Can AI take your business performance to a new level?

According to the 2019 Accenture report AI: Built to Scale, 84% of c-suite executives believed they must leverage artificial intelligence (AI) to achieve their growth objectives. Yet 76% acknowledge they struggle when it comes to scaling it across the business, so fail to realise the desired benefits. 

AI is already part of our everyday existence and its use is growing at pace. Today, you will probably have experienced AI in action through LinkedIn, Spotify or Alexa, all tailoring content to improve your personal experience. A number of industry areas are thriving and reaping the rewards from using AI. In the financial sector, as early as 1987, banks were using AI systems to detect fraud with debit cards, and AI-powered electronic trades now account for almost 45% of revenues in cash equities trading.

Covid-19 has rapidly moved consumers and businesses to digital channels with organisations adopting and scaling AI and analytics much faster than they previously thought possible, quite simply because they’ve needed new answers or ways of working. But many other organisations are still wrestling with the basic data fundamentals of having a trusted view of their customers or business ownership of the data.

In this feature, we look at how companies can start to use AI to understand how it could drive their business performance

What do we mean by AI, and why is it different to the analytics processes we’ve been using for many years?

Data analytics is the process of transforming a raw dataset into useful knowledge. AI has the potential to supercharge this by automating the data analytics process. Tasks can be performed that ordinarily require human intelligence, insights can be gained from far larger datasets, that are more accurate than traditional approaches, and often delivered in near real-time.

In simplistic terms, AI can deliver business benefits in three ways: automating business processes; gaining insight and driving action through complex data analysis; and engaging with customers and employees. These areas can broadly be defined as follows:

• Process automation: the automation of digital and physical tasks – typically, back-office administrative and financial activities using robotic process automation (RPA) technologies. Here, the ‘robots’ (code on a server) act like a human inputting and consuming large volumes of information from multiple IT systems and can fully automate, or accelerate, processes in tandem with human interventions.
• Cognitive insight: brings together a number of intelligent technologies, including semantics, AI algorithms and a number of learning techniques such as deep learning and machine learning to detect patterns in vast volumes of data, interpret their meaning, and if required, drive action. This has been likened to ‘analytics on steroids’ and is being used in many ways, including to:

1. Predict what a particular customer is likely to buy.
2. Identify credit fraud in real-time and detect insurance claims fraud.
3. Automate personalised targeting of digital ads.

• Cognitive engagement: solutions that engage employees and customers using natural language processing chatbots, intelligent agents, and machine learning. This category includes:

1. Intelligent agents that offer 24/7 customer service addressing a broad spectrum of issues from password requests to technical support questions.
2. Internal sites for answering employee questions on topics including IT and HR policy.
3. Product and service recommendation systems for retailers that increase personalisation, engagement and sales – typically including rich language or images.

What are the benefits of using AI in business?

If you are able to scale the above activities across the business, the McKinsey survey, The state of AI in 2020, identified three resultant benefits for the high performers in AI adoption:

1. Better overall performance. companies are seeing more EBIT contribution from AI experience and better year-on-year growth overall than other companies.
2. Better overall leadership. C-suite decision-making and performance is rated more highly by staff and shareholders.
3. Greater resource commitment to AI. More of their digital budgets are invested in AI and they are more likely to increase their AI investments in the next three years. This continues to fuel the benefits above, and also creates the ability to develop AI solutions in-house, as opposed to purchasing solutions, which further embeds the data capabilities in the business.

The step change in performance occurs when the business moves from the ‘proof of concept’ phase, to the ‘scaling’ phase. Accenture found in their 2019 AI report that companies strategically scaling AI have nearly twice the success rate and three times the return from AI investments compared to companies pursuing siloed ‘proof of concepts’.

What can marketers do to prepare before implementing AI?

With data, the old adage of ‘rubbish in, rubbish out’ is generally true. Organisations that are thriving through their use of AI have the core data foundations in place, namely:

• They have recognised the importance of business-critical data, identifying the key financial, marketing, consumer and master data as priorities, and have invested in the data quality, data management and data governance frameworks for these.
• They also have multi-disciplinary teams throughout the organisation that promote and enable the use of data by the business, with clear sponsorship from the top ensuring alignment with the c-suite vision. This enables faster culture and behaviour changes in the organisation. This is in contrast to those still in the ‘proof of concept’ phase, more likely to rely on a lone champion within the technology organisation to drive AI efforts, so unable to scale and generate significant benefits.

So, the starting point is to understand how data, and then AI, could most benefit the business areas or strategic business priorities. Without many of the core foundational elements above, it will be hard to move from the enthusiastic ‘proof of concept’ phase to ‘scaling’ to drive business value.

Winning hearts and minds in the business moves you from being the lone drummer, to part of a synchronised marching band. To achieve this, you need to build a hypothesis about what you want to achieve and how it helps the business, plus some evidence to show how you may realise it. Then the communication process begins to build your supporter network, along with improving the management and governance of your key data.

With that in mind, how can marketers actually implement AI to improve their business performance? This may be completely new to many marketers, so where do they start?

In building any data strategy, there’s a four-stage process:

1. Understanding the business drivers: aligning the goals of the business with the data activities, and communicating with the business to identify the long list of ‘ideas’.
2. Analysis and prioritisation: creating an agreed set of prioritised opportunities that are then investigated further, with evidence built from existing sources or through ‘proof of concepts’.
3. Assessing the current state: the ability to deliver these is assessed in parallel, including identifying key data assets and their quality/ completeness, availability of the data and existing data management practices, the business processes these drive, technology assets capabilities to deliver the specific tasks, skills, abilities and capacity of the data team to deliver the future state, and the organisational readiness for the proposed change.
4. Data/AI strategy: the above analysis is then built into a recommendation or business case.

The first stage is to identify business issues or opportunities where AI could provide a superior solution. You then move to building a prioritised portfolio of projects around these.

According to McKinsey’s 2020 AI survey, the largest AI-driven revenue increases come from pricing and promotion, customer-service analytics, sales and demand forecasting, and inventory and parts optimisation. From a marketing perspective, start by looking at the customer experience, retention and acquisition activity and see how data insights, and potentially AI could improve on your current performance.

Royal Mail’s employed AI in many areas including building models to refine customer segmentation and to predict customer churn. Pitney Bowes uses AI to increase the ROI of its prospecting activity. Amazon and Netflix built their businesses using the power of recommendation engines. If you have a digital route to market, this AI technology can now be purchased as a ‘software as a service’ (SaaS) offering for a few thousand pounds per year. Chatbots to support your customer service team are a similar price point, with both being reasonably simple to integrate with your existing data and digital platforms.

In a tech business, high growth software revenues are typically valued at over 10 times that of a tech service company. Therefore, as well as customer demand, there is a direct financial benefit to a company in creating these AI powered SaaS solutions to change its valuation metrics. This means more and more will become available, reducing the cost and increasing the ease of access and implementation of these advanced technologies.

What’s the best way of integrating AI into an existing martech stack?

Because AI technologies typically support individual tasks rather than entire processes, scale-up almost always requires integration with existing systems and processes.

It’s really dependent on the AI solution to determine how it needs to be integrated. This should be assessed by the relevant technical experts in the planning or pilot phase of the project. If the application depends on special technology that is difficult or costly to source, that could impact scale-up. But, as noted in the previous section, that risk should be reducing as more SaaS solutions become available.

What pitfalls should marketers look out for when using AI to improve their business performance?

The key to delivering significant AI driven results is being able to scale initiatives. To achieve this, you will need a supportive organisational structure including senior executive sponsorship, foundational data capabilities to allow the data to be accessed and trusted, and employee adoption of your insights. Without these enablers, it becomes very hard to make an impact on business performance.

Five tips to help you succeed are:

1. Don’t disappear into the woodwork, and then suddenly appear with a solution looking for a problem. AI solutions typically need to be developed or adapted in close collaboration with business users to address real business needs and enable adoption, scale and real value creation.
2. Start with the datasets that are easy to get and provide value. Go for the 80% solution first, or the ‘common sense’ approach. Successful programmes start with the data that are easily accessible in one system, or in systems that are already communicating well with each other. And try to deliver some value in the first three months.
3. Make sure that the insights you want to make available are convenient to access and easy to understand for the business. As AI increases its responsibility in the decision-making process, it’s important to provide explanations of the reasoning behind the decision. According to research from BCS,3 the Chartered Institute for IT, 53% of UK adults have no faith in any organisation to use algorithms to make judgements about them. Explanations help to build peoples trust in the conclusions, and may also be required for regulatory compliance.
4. Skills availability is a top barrier to achieving success in AI. You may be able to drive a car, but flying a helicopter is very different, so if you don’t have the required skills and knowledge, look at short-cutting the process by bringing in third-party expertise to work with your teams. Speed to market, or speed to scale, are one of your big risks to overcome.
5. Wherever your data is flowing, particularly when using third-party suppliers, make sure it meets your corporate security and compliance requirements and that you have the appropriate controls in place. Evidencable assessments of this are a good best practice approach.

How quickly can marketers expect to see ROI? Does AI take time to work its magic, or can it start showing results immediately?

To successfully scale, companies will generally need defined data processes, owners with clear data accountability and established leadership support with dedicated AI champions. Creating this structure and governance doesn’t happen overnight.

But you need to start somewhere, and with any data activity, you should challenge yourself to create some value within the first three months.

The McKinsey 2020 AI survey observed: “It’s also clear that we’re still in the early days of AI use in business, with less than a quarter of respondents seeing significant bottom-line impact. This isn’t surprising, achieving impact at scale is still elusive for many companies not only because of the technical challenges, but also because of the organisational changes required.”

How soon significant benefits come through will depend on the opportunity you are pursuing, but the Covid-19 challenge has shown how fast companies can change and implement new ways of working when they have to.

What advice do you have for any marketers thinking about using AI? Is it always a good move?

The key is to understanding whether AI will benefit your business performance is to initially answer the following questions for each of the project ‘ideas’:

• How well does a project idea align to the business strategy (strategic alignment)?
• What is the positive value impact for customers and/or profitable financial growth for your business (value)?
• How difficult and costly will it be to implement the proposed AI solution, both technically and organisationally (cost)?
• Would the benefits from launching the AI solution be worth the effort (opportunity size)?
• Would the business or customer be exposed to risk if the activity didn’t happen (risk)?

If an opportunity exists, learn to walk before you can run. Establish the basic data foundations, start small, fail fast and consider how you’ll scale. Building a ‘data culture’ where data is at the centre of ‘the collective conversations of an enterprise’ will almost certainly improve business performance and is key to building a data-driven business. Whether that requires AI will depend on your specific business opportunity.

In general, if you’re a large company, or need to access a lot of data, you should be exploring the potential to use AI to drive your performance, as there will be so many opportunities.

If you’re an SME with lower data volumes, there’s the potential to create game changing advances for your business. Your key question is ‘what game do you want to change?’

Overall, the journey won’t be simple, but with the right planning and development, AI should drive a new level of work satisfaction, productivity, and prosperity. 

Sources

1. ‘AI: Built to Scale’, Accenture, (November 2019).
2. ‘The State of AI in 2020’, McKinsey Digital, (June 2020).
3. The Public’s Trust of Computer Algorithms survey conducted by YouGov for BCS, the Chartered Institute for IT, (September 2020).
4. DAVENPORT, T. H. and RONANKI, R., ‘Artificial Intelligence for the Real World. Don’t start with moon shots’, Harvard Business Review, (January–February 2018).