Three strategic steps to get started with AI
- Posted on December 21, 2017
- Estimated reading time 5 minutes
The following blog post was written by Avanade alum Sarah Adam-Gedge.
Science-fiction writers and Hollywood film makers have, for many years, given so much to the development of technology by bringing fictional worlds to life. Often filled with incredible inventions that creatively solve challenges facing the real world at the time of writing, these futuristic stories have the capacity to inspire inventions that advance our civilisation.
While fictional, these stories have occasionally contributed to the realisation of real-world technological advancements. For example, Motorola’s Martin Cooper, credited for inventing the first mobile phone, was inspired by the “Star Trek” communicator. Similarly, Steve Perlman, a scientist at Apple, says he stumbled upon the idea for the company’s hugely successful multimedia program, QuickTime, after watching an episode of “Star Trek: The Next Generation”.
While science fiction has helped inventors go where no technology has gone before, maybe it is hindering technology advancements in organisations. Thanks largely to Hollywood, artificial intelligence (AI) often conjures in our minds futuristic images of humanoid robots and killer computers – think “The Terminator”, “I, Robot” or more recently, “Transcendence”.
Whether or not you’re a fan of sci-fi movies, your perception of AI has likely been swayed to some degree by pop culture portrayals of the technology.
This is partly why AI adoption is a daunting prospect for many organisations. While they are aware of the increasingly pressing need to implement AI, a perception that it’s too complex and too expensive can be paralysing. From our experience advising clients on innovation strategies, there are three tips Avanade recommends organisations consider as they embark on an AI journey:
#1: Learn what AI technologies can work for your organisation now
For the majority of today’s organisations, it makes sense to pilot AI in one of two forms; Robotic Process Automation (RPA); or intelligent automation. With RPA, machines are taught to process repetitive, high-volume, manual tasks that use structured data – like a spreadsheet. This can make it easier for organisations to start their AI journey with RPA, and that is what we’re seeing with many of our clients.
To assess the value of RPA for your organisation, we recommend conducting a thorough review of day-to-day activities and asking several questions: Are workers repeating routine tasks throughout their day? Do these tasks involve predictable decision making? Can the process patterns be taught or learnt with relative simplicity?
If the answer to these questions is yes, there is likely an opportunity to introduce some form of RPA – a move that can significantly reduce time spent on repetitive tasks and free up employees for more complex and rewarding work. Organisations can often achieve very tangible ROI with relative simplicity in a short timeframe with RPA, making it a compelling first step with AI.
Intelligent automation is typically more complex, requiring machines to mimic the learning, decision making and actions of humans in order to translate vast amounts of structured and unstructured data into actionable insights. This is often embedded in a digital agent to help humans intuitively interact with the machine to extract insights.
An example of this is Avanade’s work with a leading consumer goods company, to develop a HR digital assistant which supports with the influx of employees sending routine requests to HR employees. By implementing a fully integrated digital assistant supported by a network of human experts, 90% of queries were able to be solved – freeing up the HR team’s time to work on more complex tasks.
Another example of the power of intelligent automation is Avanade’s work with Woodside Energy, one of the world’s largest oil and gas producers. Avanade has been working with Woodside Energy to implement intelligent automation that has generated millions of dollars of cost savings, increased the lifespan of critical assets and enhanced productivity by being able to prevent unplanned site shutdowns.
To achieve sustained business value from intelligent automation, organisations will need to employ advanced analytics as well as cognitive services like chat bots, object speech recognition and natural language processing. Organisations may choose to progressively embrace intelligent automation over time, but it’s critical that they start now to focus on compiling and organising the data that will underpin these technologies over time.
#2: Just get started, but keep it simple
When starting out on an AI journey, it’s important to begin with manageable projects that deliver on their intended purpose. Once an AI project gets the green light, set out clear parameters for what it will and won’t try to achieve. We recommend starting with a basic prototype – with limited functionality – and then iterating as support and capability are built within the organisation.
For example, if a chat bot can theoretically answer 100 questions, but only understands one specific way of wording those questions, the bot will likely be seen as a complete failure in practice with users. In contrast, if the bot answers two questions and understands 50 ways each of those questions might be asked, it is much more likely to engage users – and in turn lead to them acting as internal advocates for other AI innovations. Ultimately, the more focused the project, the more rapid the results of AI are likely to be realised.
#3: Help employees understand the possibilities of AI for your organisation
In a recent Computerworld article, I shared findings from Avanade’s global research that shows 79% of business leaders believe internal resistance to change is limiting the implementation of AI technologies in the workplace. In that article, I highlighted the need for leaders to change the conversation to address the concerns of employees. Key to this process should be a focus on education and discussion about how AI can benefit your organisation specifically.
You may still be figuring out your AI roadmap, but you can start the process of engaging your employees now. This may include scenarios of how employees will be freed from routine, repetitive tasks to focus on higher-level tasks like customer engagement or innovation. In addition, Avanade recommends clients consider the ethical complexities of AI and innovation, and put in place governance frameworks to ensure the expectations of customers and employees are met on an ongoing basis.