Artificial Intelligence (AI) has undergone a drastic transformation in the last 20 years, from a concept to a science fiction staple and today a rapidly evolving domain of data science that is perpetually occupying news headlines with its impacts on industry, the world and our personal lives. Nearly every week, AI is surpassing milestones, overcoming limitations and solving a new problem. However, many business leaders fail to understand how AI can be harnessed to benefit their business and customers. For this reason, businesses are increasing looking to AI consultants who play a pivotal role in planning, developing and implementing AI solutions that unlock new opportunities for growth.
What are the barriers to adoption?
The problem is not a lack of awareness of AI: it is virtually impossible not to be even casually aware of AI and the discourse around innovation, new applications and ethical considerations. Business leaders are all too familiar with the jargon - automation, insights, data-driven decisions – around AI and the espoused benefits of adoption (efficiency, productivity, profitability). In fact, this is part of the problem: business leaders are forced to wade through seemingly endless jargon and struggle to find answers to the questions that matter:
- Which of my business problems can AI help solve?
- How can AI improve efficiency and grow my business?
- How do I prioritise opportunities?
- What is the ROI?
- How do I manage risk and avoid failure?
- How do I implement AI?
- Does my business even need AI?
The challenge facing many business leaders is not lack of awareness but rather understanding how AI is relevant to them and what is required to make the potential business benefits of AI a reality. Finding personalised answers to these questions is difficult – and the reason so much of the discussion of AI is conducted in vague terms or focused on specific case studies – because they are complex, multi-faceted and there is no one-size-fits-all answer. You will not find specific answers to these questions via a quick Google search or YouTube video. The answers to such questions require both a deep understanding of the unique commercial priorities, technology infrastructure, data strategy, business model and cultural values that define your ‘business DNA’ and the expertise to define how AI might fit into this unique ecosystem.
This is where the benefits of working with AI consultants are invaluable. By combining their expertise in AI with an understanding of your business DNA, AI consultants can cut through vagaries and hypotheticals to find answers to those complex and evasive questions.
Here are some of the main ways an AI consultant can assist your business with AI adoption.
1. Identify business challenges that are a good match for AI
When first starting out on the AI journey there is often a temptation to see AI as a solution to every business challenge: to treat every business problem as a nail and AI as a hammer.
Whilst AI has diverse capability and can potentially transform and automate several different processes and workflows across business – from forecasting, design, accounting and customer service management - that does not mean it is the best solution to every problem.
AI is an umbrella term for a vast range of techniques that enable machines to form judgements, make decisions and perform tasks that rival human intelligence. Today, AI is often referred to as ‘narrow’ AI. Narrow AI excels at performing a specific and well-defined task but will perform abysmally at anything else. For example, an AI model designed to forecast customer demand will not be able to generalize its expertise to a seemingly related task, such as predicting customer churn. Training machines to perform such tasks accurately and consistently is effortful, time consuming, expensive and is dependent on access to high-in-demand expertise and vast quantities of high-quality data. Unless the problem you are trying to solve actually requires intelligence, using AI to solve the problem may introduce unnecessary expense, time, complexity and risk to your project.
An AI consultant will aim to mitigate and manage such risk by basing recommendations for adopting AI on those opportunities that are relevant and viable for your business, and align well with the strengths of AI. An AI consultant should first begin with understanding the types of challenges and bottlenecks that limit efficiency, productivity and growth, and combine this with their knowledge of AI, knowledge of your processes and data, and the broader technology ecosystem to define which problems are great candidates for AI and those that are best solved using other techniques, such as rule-based automation.
2. Assess the value and benefits of an AI project
Before commencing an innovation project it’s important to recognize that no business has unlimited time and resources to pursue every possibly opportunity to them. AI is no different. When understanding the art of the possible, part of the challenge for business leaders is assessing the array of opportunities that present themselves and prioritise according to business objectives, value and risk: prioritizing and quickly executing against high value and low risk opportunities (“must-dos”), carefully evaluating and planning high value and high-risk opportunities (“strategic initiatives”) and deprioritizing or completely avoiding those high risk or low value opportunities.
Assessing the risk and value of pursuing innovation opportunities will require pooling the knowledge and expertise of your business and your AI consultant partner. Your responsibility as the client is to provide in-depth insight into the business context and areas relevant to solving the business challenge, such as strategic vision, goals and KPIs, data management, infrastructure and governance, systems and operations. The AI consultant’s responsibility and challenge is to apply their expertise in AI, data and technology to your business context to evaluate which opportunities will result in the greatest value, whether value is expected to be created in the short or long term, the prerequisites to pursue these opportunities and their unique risk profile.
3. Translate problems into strategy and action
Before you engage an AI consultant, it’s important to have clear expectations regarding the skills and competence expected of a consultant and the outcomes from working with one. An AI consultant is not someone with only elementary knowledge of the principles of AI and enough familiarity with different AI use-cases that allows them to provide generic observations and recommendations as to which AI domain is broadly relevant to a particular challenge. After all, anyone that has read a few articles on artificial intelligence or machine learning would likely be able to tell you that the technique for identifying and classifying the contents of an image is machine vision – but this offers very little practical value to your business.
AI consulting is all about enabling you to leverage AI to improve your business. Therefore, an AI consultant is an expert who armed with the knowledge of your business DNA and AI capability will be able to provide tailored insights and recommendations as to where and how AI could be applied and is able to create a strategy and roadmap to turn a concept into a real-world solution. Working with an AI consultant will accelerate and de-risk AI adoption by focusing on two critical elements of the AI project workflow: strategy and build and implementation.
An AI strategy should clearly define the scope and nature of the problem(s), explain the relevant solution domains and the specifications of the proposed solution. It should also outline data pre-processing requirements to transform raw data into a viable input for AI algorithms, evaluate the project risk factors and mitigation tactics, and importantly, the time, human resources and investment required to execute the project. How well established and validated your use-case(s) is will influence the project duration, complexity and risk. For example, strategies for business challenges that have many validated use-cases or leverage off-the-shelf technologies are likely to be more detailed, present lower risk and be cheaper in the short term. Conversely, overcoming a challenge that is highly complex and innovative will likely require significant research and development (R&D) to develop this new capability. Projects involving significant R&D generally involve exploring how existing techniques can be adapted to solve a different problem or developing a new technology from scratch. This approach entails less certainty and higher risk as feasibility is not validated, longer project times, and higher upfront cost, though if successful will also result in IP creation and ownership. An AI consultant should be able to anticipate and clearly communicate the main areas of project risk, the impact on project viability and advise on how to manage this risk.
Finally, a consultant should have a team with all the skills to build and implement the strategy. Depending on your scenario and your requirements, this may be a relatively straightforward task of integrating third party off-the-shelf solutions into your existing data infrastructure and business workflows, or a custom approach that requires training a new custom model from scratch. If your aim is to develop a custom AI model, the development roadmap should begin with a prototype or proof of concept (PoC) that aims to validate technological feasibility; essentially proving the idea works in practice via a model able to generate accurate and relevant output that addresses the core problem. Once the PoC validates the solution to the business problem, the knowledge acquired from this initial development stage will inform the data engineering and software development requirements to productise and integrate the model into your business operations. Once deployed, focus then shifts to managing the model lifecycle via machine learning operations or MLOps (i.e., iteratively retraining, redeploying, and retiring the model according to performance or event-based triggers) to ensure the model is continually learning, improving and performing to benchmarks.
Artificial intelligence is a complex, high potential and rapidly emerging technology domain that promises to transform business, the world of work, governments and society. Whilst many businesses are already beginning to leverage this technology to generate business value, many more struggle to decipher the jargon and understand the requirements and process for adopting this capability. This can be attributed to AI strategies and projects being at the intersection of many different commercial and technical elements, with company vision, operating model, business priorities, data strategy, data governance, infrastructure, security and cultural adoption being a handful of key elements. Without sufficient knowledge and expertise of these areas, should you manage to launch an AI project there are critical blind spots that can undermine project success. By working with a trusted AI consultant or consultancy, you can cut through all the jargon, generalisations and speculation surrounding AI and start having tailored conversations about where and how AI can create value for your business and craft a tailored plan of action.
Written by Clayton Black