Driving transformation with AI in business: A step-by-step strategy for implementation!

AI (Artificial Intelligence) is changing our life, the way we work and the way we do businesses.
As a manager, entrepreneur or business owner, what does this mean? How can we use AI in business, leveraging its positive impacts while minimizing its risks?

To start with, it is important to close “our eyes and ears” for what is being published, specially in social media, and look for credible sources to learn and make our own mind about it. There is a lot of misleading information, and several scams too.

An idea of what AI means in theory is almost self-explanatory, but what does this mean in practical terms? How can we use this disruptive technology to create real positive impact?

Without going into the history of Artificial Intelligence, or into a deep discussion on how it works, let’s dig instead on the status of this technology development.

The Evolution of AI: From Narrow AI to Super AI 

Specialists tend to agree with a technologic development classified in 3 stages:

1. Narrow AI – Artificial Narrow Intelligence (NAI)

This state represents AI systems that are designed to perform a single task or a limited range of tasks. Examples of narrow AI include: chatbots, recommendation systems and voice assistants.

2. General AI – Artificial General Intelligence (AGI)

AGI, represents an AI that has the ability to understand, learn, and apply its intelligence broadly and flexibly, similar to a human being. Although there are not yet any instances of true AGI, there are many advanced applications of specialized artificial intelligence that go already in this direction. Examples of this can be: systems for healthcare diagnosis and treatment prescription, self-driving cars and drones that avoid obstacles and make decisions in real-time or financial services with systems that analyze market trends, predict stock movements and automatically buy and sell.

3. Super AI – Artificial Superintelligence (ASI)

ASI represents a stage of AI development where machines surpass human intelligence and capability in all respects. This includes creativity, general wisdom, problem-solving, and autonomous self-improvement, leading to rapid advancements beyond human control or understanding.

As we can understand from the above description, we are still in stage 1 but moving at a fast pace to stage 2.

It is important also to consider that this  technological evolution is limited not only from mathematical algorithms, data availability and software development, but especially from hardware constraints that blocked AI in the past and still impacts further developments. This means that  if it ever happens, we are still far away from stage 3 and the risks of the so-called “singularity event” where the “machines” would control human beings.

AI in business: Key considerations for AI integration

Technology apart, let’s focus on how AI in business can already be used to create a positive impact.
While adopting any new technology, we need to look first for improvement areas, where AI could be relevant. We can start by answering questions such as:

• What areas do we have in house that need improvements?
• Where are our main difficulties and struggles in the market?
• What is blocking our growth?
• Is our competition using AI, how?
• What areas of my business are not profitable or sustainable?

With these answers in mind we can then identify areas where AI usage can make a difference and find priorities and alignment with our business strategy.

For Ai in business to be successful and independently of the areas we are going to use it, please consider at least 8 prerequisites that need to be in place for its effective implementation.

1. Repetitive activities/decision processes that can be automated;

2. Activities/decision processes where real-time info can deliver significant business impact;

3. Enough quality data (recurrent and unbiased info so that AI algorithms can learn and be trained). This data can be partially external too. Defined.AI, a portuguese startup founded by Daniela Braga, is a data marketplace for several AI and ML (machine learning) solutions;

4. Tools, software and hardware requirements identified by an internal/external team;

5. Human resources with business knowledge to specify the needs and validate the results (in the initial phase and during usage, since algorithms tend to get stuck, especially if fresh data is not fed into the system).

6. Human resources that can design, develop and implement the needed AI solutions (it can be an external team also.)

7. All the numbers to do the math, so even without considering the setup investments, applying AI will be a cost-effective solution.

8. In-house or external support, with a clear understanding of AI usage risks and how to mitigate them.

Understanding AI’s role in business transformation

While implementing AI in business we can also look at 3 main drivers for which its impact will be our new Superpowers!

For further dive into How Businesses Are Using Artificial Intelligence In 2024, check on this study from Forbes Adviser.

Step 1 – Usage of Off the Shelf AI tools

If you are entering in this area, start slowly with contained business cases and use the off the shelf solutions. These tools can significantly enhance productivity by automating routine tasks, optimizing workflows, and providing insights through data analysis for the decision making processes. In short, these tools can help us on our Superpower 1.

Examples of applications of AI in business, already available today, can be:

CRM Systems with AI Integration:

1. Salesforce Einstein: Uses AI to analyze sales data and predict customer behavior, recommend actions to sales reps, and automate tasks like data entry.
2. HubSpot: Uses AI for lead scoring, predicting which leads are more likely to convert, and for content optimization to improve engagement rates.

Chatbots and Virtual Assistants:

1. Intercom and Drift: Use AI-powered chatbots to interact with website visitors, answer common queries, book meetings, and route complex issues to human agents, enhancing customer service efficiency.
2. Google Dialogflow: Uses AI to build conversational interfaces for websites, mobile apps, and messaging platforms, improving customer interaction.

Human Resources and Recruitment Tools:

1. LinkedIn Talent Insights: Uses AI to analyze large datasets to find suitable candidates and predict hiring trends.
2. HireVue: Uses AI to assist in screening candidates by analyzing video interviews, assessing candidates’ performance identifying the best fits for a role.

Marketing Automation Tools:

1. MarketMuse: Uses AI to analyze content and suggest improvements for SEO, helping marketers create content that ranks higher in search engines.
2. Adext AI: Uses AI to optimize advertising campaigns on platforms like Google Ads and Facebook by adjusting the ads’ parameters to reach the most relevant audiences.

Financial Analysis and Fraud Detection:

1. Kount: Uses AI in fraud prevention by analyzing transaction data to detect and prevent fraudulent activities in real-time.
2. Xero: Uses AI for automated bookkeeping, categorizing bank transactions, and reconciling accounts, which saves considerable time for finance teams.

Supply Chain and Inventory Management:

1. IBM Watson Supply Chain: Uses AI to predict supply chain disruptions and optimizes logistics to a more efficient supply chain management.
2. Blue Yonder (formerly JDA): Uses AI in inventory management to forecast demand and adjust inventory levels accordingly, reducing overstock and stockouts.

There are several solutions to apply AI in business in the market and each day new ones are launched. The famous ChatGPT is also a powerful tool that can assist employees on daily activities. OpenAI has an enterprise version which is important to consider since it comes with proprietary data protection.

Step 2: Usage of low code, no code tools

In a second phase (Superpower 2), we could start customizing and developing our own tools, tailored to our unique business challenges. Also here we have market tools, data and frameworks to assist us, such as:

Microsoft offers several low-code and no-code tools that enable us to develop AI applications with minimal programming expertise. These tools are designed to democratize AI development, making it accessible to a wider range of users, including those without deep technical backgrounds. Here are some specific examples:

Microsoft Power Platform:

1. Power Apps: Low-code development platform used to build custom business applications, for functionalities like text recognition, object detection, and prediction models without writing code.
2. Power Automate (formerly Microsoft Flow): Creates automated workflows between applications and services. It supports AI through pre-built AI models, enabling automation of tasks like image recognition and form processing.

Azure AI Services:

1. Azure Machine Learning Studio: Allows us to build, test, and deploy machine learning models using a drag-and-drop interface.
2. Azure Cognitive Services: Pre-built AI services and APIs that can be integrated into applications without deep AI expertise. Includes services for vision, speech, language, decision, and web search.

Microsoft AI Builder for Power BI:

• Power BI: Tool that integrates AI to provide advanced analytic capabilities like text analytics, image recognition, and automated machine learning models. We can incorporate these AI features into reports and dashboards without needing to code.

Microsoft Forms Pro:

• Forms Pro: An advanced survey tool that integrates with Office 365. It uses AI to analyze survey responses and provide insights like sentiment analysis and keyword identification, helping us to better understand customer feedback for example.

Google also offers a range of low-code and no-code tools specifically designed to facilitate AI application development, for example:

Google AppSheet:

A no-code platform used to create mobile apps from data sources like Google Sheets, Excel, and more, without writing any code. It includes AI and machine learning features, such as image recognition, OCR (Optical Character Recognition), and predictive modeling, making it easy to integrate advanced functionalities into custom apps.

Google Cloud AutoML:

• AutoML (machine learning) Tables: Enables us to automatically build and deploy state-of-the-art machine learning models on structured data, with minimal machine learning expertise. It’s a low-code solution, offering a user-friendly interface to train, evaluate, and deploy models based on a dataset.
• AutoML Vision, AutoML Video Intelligence, AutoML Natural Language, and AutoML Translation: These are specialized versions of AutoML designed for specific tasks (image analysis, video analysis, language understanding, and language translation, respectively). They provide a user-friendly interface for training custom models on our data without deep machine learning expertise.

Google Cloud AI Platform:

This AI Platform Prediction allows for the deployment of machine learning models in the cloud. It’s part of Google’s AI Platform, a suite of tools and services for machine learning practitioners, but it can be used with minimal coding if the models are already created, such as those built with AutoML.

Google Data Studio:

While not exclusively an AI tool, Google Data Studio allows for easy visualization and data analysis, which can be integrated with BigQuery ML. This integration enables users to run machine learning models directly on BigQuery datasets and visualize results in Data Studio, all with minimal coding.

Google and Microsoft, also give access to libraries and data to be used for several purposes. Note that some of the above solutions may still be in beta version and not available for all geographies.

Step 3 – Development of our own AI tools

Finally for Superpower 3 – develop our own specific solutions from scratch – we need to gather a team of highly skilled mathematicians, developers with knowledge in Python or R languages, IT infrastructure management and resources that understand very well the business challenges that we want to solve. A partnership with specialized development centers might be a good solution to help staffing the team.

This last phase, due to its costs, will probably only make sense whenever we talk about solutions that solve problems of an overall industry or for new high-end products/services.

In conclusion for AI in business: what can we expect?

Although we are in a hype phase with a lot of buzz, positive and negative, AI is here to stay, and we should take a deep look at what this technology has to offer.

As a disruptive technology it will shake our way of living, bringing opportunities and risks.

With high levels of investment being allocated into the AI ecosystem, this revolution will be accelerated in the years to come.

Although we can’t stay still, we need to approach it with a careful step-by-step strategy.

We will probably need assistance every step of the way, but the strategy of “wait and see” will be a fast walk to failure, so let’s start now!

Note: This text is an original one. Besides the definition of technological development from AI specialists, all other presented concepts represent my opinion about the topic.