AI was arguably the buzzword of 2017, with the likes of Facebook and Google dominating the headlines for their projects in this field. And with China building a $2.1bn research park focused on AI research – and revenue from AI designed for enterprise applications expected to reach over $31bn by 2025 – it is fair to say this technology will remain a focus for many years to come.
However, many of the headlining AI applications designed to date have come from tech giants, with many other industries yet to see major uptake throughout their value chain. But how do pre-digital companies, otherwise known as digital transformers, take advantage of this AI boom?
Business leaders from pre-digital companies may be forgiven for thinking that trying to emulate the likes of Google to intelligently target customers is the way to go. However, much of AI’s transformative potential comes from data collected by R&D and the industrial processes that take place during a company’s digital transformation.
The business challenges this type of data can address are completely different from those that companies like Google are looking at. To help pre-digital companies harness the transformational power of AI, we have developed the following 11 steps:
Build trust in AI
AI is going to make businesses more productive by automating repetitive operations, releasing resources to focus on higher-value tasks and remove the subjectivity from a human judgment call.
However, organisations will only adopt AI if they can trust the decisions it makes are within an acceptable level of risk. As a general rule, the greater the negative consequence of incorrect AI behaviour to the business, the more rigorous the approach to data and training needs to be.
Don’t blindly hoard data
There is a belief that all data holds value, but this simply is not true. Businesses must identify a problem that needs solving and then work smartly to identify the data best suited to solving it.
Organisations will only adopt AI if they can trust the decisions it makes are within an acceptable level of risk
In essence, quality of data supersedes quantity. This approach promotes understanding of the problem’s context and builds an informed data acquisition and management strategy for the capture, control and exploitation of the data. By taking this approach you can end up with reliable, trusted results.
Focus on user experience
AI interaction must be intuitive or it will not be taken up. Here, we can learn from ‘digital native’ companies: Google Photos, for example, runs neural networks, image analysis and natural language understanding, but all the user needs to master the application is a search bar and grid of their photos.
Your AI implementation needs to be the same: simple, intuitive and natural to interact with. Whether designing new materials with specific physical properties or scaling up production of a monoclonal antibody to treat a rare disease, any interaction with AI must be designed with the user in mind.
AI is good at automating routine tasks but cannot deal with situations outside its training. To avoid AI failure, businesses must check random samples of AI outcomes with human experts and plan for human intervention when unexpected events occur, such as extreme weather.
This is partly because of the complexity of non-routine events and also because there is often not enough data on unpredictable events to train the AI appropriately. In this situation, expert human intervention needs to be ready to step in and take over.
AI is about talent as much as technology
While digital transformers have different problems to digital native companies – and need different skills and resources – they should emulate the Google/Facebook approach of finding the right people for the task at hand, and avoid the temptation to simply throw technology at the problem.
By investing in an AI talent pool – one with the right blend of experience, problem solving and technical expertise – and combining that with a clear vision for the AI solution, non-digital businesses can begin to utilise the transformational power of AI.
AI should be designed by people who understand the problem, the underlying data and what it represents in a real-world context. The best teams include representatives from IT, operations and business teams, domain experts, AI and data analytics experts, and, critically, people who can translate between these different roles.
Seek out AI experience from other domains, don’t limit your search to your sector; your problem may have already been solved elsewhere
By following this approach, business leaders can ensure their AI projects are backed with the correct technical expertise but also that the team members involved are fully aware of the goals and the underlying business value derived from the company’s AI plans.
Look outside your organisation
Specialist AI skills rarely exist in pre-digital organisations. Look externally and embed specialists within business teams. Seek out experience from other domains, don’t limit your search to your sector; your problem may have already been solved elsewhere.
New thinking and fresh perspectives on your most challenging issues are drivers for game-changing innovation, which can result in new growth for pre-digital companies and provide them with the ability to innovate ahead of their competition.
The temptation is to cover everything, launching an all-encompassing initiative to build and implement a complex cross-enterprise AI platform. But our experience shows that starting too big, too early, undermines effectiveness and delays the impact felt by the organisation.
Companies should start by building a roadmap that identifies the business decisions that AI can inform. Focus initially on well-understood opportunities that can be executed quickly. This will build critical momentum for AI programmes.
Explore multiple AI projects
Accelerate this momentum by running multiple AI projects in parallel, ensuring the best ideas are progressed rapidly. This agility is how digital native companies deliver innovation, but it is often lacking in many pre-digital organisations.
Rapid prioritising of resources into the delivery of the most successful ideas demonstrates an AI strategy focused on the realisation of tangible value, which is crucial to building trust. Trust is vital to the success of any AI project, but particularly in pre-digital companies where staff may be experiencing this technology for the first time.
Monitor your many AI projects, checking the relative performance of each, and abandon bad ideas; use successes and failures to improve your training regimes. Expertise only comes from having many different experiences.
Sampling widely and failing fast leads to a far better trained AI that is dependent upon fewer misplaced human assumptions and is more open to innovation. Agility in your approach to delivering AI is, therefore, crucial if you want your digital transformation programme to have a positive impact on your business.
Businesses must define measurable goals and KPIs for each new AI release, e.g. increased customer engagement, improved production line quality and reduced non-productive time. Businesses should then use these to demonstrate success to financial backers and to feed back into the AI strategy.
Placing AI within such a managed framework delivers many operational and commercial advantages. Performance tracking makes it possible to dashboard and visualise the impact of AI, to manage the programme as a service and embed it into a broader DevOps environment.