Nordcloud data and industry experts Risto Jäntti and Cormac Walsh discuss what’s needed to unlock AI on an enterprise scale.
Splicing AI into the DNA of an enterprise will be one of the most challenging requirements ever faced, however the payoff will be equally huge.
And the changes AI will bring will catch most enterprises off guard. Arguably this has happened already. For example, the marketing industry that for years has tried to measure and predict the capricious nature of human behaviour, will see that this power is child’s play to AI.
Why? AI thrives on data - albeit quality data - and sample size. How little or how much data do we share about our behaviour? The generation coming online right now will be the best understood in human history. And it’ll be understood by AI.
But how do businesses really grasp this potential and get real value from all the data that’s available?
Let’s look at the considerations that we see as being key in the implementation of all things AI.
Shifting the learning curve
As it’s so new and spectacularly misunderstood, AI of all types is prone to misalignment between capabilities and requirements. This misunderstanding - combined with misalignment - means that there will be a period of rapid changes of opinion and expressed requirements from business, resulting in confusion.
Maintaining the fine balance between what can be done and what should be done is vital. This will mitigate this confusion and keep focus when building a data and AI strategy.
So, how to start. Firstly, with the realisation and acceptance that AI will be ubiquitous. It will be and needs to be everywhere: a truly pervasive technology embedded-in or rewiring everything that an enterprise does.
It should be seen as a general purpose tool, a universal Swiss Army Knife. In this respect, the realisation that AI in daily business operations will be nothing special, it will be a given. Just like air conditioning, electricity or your messaging app, a totally transparent technology that gives you value and is ever-present, but never gets in the way.
Once you get to this stage, the learning curve can be moved, from the user to the systems. And it’s better if the systems learn the user and how to benefit them the most (remember, it’s smart enough to figure this out), to shape a more efficient learning curve that’s in the right place, used in the right way.
It should be seen as a general purpose tool, a universal Swiss Army Knife. In this respect, the realisation that AI in daily business operations will be nothing special, it will be a given. Just like air conditioning, electricity or your messaging app.
Getting to this place takes a few steps
And getting here relies on public cloud. You’ll need a stack of tech and corporate culture traits, in order to work and continue to grow.
Let’s give you an example. An enterprise client of ours in the UK knew that they had to get ahead of the curve in all things AI. But given the size of their organisation and the vast and shifting landscape of cloud-based AI tools and technologies, they knew that it would be a struggle to arrive at the “perfect” use-case with the “best possible” technology.
Our advice was to focus less on the technology itself and the use-cases, but more on the support model, the vision and use-case capture mechanism.
- Kick off with a single crystal-clear and well-communicated demonstration of the power and ease of use of AI technologies.
- Combine this with a natural human interface and the processes behind it to capture the ideas that should emerge from daily use.
This approach has the benefit of developing user participation and investment from across the enterprise.
Responsible, ethical & trustworthy AI
The recent AI Safety Summit hosted by the UK government in early November highlighted the growing importance of addressing potential risks and responsibilities associated with AI usage.
The summit brought together AI experts, researchers, government and industry leaders to discuss and address the challenges posed by AI, including ethical considerations, transparency, and the potential risks associated with AI usage.
As public cloud providers introduce numerous ready-made AI models for organisations to implement across various use cases, it’s essential to consider these concerns in the context of adopting AI technologies.
These models, built from vast amounts of data, primarily have their development responsibility lying with the hyperscalers or the generational AI developer. However, when organisations enhance these models with their own data, questions and concerns can arise.
As an anonymous web user aptly put it, "With great power there must also come great responsibility." With the increasing adoption of powerful AI technologies, organisations must be prepared to adapt and implement new measures to ensure responsible and ethical use, reflecting the discussions and insights from the AI Safety Summit.
Concerns surrounding AI usage include potential harm to people or the environment, as well as reputational risks that could arise from misuse or miscalculation of outcomes.
To address these issues, organisations should consider the following:
1. Mandatory AI model governance
Establish a clear framework to guide the development and deployment of AI technologies. This framework should outline bias recognition and avoidance, ethical guidelines, risk management strategies, and compliance requirements. This governance process should be easy to understand and follow, otherwise people will try to circumvent it (don’t fight human nature).
2. Embed a corporate culture of AI ethics
Grow a culture that emphasises responsible AI use, where employees are educated to the potential consequences of their models and are committed to acting ethically, transparently and in a non-biased way.
3. Assess AI risks and impacts
Build a cadence of regular assessments to identify potential risks and negative impacts associated with AI technologies, be open and honest, this is not a naming and shaming exercise. Use these assessments to inform decision-making and to mitigate potential harm and to improve the quality of future models.
4. Avoid silence
Maintain open and regular communication with internal and external stakeholders to address concerns, share best practices, and collaborate on AI-related challenges.
We can help navigate the curve
The integration of AI into enterprise operations is a complex but hugely transformative step. And it’s one that sooner or later you’ll need to embrace.
Recognising this ubiquitous role, and making it a universal tool embedded seamlessly in daily operations, is essential for navigating the learning curve effectively. Public cloud will power the practical approach to AI adoption. And the need for responsible and ethical AI practices, makes a framework for governance, ethics and risks super important.
If you need help better understanding the opportunities and early use cases for AI in your organisation, use the form below or reach out to Risto or Cormac directly.
Get in Touch.
Let’s discuss how we can help with your cloud journey. Our experts are standing by to talk about your migration, modernisation, development and skills challenges.