rather than hoovering up every piece of information available – should be the priority for businesses. Without a clean dataset, the true value of the information can't be realised. 

If the data has any anomalies within it, those will be reflected in the way any algorithms learn and the conclusions they arrive at. To address this, consider hiring a higher ratio of data engineers, who can keep datasets truly clean, to highly-trained data scientists, who will design the algorithms.

How should you structure analytics within the business?
There's no clear consensus yet on whether the analytics should be central or tightly-bound to a commercial function. Being central means that an analytics unit can have close contact with the heads of various functions, examining requirements and KPIs across the business. However, placing the unit within the finance department, having it report into the CTO, or locating it within product team allows for a greater level of focus. 

A central function that's aligned with commercial functions may offer the best of both worlds.

Where do you find the right data engineers?
It's common for data engineers to move on from the role relatively quickly, creating a turnover problem for businesses. One solution is to hire junior people, framing the role as a stepping stone to becoming a data scientist, or otherwise plotting a clear career path for new starters. 

Another option is finding more challenging tasks within the business – that involve building rather than simply maintaining – to provide extra variety in the data engineer role. Platforms orientated around data engineering are increasingly coming onto the market, and may ultimately affect companies' requirements around data engineering in the future. 

What do you need to consider when introducing new tools?
It's not just functionality, ease of use, and cost that companies need to bear in mind when choosing software: talent need to be at the top of the agenda too. 

Certain data science tools have larger pools of developers; when planning for the future, businesses need to factor in the widest availability of talent when selecting which applications to use. 

Could GDPR really be a good thing for data science?
The upcoming GDPR regulation is throwing the data debate into sharp relief in businesses across the country. Thanks to its arrival, companies are thinking about the risk and rewards associated with collecting each individual piece of data. GDPR has also raised questions of data - what's being collected and why - to board level. As result, companies will be increasingly selective over the data gather, and data scientists will only focus their attention on the most valuable information.

Can data gathering strengthen relationships with customers?

On the back of new regulation, companies' data gathering practices will not only come under scrutiny from the board, they will be examined by customers too. Those businesses that choose an ethical path to data gathering and interrogation, allowing users to control their own data destiny, will be those that are best able to turn upcoming regulation to their advantage, by using it to build trust with customers.

 

For more information on the event, or how you can begin commercialising data and AI within your organisation, get in touch to learn how we can help.

02079322091 - barty.isola@lafosse.com - LinkedIn

02079321687 - rad.akbari@lafosse.comLinkedIn

Where to next?

Search for BI, Data and Analytic jobs here

Tagged In: Recruitment
Add new comment
*
*
*