Artificial intelligence has evolved to become one of the most overused and misunderstood terms in business while also offering the potential to be the driving force in business decision-making, automation, and scalability.

Now is the time to develop strategies that set your business up for success for years to come. This article attempts to shed light on the origins, definition, types, business applications, and how senior executives can approach introducing AI to their business.

Defining AI

Artificial intelligence (AI) is a broad term and describes technology’s ability to perform intellectual tasks typically only performed by humans. Technically speaking, a spreadsheet that helps calculate insurance rates based on a range of inputs can be classified as AI.

Applied in a business context AI can describe what happened based on historical data, anticipate what is likely to happen in the future, and provide recommendations on what to do to achieve goals.

The process that made AI the powerful technology it is today is machine learning (ML). It describes the ability of a system to analyse data, identify patterns, and make recommendations by processing data and experiences without explicit programming instructions. ML models adapt and become more accurate over time.

Examples of ML include:

Talent management – organisations identifying which employee traits are correlated to high performance based on CV information and performance review data.

Pricing – ride share services adjusting pricing based on estimated customer propensity to pay a higher price.

Navigation – courier services planning delivery routes based on weather, traffic, and fuel costs.

The latest advancement of ML is deep learning which is a technology that requires even less human guidance and is more accurate than most ML methods. Areas deep learning have helped to evolve include challenging tasks such as image recognition, sound processing, and natural language processing. Google Assistant, for example, is a product of deep learning advancements.

From zero to Skynet in 200 years

The fact that AI has evolved into the most disruptive technology since the introduction of the internet is based on the evolution of three major trends.

Big data – The digitisation of our economies and the associated data volumes have been crucial in creating data sets required to effectively train machine learning algorithms. According to Globalwebindex, there are now over four billion people online globally generating vast amounts of data every minute of the day.

Algorithms – Researchers have paved the way for AI by gradually improving algorithms. Theoretical work in the 1800s was brought to life when American scientist Frank Rosenblatt developed the very first machine learning model in 1958.

Computing power and storage – Since Amazon has brought cloud computing and storage capabilities to the mainstream, the costs and ease of access have improved significantly.

These three trends have led us to AI today, a technology so powerful that it already outperforms humans at certain tasks.

How it works

The machine learning process roughly works the same in each case

  1. Business objective: The business objective is defined and AI might be identified as the way to achieve the objective.
  2. Data preparation: Training data is processed (cleaned and standardised) to make it suitable for the model.
  3. Model draft: A first iteration of the machine learning model is created.
  4. Model training and optimisation: Based on a training data set, the model is fine-tuned to generate better outputs.
  5. Business rules: Business rules are defined to do something with the output of the ML model.
  6. Model deployment: Once the accuracy of the model is satisfactory and the business rules are defined, the model is deployed, which means that “real-world” data inputs (i.e. not training data) can be used to return results.

Rule of thumb here is: more data –> better model –> higher accuracy.

A simplified example of a company going through this process could be a retailer wanting to increase the lifetime value of their online customers.

  1. Business objective: Increase the value of products purchased online per transaction by 25 per cent.
  2. Data preparation: All online and offline purchase data captured via the loyalty program is standardised and transferred into a central database. This data includes customer gender, age, product, product category, and date of purchase.
  3. Model draft: The initial model is created to identify customer segments and the products they are likely to buy together each season.
  4. Model training and optimisation: The initial outputs are compared to the latest real-world data and variable tweaks are needed to make the predictions of the model more accurate.
  5. Business rules: When checking out, each customer should be presented with a last-minute product recommendation that amounts to a minimum transaction value increase by 25 per cent.
  6. Model deployment: The check-out product recommendations are now visible to each website customer and the system will optimise recommendations dynamically based on the latest customer purchase behaviour.

The people needed to make it happen

Larger scale companies who are experienced in ML typically involve a wide range of personnel in projects. Here are some examples. Please note that the job titles and responsibilities might vary greatly across organisations.

Business Analyst – Understands the business needs and determines the outcomes to be achieved.

Data analyst – Defines and sources the data required to solve the business problem.

Data engineer – Establishes the connection between the data sources and the database. She also defines the database structure to ensure efficient access.

Data designer – Defines database structure to ensure efficient access.

Database administrator – Manages the storage facility including performance and security backups.

Data architects – Is across the big picture of data flows and defines the data architecture in collaboration with the data designer and the data engineer.

Data scientist – Uses statistical analysis and data visualisation tools to explore data and creates machine learning models based on findings.

ML Engineer – Deploys the ML model and ensures that IT resources such as processing power and storage are appropriately allocated.

If you have not started introducing AI into your business and you don’t want to hire a whole team from scratch you might want to consider sourcing a vendor with AI capabilities. There is an increasing number of vendors out there and if offshore vendors are an option you’ll be able to find highly qualified talent at a fraction of the cost of western markets.

Making AI part of your business DNA

Just like any other new technology, ensuring a widespread adoption within your organisation and its culture is a challenge. Considering AI is the most powerful technology known to mankind it has never been as important as it is today to create an effective adoption plan.

The following blueprint for AI adoption can be applied to most businesses.

Stage 1 – Discovery

This is the early stage of AI adoption most business will find themselves in today.

Here it is important to make the most out of your existing resources. Start thinking about the problems you might be able to solve and what data may be required.

Engage some of your existing engineers and ask them to learn about AI and set up an AWS environment to experiment with model templates. Once they feel confident creating basic machine learning models, work with them on designing an MVP. Engage your most loyal customers to test the MVP and capture feedback.

Now you’ll be able to communicate the value AI can bring to the business using the findings of your MVP experiment and align your key stakeholders.

Stage 2 – Engagement

Engage your key stakeholders to map out how AI can help achieve department objectives. Map existing processes to understand where AI can add value, which employee roles will change and how customer experiences can be improved. The existing prototype may be able to offer improvements already. If new AI capabilities are needed, create a data strategy that prepares the business for future advancements. Data partnerships will help to realise your data strategy faster.

Developing an understanding of how AI relates to other technologies is crucial to ensure future relevance. It may be the most powerful technology of them all but you don’t want to miss out on synergies with others such as IoT, VR, big data, and blockchain.

Change management will be required to take your staff on the AI journey. There will be anxiety around job losses. Sure, some jobs may not be required anymore but AI and general company growth will create new ones. Offering transparency around the use of AI within your organisation and training to the roles likely to be affected by changing job requirements will ensure a smooth transition to the next stage.

Stage 3 – Execution

Consider how AI will impact your strategy. Adjustments may be needed now that you have broader buy-in and established capabilities.

Engage your HR team to develop a learning framework so your employees understand current and potential future applications of AI in more detail. Use the fact that staff are now part of the AI journey to develop outstanding customer experiences, efficient business processes, and IT upgrades that set your business up for success for years to come.

This overview is very brief and only covers the basics of AI. If there are any nuances of the topic you would like to hear more about or if you would like to share general feedback, please don’t hesitate to reach out to me on LinkedIn.

Previous post

Brick-bait: three tricks up retailers’ sleeves to lure you back to physical shops

Next post

Los Angeles sues IBM’s Weather Channel App for “deceptively” profiting from user data

Join the digital transformation discussion and sign up for the Which-50 Irregular Insights newsletter.