Cloud

Artificial Intelligence – a team sport

Share this post:

What does it mean to deploy artificial intelligence into your business and what makes an AI project successful?

These are some of the many questions we will be tackling in depth at our upcoming event on the 21st May at The Groucho Club, but until then, here are some topics you might wish to consider.

Over the past five years of working in AI with IBM Watson, I have been asked many times what does it mean to get started with AI? What can I do with AI in my business? I like to turn the question around and ask “what aspects of your business do you want to improve?” or “what are you trying to achieve”. The answers I get are many and varied but broadly can be put in these categories:

  • I want to allow my customers or employees or other stakeholders to answer their own questions.
  • I want to make my employees more productive and make their jobs more interesting.
  • I want to automate parts of processes that currently require laborious human intervention.
  • I want to better understand what is going on in my business.

Often there is an implicit or explicit desire to reduce costs. The good news is that a well implemented AI solution will usually help with that goal as well as delivering many other benefits such as improved customer and employee satisfaction and increased revenues.

How is the AI going to enhance the experience of my customers and/or employees

I cannot over emphasise the importance of great design to the success of AI projects. Even the process automation implementations need to integrate seamlessly with the rest of the process and handle exceptions gracefully.

In the case of AI systems with user interfaces the expectation of users is extremely high. The proliferation of mobile applications and personal assistants has raised the bar. Users will expect to be able to ask questions (e.g. of a chatbot) in a very natural way and to be understood. They will expect to be offered choices and suggestions that meet their needs and be presented with information in formats and styles that are engaging, appropriate, great to look at and easy to interact with.

Integration with other systems

If somebody wants to know their account balance you need to connect to the appropriate system(s), gather the data and display it to the user. Integrating with existing systems is often one of the more challenging parts of the implementation but can really make the difference in terms of the capabilities of the system.

The fuel for AI is data

AI projects will only succeed when underpinned by a solid foundation of trusted, reliable information. This applies to both the data used to train the AI and to the data used as input on a day to day basis. The AI Ladder is a good model that explains the relationship between AI and its underlying data platform.

The AI Ladder

Another thing to consider is where will you build, train and run your AI. AI workloads often lend themselves to running in the public cloud but you may have good reasons (security for example) to run them in a private cloud or on premise. It is highly likely that your data will be in multiple locations and on multiple clouds. You need to have a clear strategy on how you will manage this complexity and optimise your implementation.

 The AI itself needs to be accurate, trusted, reliable, transparent and unbiased.

You need to be able to explain what it is doing and why to a customer or even a regulator when asked. You need to train the AI using representative data and you need to get the people who really understand your business and your data involved in the training. AI is a team sport. You need data scientists, developers, business experts and operational teams to all work together.

The IBM Watson Platform, alongside other IBM capabilities, has very strong support for all the concepts mentioned and IBM has vast experience of real world AI Implementations.

Join me  on 21st May, at our #WatsonAI breakfast briefing to learn more.

Data Science & AI Architect

More Cloud stories
By Mark Restall on 5 November, 2024

Impact on Data Governance with generative AI – Part Two

Many thanks to, Dr. Roushanak Rahmat, Hywel Evans, Joe Douglas, Dr. Nicole Mather and Russ Latham for their review feedback and contributions in this paper. This blog is a continuation of the earlier one describing Data Governance and how it operates today in many businesses. In this blog, we will see how Data Governance will […]

Continue reading

By Mark Restall on 28 October, 2024

Impact on Data Governance with Generative AI – Part One

Many thanks to, Dr. Roushanak Rahmat, Hywel Evans, Joe Douglas, Dr. Nicole Mather and Russ Latham for their review feedback and contributions in this paper. Introduction As artificial intelligence (AI) and machine learning (ML) technologies continue to transform industries and revolutionise the way we live and work, the importance of effective Data Governance cannot be […]

Continue reading

By Steve Moe on 14 October, 2024

Accelerating the creation of AI-infused solutions in a hybrid environment

As a global leader in software for banks and financial services organisations, Finastra aims to bring generative AI (gen AI)-enriched solutions to its clients without limiting their options around choice of platforms. Steve Moe, Head of Technology for the Lending business at Finastra, explains how a collaborative initiative between IBM, Microsoft and Finastra, using the […]

Continue reading