Menu Close

Rewire your company to be AI ready

Tim Lewis

Artificial intelligence (AI) presents opportunities for huge growth in the next decade, but if you want to take advantage, you must “rewire” your organisation now, write Tim Fountaine, Brian McCarthy and Tamim Saleh for Harvard Business Review

Artificial intelligence (AI) is already starting to reshape business and, as the technology required to enable it becomes more affordable, AI is going to be revolutionary. McKinsey partners Tim Fountaine, Brian McCarthy and Tamim Sale estimate that AI will add US$13 trillion to the global economy in the next 10 years. 

However, their research suggests only 8% of companies “engage in core practices that support widespread adoption”. The problem is a failure by business leaders to break down the cultural and organisational barriers preventing their companies taking full advantage of the opportunities offered by AI. 


The authors outline three essential preliminary “shifts” you must initiate if you want to “rewire” your company: 

1) From siloed work to diverse teams. AI should be developed by “cross-functional” teams made up of both analytics experts and those responsible for the day-to-day operation of the business. These teams will focus on “broad organisational priorities” and consider the operational changes required for new AI applications to succeed.

2) From top-down decision making to devolved data-driven decision making. For AI implementation to succeed, staff at all levels of the hierarchy must be empowered to make decisions based on the algorithms’ suggestions.

“If employees have to consult a higher-up before taking action, that will inhibit the use of AI.”

3) From perfectionism to a “test-and learn mentality”. Forget the idea that a tool needs to be perfected before it is launched. Banish fear of failure, launch your AI application, collect user feedback and learn from your mistakes. This agile approach will enable your teams to create a minimum viable product (MVP) in a matter of weeks. 


It is vital you “get employees on board” right from the start and prepare the ground for the successful adoption of AI. Fountaine, McCarthy and Sale advise taking these four steps: 

1) Explain why. Your employees might be concerned AI is going to make them surplus to requirements. It is important to provide a vision that helps them understand why AI is important to the business and where they will fit into the new “AI-powered organisation”.

2) Anticipate unique barriers to change. Each organisation has its own distinct culture and therefore its own set of hurdles that must be jumped on the way to AI success.

In order to identify and overcome these barriers to change, look at and learn from how past change-initiatives tackled the same problems.

Sometimes you can turn a negative into a positive, like the financial institution with a strong emphasis on the human touch that produced a booklet to show relationship managers how combining their skills with AI’s bespoke product recommendations could improve the customer experience.

You could also employ an analytics translator able to “bridge the data engineers and scientists from the technical realm with the people from the business realm”.

3) Budget for adoption and integration as well as technology. In one survey conducted by the authors, nearly 90% of the companies that had achieved scale had spent more than 50% of their analytics budgets on driving adoption of AI.

4) Balance feasibility, time investment and value. You will not be able to achieve everything immediately. Take a long-term view – typically three years – and develop a set of initiatives with different “time horizons”. Automated processes, e.g. AI-assisted fraud detection, can provide quick wins, while processes that combine AI and human involvement, e.g. AI-supported customer service, will take longer.


In order to get AI up to scale you must decide where to focus your AI and analytics capabilities. Successful AI-enabled companies divide the majority of tasks between a centralised “hub” and decentralised business units or “spokes”. Everything else falls into a “grey area”, with a company’s unique characteristics determining where it should be done.


Data governance, AI recruiting and training strategy, work with third-party providers of data; AI services and software and systems and standards related to AI should be handled by a hub overseen by the chief analytics or chief data officer.

Companies that have successfully implemented AI on a large scale are three times as likely as their peers to have a hub, according to research conducted by Fountaine, McCarthy and Sale.


Tasks related to adoption, including end-user training, workflow redesign, incentive programmes, performance management and impact tracking should be handled by the spokes.


Other tasks, such as setting the direction for AI projects, analysing the problems they’ll solve, building the algorithms, designing the tools, testing them with end users, managing the change and creating the supporting IT infrastructure, can be handled by the hub or the spokes or shared.


Create a team combining analytics, business and IT leaders to oversee your AI journey and teams – including the manager responsible for the initiative’s success, translators, data architects, engineers and scientists, designers, visualisation specialists and business analysts – to handle the development of each individual AI initiative. 


To ensure the successful adoption of AI, set up an internal AI academy. The academy should provide leadership with a “high-level understanding of how AI works”, sharpen the skills of employees responsible for analytics – such as data scientists, engineers and architects – and analytics translators, and provide on-the-job training for frontline workers.


To maintain momentum, the authors suggest business leaders do these four things: 

1) Walk the talk. You are a role model. Demonstrate your commitment to change at every opportunity. For example, attend the AI academy.

2) Make business accountable. Analytics staff are often made “owners” of AI initiatives, but it is important that business units take the lead and are made responsible for their success.

3) Track adoption. Compare processes before and after AI to motivate staff. Monitor implementation to ensure it is easy to change course if necessary.

4) Provide incentives. Reward employees who participate fully and successfully in the process of AI transformation. 


Most AI transformations take 18 to 36 months, with some taking as long as five years, but if you can persuade all of your employees to back the change, it will pay off in the long run. 

“Companies that excel at implementing AI throughout the organisation will find themselves at a great advantage in a world where humans and machines working together outperform either humans or machines working on their own,” write Fountaine, McCarthy and Sale. 

Source Article: Building The AI-Powered Organisation
Author(s): Tim Fountaine, Brian McCarthy and Tamim Saleh