Menu Close

What is an AI strategy and how do you use one?

To gain an edge over the competition, organisations are now implementing AI at scale. But to transform AI from technological solution to strategic asset, it’s crucial to have the right plan in place.

Human finger touches AI robot finger

Whether you’re anti-AI or an AI evangelist, there’s no denying the technology is completely changing the way we work, interact and go about our lives. To stay competitive in an increasingly crowded marketplace, organisations are implementing AI at scale. This has seen businesses move beyond simple forecasting and into detailed anticipation of customer movements, and proactively evolving with them.

However, many businesses adopting AI haven’t added the key ingredients required for its success. They’ve failed to understand that the competitive advantage of AI assets is a unique combination of an organisation’s culture, customer view and data. For example, replicating chatbots from a competitor will not generate a successful outcome.

A high-level AI strategy must translate company vision into measurable business objectives and put in place the necessary operational guardrails. It should foster rapid, multi-disciplinary experimentation of hypothesis towards meeting the objectives, and provide a well-defined process towards improving and productising validated initiatives.

The result is an overall capability – not just a technology stack – that will achieve set objectives, with key AI assets that provide the business with a competitive edge.

Measurable objectives

The AI strategy, through its canvas, provides the objectives, tactics, goals and measures for success.

There are ‘hard benefits’ around cost avoidance, and then there are less visible, financial tangibles involving ‘soft benefits’. Leaders need to understand that today’s soft benefits are tomorrow’s hard benefits involving cost out and should start to evangelise them accordingly.

All AI initiatives need to align with business objectives and be prioritised. Measurements may be in the form of ROI, or the customer’s user experience. Detailed metrics help greatly with the line of sight between strategic ROI objectives and the execution that’s happening on the ground – everyone is on the same journey; top down, bottom up.

Governance guardrails and AI

People, process, ethics and legal are key AI considerations that serve to underpin the strategic approach.

As part of the May 2018 budget, the federal government has funded the creation of the AI Ethics Framework for Australia, which forms a good starting point for an organisation-wide AI governance framework. Also due to the cutting-edge nature of artificial intelligence and the lack of reference governance models, it is always prudent to consult with other industry experts. Recent developments in Consumer Data Rights and open data will also introduce new privacy complications that need to be considered.

Human bias can present immense challenges within the discipline. Multidisciplinary teams help overcome culture, gender and geographical bias, input of data and training of algorithms for your AI solution.

Rapid experimentation and validation of hypothesis

During the implementation of an AI strategy, the biggest challenges are not technology related. Business leaders often make the mistake of separating the vision from the execution, often with a ‘build it and they will come’ mentality. This can result in disjointed efforts and a heightened risk of failed projects, due to a lack of stakeholder engagement, cultural anxiety, misalignment of objectives, and potential skills loss.

Instead, the implementation of an AI strategy should be first aligned to the customer, and work backwards to deliver agile, iterative value to the business. Such agility enables feedback for the strategy to evolve over time, changing as the organisation changes and allowing for input from all levels of the organisation. Together with democratised open source frameworks and high-level solutions provided by cloud vendors, your business, data science and engineering teams can focus on quickly solving AI problems instead of reinventing established platforms or algorithms.

Lighthouse projects (a minimal viable project that acts as a beacon for larger-scale capabilities) will need to be proven to work for a variety of key internal stakeholders. Your multidisciplinary teams will need to be upskilled. Processes will need to be streamlined. There will be mistakes made and lessons learned – but that’s okay, there have to be levels of experimentation in order to succeed. Business leaders need to focus on a culture of learning and continuous improvement with employees at the centre, to drive true transformation that radiates to the customer experience.

Read next: Do you need a Chief AI Officer?

Leave a Reply