AI and machine learning – why should you care?

Are you tired of the hype yet?

After all, we’ve been inundated for years with endless rhetoric concerning the promise of artificial intelligence (AI) and machine learning (ML). Positioned by many as the silver bullet for all your business ills and personnel shortcomings, AI and ML have an awful lot to live up to. The question, have they lived up to the hype?

Where are the results? Hard, quantifiable and reproducible benefits to businesses, organisations and communities – where are they? With approximately 90% of data science projects failing to make it to production, it seems that substantive results are, well, lacking.

For organisations seeking to gain value from their prized troves of data, this isn’t good news at all. Procuring and curating large stores of information is expensive and complex, requiring skilled people and pricey technology. Leadership recognises the potential of their data to transform the way people live and work, generate extra streams of revenue, and increase business value. They have a rich vision for obtaining considerable benefit from their deep wells of information, and that fire is fuelled by technology vendors. Unfortunately, the gap between vision and execution is often enormous.

In principle, the transformation and empowerment potential of AI and ML is immense. In practicality, many obstacles have derailed the realisation of this potential, including:

  • Poorly defined goals and objectives
  • Incomplete, inaccurate data
  • Inadequate human resources
  • Internal misalignment
  • Ineffective change management
  • Poor training and adoption
  • Incorrect and inappropriate technology
  • Disconnected silos of information

The good news that you can energise your data with intelligence:

Thanks to Peter Coffee at Salesforce for this powerful acronym

Artificial intelligence and machine learning have the power to supercharge your data initiatives, uncovering valuable insights and providing perspicacious predictions. Here are a few real-world examples of the value that can be realised from AI and ML:

  • A telecommunications giant leverages a churn model to identify customers in danger of falling away and drives engagement to mitigate those risks.
  • The administrators of a two-hundred-million dollar array telescope uses ML to uncover hidden insights in a mess of operational data that improve performance,
  • A global radio communications business deploys ML to flag deals at risk, enabling the sales team to optimise resources and focus on opportunities that require their attention.
  • A highly-impactful charity identifies one-time donors who are likely to become regular givers, allowing them to focus valuable resources on reaching out with appeals.
  • Health care executives can better decide if they should expand care programs to other populations through the combination of predictive analytics and statistical modelling.
  • A world leader in the supply chain and logistics space analyses historical data to build a predictive model that proactive warns hub managers that demand may outstrip capacity.
  • A high-tech business in the healthcare industry analyses and models their customer engagement data to mitigate churn and boost their bottom line.
  • A customer-focused technology business employs ML to analyse historical data so that service cases can automatically triaged, routed, and escalated.

We all need a balanced perspective on ML and AI. Do not be fooled by the hype, and do not be dissuaded by the failures. As the technology improves, data grows, and business mature, the likelihood of success with ML initiatives continues to grow. This is especially true as we see the development of no-code AI platforms, enabling rapid time to value and decreased reliance upon scarce specialist resources. Many organisations could achieve tremendous results and drive significant value with a well-delivered ML initiative. The question is, how?

Five recommendations on how to deliver a successful ML project:

  1. Begin with a clearly-defined use case. A lean, valuable and achievable MVP approach has a far greater rate of success than a bloated, optimistic beast.
  2. Consider a pilot build as opposed to a proof of concept (POC). Way too many organisations have invested valuable time and resources in expensive and useless experiments, lacking a strategic road map for development and deployment.
  3. Collate, analyse and prepare your data before you build anything. The data used to train (teach) your ML model is critical to project success, and therefore must be thoroughly prepped prior to model building. Note: this process of data remediation and augmentation never stops in successful ML initiatives.
  4. Consider a platform-based approach. The traditional approach to data science has delivered much value, and will continue to do so, but a low-code platform approach has the ability to deliver quickly and cheaply, freeing up data scientists for high-value work that only they can perform.
  5. Invest in the best people right from the start. When you skimp on people, whether internal or external, you sound the death knell of your project, so be sure to invest in a dream team of technical and business players if you’re serious about success.

Consider your organisation – is there potential for machine learning to extract hidden value from data and impact the way that your team live and work? Could AI and ML transform the customer experience and enhance operational efficiencies? Perhaps artificial intelligence and predictive insights could even elevate your organisation to the place of an industry disruptor? You’ll never know if you don’t have the vision and courage to embark on the ML journey – just be sure to have a strategic approach and follow sound advice as you do so.

Do you have a question about this post or the subject of machine learning in general? Please send me DM, or comment on this article.

Addendum: What is Machine Learning?

Machine learning (ML) is analytical model building that can be automated, where the systems “learn” from data and identify patterns, resulting in meaningful conclusions and predictions. I like this definition of ML from the SAS institute web site – “Machine learning is a method of data analysis that automates analytical model building. It is a branch of artificial intelligence based on the idea that systems can learn from data, identify patterns and make decisions with minimal human intervention.”