Machine Learning has already changed our daily lives in the form of products and services we use on a regular basis. This all too pervasive technology is not just a household name (hello Alexa!), but it also powers the predictive text features we use on our devices, chatbots that we engage with, personalised Netflix recommendations that we rely on, the ads we see on social, the list goes on….
Why ML and why now?
Explosion of digital channels and the exponential growth of data collected presents several opportunities for organisations. To make sense of the available data it is imperative for organisations to embrace new technologies and advanced analytical techniques — to mine and contextualise data. ML-led initiatives offer opportunities to support customer retention, growth, increased productivity, and reduction of operational costs.
While everybody doesn’t need to understand the technical details of machine learning, it is impossible to ignore what ML does and doesn’t do from a business perspective. We can no longer afford not to know how ML can add value to our organisation.
Data … an essential pre-requisite
At the heart of any successful ML initiative is ‘data’ — that can be images, receipts, text, customer records, time series data from devices etc. These strategic data assets represent a cumulative imprint of an organisation’s processes, procedures, and interactions with their consumers. Through application of algorithms, insights can be drawn from these data sources by identify patterns and building contextual relationships. These insights can not only help explain what has happened, but also predict what may happen and through data suggest what actions to take should something happen.
Machine learning works with ‘data’ — and is best suited in situations where there is lots of data. This data needs to be reflective of the real-world situation, issues and challenges a machine is expected to tackle. As new data becomes available or business processes change — the models need to be continuously trained and improved to automate decision making. When trained properly these machines promises to take over non-routine, repetitive, and cognitive tasks, but also process large volumes of data which might not be humanly possible without machines.
Conversely, if there isn’t a lot of data (quality and volume) — it is important to identify necessary steps required to augment existing data or improve data collection processes in place.
How do I get started?
Organisations looking to kickstart their AI/ML journey and could start to look at the following areas to identify strategic challenges for ML opportunities:
- Growth — Rising collection and analysis of customer data using ML can help identify new growth opportunities — in the form of product positioning, new offers, new channels and delivering personalised offers at the right place at the right time. ML has already transformed manufacturing, logistics and financial services to name a few.
- Improving Customer Experience — The ability to mine qualitative and quantitative data from several sources (CRM, financial, procurement, online, social and 3rd party sources) helps brands understand their customer pain points and put efficient processes in place to increase customer loyalty and relationships.
- Operational Efficiencies — automate key manual and repetitive operational processes to increase efficiencies and reduce waste.
And don’t forget
Strategic vision and operational execution go hand-in-hand, to ensure alignment from a delivery perspective, it is also important to evaluate the following areas:
- Organisational Context — by building an understanding of how various functional lines operate, key business processes that overlap and how data is collected across these workflows helps to establish clarity on objectives and identify roadblocks before commencing on ML journey. A clear organisational evaluation also helps understand what problems can be solved with ML and how learning can be embedded back within business for organisational success.
- Technology — a end-to-end evaluation of existing technologies and tools required to deploy ML solutions, building a roadmap of initiatives if there are gaps.
- People — building the right team is crucial for success of ML initiatives. Assessing teams for required skills to not only identify ML use cases, but also to build and deploy solutions is essential. In cases where there are gaps in skills within existing teams, organisations could look towards consulting organisations to augment their existing teams.
Building the requisite foundational elements using right technologies, teams and organisational alignment sets the initiatives for long-term success.
