Harnessing the power of artificial intelligence to future-proof your business

Artificial intelligence (AI) is now gaining the same kind of acceptance as we have for email, word processing and spreadsheets. Now proven in the eyes of many are technologies from the heuristically rigid, reactive algorithms capable of beating chess grandmasters to today’s adaptable machine-learning systems used in chatbots and autonomous vehicles. Entire nations now stake their futures on AI.

In October 2017, the UAE government launched its Strategy for Artificial Intelligence, established an AI Council, and appointed the world’s first minister of state for AI. The government – a perennial regional leader in technology strategy, adoption, and implementation – is confident the program will reduce costs, enhance operations, and preserve public and industrial safety in sectors from transport and construction to education and healthcare.

Independent observers share the UAE government’s outlook, with some estimates putting AI’s GDP contribution at 14 percent of the nation’s 2030 GDP, and others predicting that AI will contribute more than $90 billion to the economy by 2035, including $37 billion to the FSI sector and $22 billion to healthcare.

But as regional private enterprises look to board the AI train, their use cases are focused on the maximisation of value in their data. And this presents some challenges. For a start, all data is not created equal. Some is useful; some is not.

And the data that is useful is not useful to all stakeholders in the same way. So, success becomes about getting the right data to the right people in the right format at the right time. When examining this challenge, savvy technologists and seasoned business leaders will recognise the need for a culture change. And they will follow these three steps to deliver it.

Eliminate silos

Depending on the scale and type of business, skills, tools, and goals will vary, as will the nature of the data on hand. Its sources will be a factor in how it is handled and in how it should be homogenised. The goal should be a centralised repository of standardised formats that can be repurposed in real time to give the most useful view to the user – be they a finance executive, salesperson, customer service agent, or C-suite officer.

To get to this point, silos must be connected into a cohesive whole and “cleaned”. Text data will have inconsistencies; dates may be in a variety of formats; different business units may have categorised their data using proprietary labels; and some fields may be unpopulated. For data to be useful to AI algorithms, and hence to the business, all these issues must be addressed.

After this, analysts and data scientists can step into make the data come to life – first by deriving new views from raw bytes, then by enriching the ecosystem with intuitive dashboards and reports.

At all points, skills will be critical. The data-science pipeline needs experts on the business, as well as specialists in the technologies and data being reformed. Failure to pay due attention to skills gaps along the way can lead to subpar results.

Prove value

Having addressed skills and opened up the data ecosystem to potentially provide value for all business units, project leaders must now quickly demonstrate that value before confidence wanes and silos reemerge.

Transparency is critical, so data teams should concentrate on picking some quick wins for each department and building use cases that can be easily tied to measurable KPIs. Nothing garners buy-in like success, so if the data team can make the KPI dials glow hot, then they will win hearts and minds, and culture change will be on the cusp of realisation.

If expectations are sensibly managed and wins convincingly demonstrated, people across departments will feel the value of AI and ask for more. Line of business itself will now propose use cases, and the culture change can grow from there. Data leaders must ensure that experts from all sides come together to design each KPI, and ensure it is compatible with what is deliverable and what will impress end users.

Scale up

When skilled employees move on, sometimes they leave behind a project that, while useful, cannot be amended because the skills to do so have departed. In such cases, artifacts can quickly become relics. If silos are to be eliminated for good and value from a data ecosystem is to be sustainable, reusability of artifacts must be a key focus of the data team.

In an agile business, new users must encounter shallow learning curves. Their quick adaptation will allow them to build on previous work and produce their own assets, even if the initial creator is no longer part of the team.

Slow and steady wins the race

Share assets and success stories. Reuse successful assets and discard others. Build a common vision and communicate it to all. Skill, upskill and reskill. Assess results fairly and act decisively on the story they tell.

These are the guiding principles of AI culture change. Getting data scientists, analysts, data engineers, business executives, finance leaders and others to become a holistic organism that pulls in a single direction is more easily accomplished in a step-by-step plan.

But as AI culture permeates the corporate space, data leaders must not forget to address issues such as responsible AI, governance, and technology resilience in the age of hybrid IT. There is no doubt that change is a process, and culture changes can be some of the longest. But in the case of AI, the value is inescapable.

The Arab Gulf region is in economic recovery. Competitive edge will determine who shares in that success and with what measure. And AI has proven itself the surest route to that edge.

 

Source:Arabian Business