AI now and into the future
As organisations continue to evolve in the digital age, artificial Intelligence (AI) technology is playing an increasingly pivotal role across a large number of sectors.
This summary report highlights how AI-driven solutions are transforming healthcare, retail, finance, and manufacturing industries worldwide, along with key statistics, insights and approaches to inform your AI strategy.
Key AI Insights:
- AI-powered cybersecurity systems can analyse vast datasets and detect anomalies faster than human experts, which is essential due to (ironically) increased threat from AI attacks according to Forbes
- Accenture estimates that AI has the potential to increase labour productivity by up to 40% by 2035, primarily through the automation of routine tasks, allowing employees to focus on higher-value work
HEALTHCARE
AI-Driven Diagnostics and Predictive Analytics:
AI technologies are revolutionising patient care and treatment outcomes in healthcare. The integration of AI has paved the way for precise diagnostics and personalised treatment plans. Examples include:
- PathAI employs AI to assist pathologists in diagnosing diseases from medical images, particularly enhancing the accuracy and efficiency of diagnosing diseases such as cancer
- Tempus employs AI-driven analytics to assess large volumes of patient data and recommend personalised treatment decisions based on patient-specific data
RETAIL
Recommender Systems and Enhanced Customer Experiences:
Retailers are harnessing AI-driven recommender systems and chatbots to offer tailored product recommendations and exceptional customer experiences.
- Amazon’s recommendation system, driven by AI, contributes to an astounding 35% of its revenue*, exemplifying the impact of AI in the retail sector
- Sephora’s AI-driven chatbots and virtual shopping assistants provide customers with personalised guidance, helping them explore and select beauty products effectively
TECHNOLOGY
- DeepMind AI reduced Google’s data centre cooling bill by 40%, highlighting the profound impact of AI in energy efficiency
- AI technology allows systems to react more swiftly to real-time changes, such as weather, enhancing overall efficiency
- Custom-tuned models for AI are essential, as each data centre is unique in its requirements and operations
FINANCE
Fraud Detection and AI-Powered Robo-Advisors:
The financial sector is benefitting from AI by using it to identify fraudulent transactions and offer personalised investment advice.
- JPMorgan Chase employs machine learning algorithms to continuously monitor financial transactions and identify fraudulent activities in real time
- Wealthfront offers AI-powered roboadvisors that provide customised investment advice, allowing individuals to create and manage diversified portfolios based on their financial goals and risk tolerance
MANUFACTURING
Predictive Maintenance and Supply Chain Optimisation:
AI-driven predictive maintenance and machine learning algorithms are improving manufacturing efficiency and supply chain logistics.
- Siemens uses predictive maintenance powered by AI to optimise equipment uptime and reduce unplanned downtime in manufacturing facilities. This AI analyses sensor data to predict maintenance needs accurately
- Flexport utilises machine learning algorithms to optimise supply chain logistics, enhancing routing, reducing transportation costs, and minimising delays
- Applying AI-driven forecasting to supply chain management can reduce errors by between 20 and 50 percent—and translate into a reduction in lost sales and product unavailability of up to 65 percent
AI technology is transforming organisations globally, offering significant improvements in
diagnostics, customer experiences, fraud detection, and operational efficiency.
These technologies are helping businesses thrive in an increasingly competitive and datadriven world, and the United Kingdom stands as a significant player in the global AI landscape. CIOs and CTOs should continue to explore and invest in AI to keep their organisations at the forefront of their respective industries.
IDENTIFYING AI OPPORTUNITIES AND CREATING AN AI STRATEGY
In our insight piece ‘How to Build a Game-Changing AI Strategy’ we outlined a four-step process to building your own AI strategy. It’s always valuable to see processes come to life, so we have mapped this onto a construction company’s AI discovery journey:
- An employee with knowledge of key AI technologies & the sector led the AI innovation process.
They provided applicable technology summaries to upper management to help align stakeholders with some of the required knowledge and terminology - Peer use-cases were reviewed and workshops conducted with key internal stakeholders and domain experts to ideate and review the case studies.
- Where are the opportunity areas in this company?
- It collects a huge amount of IoT sensor and user app data, but has an inadequate firepower challenge – lots of data with potential insights that aren’t extracted or utilised strategically
- Bottleneck in flow of information – data exists from sensors etc but not all integrated together to be maximally useful, some information silos
- These represent great opportunity to leverage data science & AI techniques
- Some of the key use-cases identified were:
- A virtual chat-bot assistant to help with administration tasks like room and desk booking, using natural language input
- Scanning building schematics & information documents to extract useful data using computer vision
- Automatically detecting faulty equipment and notifying the relevant parties, based on live sensor data (predictive maintenance)
- Automatically adjusting various building climate controls based on a complex predictive climate model & live sensor data
- Evaluate example use cases:
- For example, evaluating “Automatically detecting faulty equipment and notifying the relevant parties, based on live sensor data (predictive maintenance)”
- This is a large pain point that causes customer frustration & costs due to equipment failures and downtime, alongside significant internal costs to the building services company in supporting the customer with initially identifying the fault from diagnosing downstream technology failures. It also increases the energy use for the customer due to operating non-optimal equipment, which is of huge value to the customer. High demand
- The company has recently hired specialists in AI & data science, so they have the skills to tackle the challenge. They have the datasets available for training predictive models, but there is a challenge to ensure all data is correctly labelled to make it applicable for this use case. At this point would need prototype testing to understand how accurate a predictive model would be, given the data available. Assuming good results, the company could quite easily integrate the predictive model into their app platform and workflow. Moderate feasibility
- This solution would provide a significant competitive edge and cement the company as a leader of tech innovation in this sector, contributing to long term growth. It would provide value for marketing and sales, given the exciting nature of the innovation. There is scope to upsell customers for this as an additional service. There would be significant labour costs in testing and eventually deploying the solution, but any developed solution would be highly scalable. High commercial viability
- Select the appropriate (broad) AI tech for each use-case above:
- For example, assessing the appropriate technology for some of the identified use cases:
- “Automatically detecting faulty equipment and notifying the relevant parties, based on live sensor data (predictive maintenance)” – likely an open-source supervised ML classification model, trained from manually labelled historical data, deployed into the existing app microservices ecosystem
- “A virtual chat-bot assistant to help with administration tasks like room and desk booking, using natural language input” – likely a pre-built natural language processing chatbot solution / platform that is customised and configured for the company’s specific needs, such as Amazon Lex, Dialogflow (Google), or Azure Bot (Microsoft)
- Where are the opportunity areas in this company?
- Launch pilots
- The company is planning AI applications to be a large growth area for them, and so have created a dedicated team to manage AI pilot projects (a “centre of excellence”). Minimum Viable Business Cases were created for each potential pilot, to wholistically evaluate all aspects of the idea in the specific business context
- They launched several concurrent proof-of-concept pilots to trial the top priority use cases, which were self-contained and not initially interacting with any live data or services. This involved remapping workflows using end-to-end process maps, monitoring if it was accurate in practice and determining the amount of human input / oversight needed for the new process
- For example, for the “predictive maintenance” pilot, the new workflow had a dedicated fault prediction engine running on the collected building sensor data, that would automatically detect any potential anomalies and push notifications to the configured persons who could then review the data and determine the next steps. Initially this ran in parallel to the existing regular maintenance checks and reactive fault-reporting process, for just one trial building
- Scale up
For example, scaling up the “predictive maintenance” pilot to the wider business involved discussing implications with all potential stakeholders and technology experts whose processes would be affected, trials at more buildings, then fully integrating the fault prediction engine into live production services whilst continuing to monitor early performance and any pragmatic workflow modifications needed
“We’re living through another
industrial revolution and as IT
leaders we’re on the front lines.
I hope this summary report will
prove valuable to you in
developing and building your
own AI Strategy.”
Christopher Renn
Director, ImprovIT
Sources:
