How do you implement AI?
Artificial Intelligence (AI) is able to improve business and customer facing processes, automate repetitive data tasks, ease workload from staff and ultimately improve efficiency. Dr Janet Bastiman, Chief Science Officer at StoryStream, returns to help demystify how to implement an AI strategy into your business.
How does a business start using AI?
It’s not a magic wand, it’s a tool, you will need to ask yourselves – “what problem do I want to solve?” and make sure you have the data that defines that problem. This could be inwards facing, focusing on your internal systems or even outwards facing and helping improve the relationship between you and your customers.
At StoryStream, we use AI to solve many problems within StoryStream. From classification of images and text through to predictive analytics. Wherever we can identify a way of using AI to make things more efficient for us or more beneficial for our customers then we apply it.
What to know before implementing an AI strategy:
Information on AI varies from hugely accessible through to cutting edge research and everything in between so you need to know more about the solution you are trying to gain. From there talk to experts, attend meetup groups and the community. This is often a way to gain great insight into tried and tested AI systems. We’ve recently released a high level infographic for AI in marketing to help demystify some of the concepts. If you know what problems you want to address then it’s easier to get clear advice.
Questions your business should consider before implementing an AI solution:
- What is the scope of the project that AI can perform?
- Is the problem well defined?
- Do you have all the data needed?
- Would AI ultimately increase efficiency or drive profit?
To drill deeper you then need to consider whether you need a dedicated in house team, consultancy or third party service to assist you in deploying the AI system. If this managing in hour is an option for you, consider what training your staff may require, paying close attention to the new GDPR regulations and how you are collecting, storing and processing personal data. Never lose sight of the risks and benefits and of course how you keep all stakeholder informed.
Test, pilot, learn, improve…
A great way to introduce an AI system is to start with a small, defined project within your businesses where you can get a fast return on investment as a gateway to a larger project. More than likely, you and your teams will have much to learning from these and to get right for next time.
How do you know if AI will benefit your company?
The first is in automating repetitive processes to ease the workload of existing employees, think of labelling large amounts of data, finding similar profiles or even extracting important parts of text from large documents. This is all about efficiency and ultimately driving profit.
The second is to use the data you have to do new things, things that wouldn’t be possible with your human workforce. Whether this is discovering correlations in data that you may not have known where there or providing a service to people who cannot otherwise access a human operative.
With technology rapidly changing the right time to invest in AI is now. Because it’s changing so rapidly, the sooner you start seeing the benefits the better. This isn’t one of the large scale IT investments that we saw in the last century, think of AI as plugins to drive your business. With a third party service you can be up and running very quickly. Even an in house team should be able to get you something within a few months (depending on the complexity of the problem). Make sure you have the data you need and the problems clear and any partnerships are with companies who can update you with the latest AI techniques as appropriate.
The more coordinated your systems are the easier it will be to acquire the data to drive any AI solution and also act on the results.
Watch our Leaders in Tech: Reading Meet up page for further events.