Augmented analytics, deep learning, predictive analytics, machine learning - exciting new areas of digital innovation and transformation that aren't so far off into the future as they once seemed.
The latest advancements in cloud computing and many ongoing artificial intelligence-led transformation initiatives are increasingly proving to go hand-in-hand when it comes to enabling success. Approximately 85% of AI applications currently deployed and in production will leverage public cloud platforms by 2021, according to Gartner.In addition, a extensive survey by PWC revealed 67% of business leaders actively identifying the future potential of AI to automate processes and optimise efficiency and productivity, with 34% of respondents already seeing successful results by using some form of digital assistants.
While IT is the department most interested in leveraging AI technologies (67%), customer service (44%) and lines of business (33%) across industries are also pushing for AI.
If your ongoing digital transformation initiatives is heading towards AI, it’s essential to get your cloud in order - or begin examining its AI-based hosted services and solutions.
Here’s 6 essential steps to prepare for your AI adoption and keep competitive.
#1 - Take the time to foster a cultural and technical shift
Organisational agility, improving customer experience and reducing costs are the top 3 reasons businesses are investing in AI, according to Gartner’s annual enterprise survey.
Artificial intelligence is such a big ticket transformation item that over 81% of respondents planning an AI project were confident they would launch their initiatives by 2019.
Unfortunately, while the interest and intentions have been high in the past few years, not many of those companies had real success.
The same survey revealed just over half (54%) of data science projects between 2015 and 2017 focusing on driving AI initiatives never went beyond pilot or reached full production state. Why? The reality is many organisations underestimate the complexity of AI technologies.
Yes, it’s true AI can render new insights, transform our decision-making and drive better business outcomes. But using AI to add value demands a thorough strategy to deploy it and an organisational-wide cultural shift to help the whole company understand how it aligns with their desired outcomes.
As leaders in the business, you must foster the mindset shift required for AI to be embraced as both a business process and driver of future innovation.
Be proactive in managing challenges that arise, build trust for the AI field gradually and support those with new roles and responsibilities in the space by fostering open dialogue between end-users and management and let them see its big value beyond technology enablement.
#2 - Be aware of internal skills gap and identify your leaders
Many digital-led organisations in 2019 will most likely not have the capability upfront to realise their AI initiatives without external service providers, vendors and new talent.
Approximately 47% of respondents in Gartner’s 2018 CIO survey revealed the lack of specialised skills in artificial intelligence and machine learning is their biggest pain-point going forward, making AI talent acquisition the main roadblock to AI adoption to overcome.
Beyond hiring new specialists, businesses need to plan for the transfer of knowledge from partners to your IT and end-users throughout the implementation of first-wave AI projects to ensure internal skills are cultivated, developed and able to be used for future initiatives.
Having an awareness of the internal skill gap and adequately preparing for knowledge transference is crucial to ensuring AI and ML projects continue to be realised and that their potential value is clear to all. Leaders also need to nurture new talent to guide existing teams through best practices and model development and act as a champion for its future benefits.
One other key element of a smooth AI adoption is identifying existing leaders in your organisation with the hands-on experience and knowledge necessary to champion departments past the initial transition phase and navigate users through AI’s many adoption pain-points.
Having these employees on-board at the forefront of your pilot alongside newly hired AI specialists is invaluable to driving acceptance and long-term success with your AI solutions, and the importance of leveraging and using training resources as part of your overarching strategy.
#3 - Cleanse, integrate and store your data before AI adoption
Deloitte’s State of AI in the Enterprise survey revealed over 39% of companies believe data management issues are in the top 3 biggest challenges they face with AI projects.
To fully realise the benefits of AI-fueled transformation, businesses need to focus on data management, data governance and data storage - core building blocks in building your enterprise AI engine, deriving autonomous insights and delivering reliable data for algorithms.
AI enables us to process and act on data in ways not possible with conventional solutions. It’s therefore crucial that data is being appropriately tagged and cleansed to feed reliable information for your AI models. Many organisations without data analytics maturity will also have collections of structured and unstructured data, spread out in disparate silos across on-premise (legacy) and cloud-based (digital) sources, often managed by different teams.
Buried within these assets is both valuable information, and redundant or otherwise low-quality data that can skew your machine learning & AI implementation. It’s important before kick-starting any AI project to adequately consolidate your data sources and organise them into an optimised data lake or data warehouse for enterprise-wide data cleansing, curation and storage.
Philip Pokorny, CTO at Penguin Computing, an AI and ML specialists, emphasises as such in the company’s white paper - “Improving algorithms is important to reaching research results”. According to Pokorny, “Without huge volumes of data to help build more accurate models, AI systems cannot improve enough” to achieve their AI objectives, which is why “fast, optimised storage should be considered at the start of AI system design.”
Taking the time to optimise AI storage means high-quality datasets can be integrated together to iron out inconsistencies in your samples and obtain the accuracy required for ML projects. Without this basic level of data preparation, you run the unnecessary risk of producing result based on inaccurate, unreliable information - potentially devaluing your AI initiatives with users.
#4 - Don’t worry about ROI out of the gate - start small
It’s easy as business leaders to focus on the Total Cost of Ownership (ROI) and Return on Investment (ROI) when adopting new (and expensive) technology.
But with AI being such an advanced and ever-evolving field of digital transformation, it’s best practice to take the necessary time to pilot new projects with the right strategy, assess which AI services fit your business’s objectives best, and allow your teams (end-users) to be involved in the process to remove the pressure of quick ROI.
Implementing artificial intelligence into your company’s day-to-day operations or service offerings should start small to help staff get familiarised and learn the in’s and out’s.
Gartner recommends exposing business users to available AI options to get employees on-board with the potential of AI - Alexa, Cortana, Siri, and drones - and give them a pressure-free environment to experiment with the tech without expectations for an immediate return on investment.
Gaining ‘soft’ AI-driven improvements in processes and customer experience are a lot more achievable in this earlier phase of transition than ‘hard’ outcomes such as direct financial gain.
#5 - Digital transformation comes before AI adoption
Companies that have adopted both cloud computing and big data have been found to have experienced the most success in their AI solutions, according to McKinsey.
Cloud computing and artificial intelligence go hand-in-hand, with 85% of AI applications deployed and in production set to use public cloud platforms by 2021, as per Gartner.
The fact is companies cannot jump into deploying AI solutions without having first modernising their infrastructure and ways of working through digital transformation. McKinsey found that businesses that have moved into the digital and have the necessary experience have 50% higher odds of generating profit using AI compared to those that don’t.
There’s no mistake that AI is one of the biggest game-changing technologies out right now.
But the reality is the best success comes with a strong cloud and data platform foundation first.
#6 - It’s not about the AI platform, but your desired outcomes
A common pitfall in adopting AI technology solutions is the lack of focus on which areas of business to explore using AI, deep learning, augmented analytics and machine learning.
On top of that, many organisations get caught up on which service, tool or cloud platform has the best AI solution to leverage, and lose sight of their main objectives and reasons for AI - to deliver their desired business outcomes.
As McKinsey notes, a strategic approach is key to accelerate your AI journey. The sheer range of AI tools in the modern digital transformation era have long proven to solve a variety of business problems (automated chatbots for customer experience, pattern detection for predictive maintenance, etc) which makes it easy for companies to consider a portfolio-based approach to AI adoption across three distinct time horizons:
- Short-term: Use cases with proven technology solutions today that can easily scale across the modern business to drive meaningful bottom-line value.
- Medium-term: Use cases with experimental technology that’s promising but still developing (such as deep-learning video recognition) that can be used to support AI’s broad future value.
- Long-term: Use cases developed with experts or partners that focus on relieving very specific pain-points using bleeding-edge AI technology, and that are ideal for first mover’s advantage on your unique business problem.
It’s important to keep in mind that the reasons for adopting AI is to realise business outcomes and not technology enablement.
Having an understanding of the main use cases to adopt AI in your business will keep things on track and the value of the technology for your outcomes clear.
How to prepare for AI adoption: Next steps
Choose your AI solutions based on compelling use cases, such as leveraging automation in data exploration processes, delivering exceptional customer experiences with chatbots and natural language processing (NLP), or deriving new value from analytics from machine learning algorithms not previously possible - not just to have it.
Adopting AI in any business is a journey that's best bolstered by adequate planning and assessment. If you need expert assistance in assessing your organisational readiness for AI or to prototype potential AI-powered solutions in your business, reach out to the team at Xello and view our tailored AI service offerings to get started.