Bringing AI from the Boardroom to Operations: Your Three Options
Generative Artificial Intelligence & advances in data analytics have become the talk of the town. It has become glaringly evident for boardroom executives that Artificial Intelligence (AI) and data science will play a pivotal role in determining their business's competitive advantage. In the UK we are currently one of the global leaders in AI, ranked highly in AI readiness. The UK government will play a pivotal role in supporting the adoption of AI via its National AI Policy, to ensure productivity, growth and innovation can be harnessed by all businesses. However, overall AI adoption remains low at only 35%. The degree of AI adoption effort across sectors varies significantly with legal, technology, IT and telecoms sectors leading the way; compared to hospitality, health and retail lagging behind. Size also matters. Larger organisations have been able to adopt AI & data analytics quicker than small & medium sized businesses. Nonetheless, it is genuinely awe-inspiring to witness the fervour and resources being poured into these cutting-edge technologies. Some facts:
· 68% of large companies, 33% of medium-sized companies, and 15% of small companies have incorporated at least one AI technology
· 1 in 10 businesses plan to adopt AI this year
· Spend on AI & data analytics to rise to £35B by 2025
But here's the rub. Despite skyrocketing interest and investment, many top-level decision-makers find themselves grappling with the implementation of AI and data science into their processes. The challenge lies not in realising potential benefits but in the practical application and integration of these transformative tools. The questions you may need to ask yourself is: Can your business afford to delay AI adoption? If not, which option works for me right now?
Since Emneo’s inception this year, we have spoken to nearly 50 executives from a range of enterprises across multiple sectors. Our proposition is specifically focused on the AI & Data Fabric but our conversations varied based on our expertise of AI implementation options. In our view, figuring out how to operationalise AI and data science effectively, is the elusive key that unlocks the doors to untold possibilities (and prosperity) for these companies. We believe AI & data analytics together should deliver faster processes, personalised data and better experiences for business users and customers alike. To bridge this gap effectively, it's essential to transform the realm of AI and data science from abstract, high-level discussions into tangible, real-world capabilities that can fuel critical workflows, elevate the workforce, drive informed decisions, improve productivity & compliance and empower various applications. Within this context, we believe organisations have three primary avenues to explore, for seamless integration of AI, from boardroom to practical operations:
Option 1: Internal Innovation with AI. Stand up entirely new internal teams for AI development and data science.
An approach once required for the early adopters of AI. This option includes building an AI innovation function focused on critical business problems or objectives. Requires an independent function that consists of AI and Machine Learning (ML) experts that work closely with decision science experts to build unique models and solutions for specific requests. The internal team takes the organisation through a cycle of model development, deployment, scaling and continuous learning. The process can be time-consuming, with a long horizon on benefits realisation and requires discipline in delivery.
The benefits are significant and include bespoke AI innovation, with focused business outcomes delivering solutions to functional or operational issues, improvement of business data and standardised execution of processes. A working AI model depends heavily on consistent and reliable business data, generating a competitive advantage if all the pieces fall into place. Effective AI models can lead to targeted customer experiences and workforce elevation.
In addition to a clear strategy, skills and tech required include an understanding of several Artificial Intelligence tools like Machine Learning, data generation & deep learning models with implementation frameworks, hardware and core computing power all in a project and improvement-driven control centre. The sheer amount of data required to train bespoke models can also be unavailable in some organisations.
Significant capital investment is required to resource the function, in line with an R&D approach and boardroom understanding of being flexible and failing fast. Standing up an AI DevOps and Data Science function can be a costly investment. Business needs to define strategy and provide infrastructure, which as you can imagine, takes time to stand up and implement. As mentioned before, discipline in where to focus the energy of the AI function is critical to success. This option isn’t for the faint-hearted and may not even be an option for SMEs unless you have a taste for large R&D spend.
An alternative option to internal DevOps teams would be to outsource the AI/ML & Data DevOps to a provider, a decision dependent on ROI and business appetite.
Use cases include code writing for application development, predictive models for pricing & costs, personalised product recommendations for customers, sales forecasting.
Option 2: AI & Data Fabric Layer: Utilise pre-built Intelligent Automation platforms and infrastructure from vendors.
Our option of choice for large, medium and small enterprises alike. This option includes creating an AI and Data Fabric around your existing technology stack , without rip-out-and-replace of already sunk technology investments or expensive AI transformation projects. This option is all about our core beliefs of eliminating waste, reducing risk and creating better experiences for customers and staff alike, by wrapping existing enterprise technology platforms in next-generation Automation and AI capability with a rapid ROI.
Businesses can leverage the readily available, tried and tested cloud-enabled Intelligent Automation platforms. Intelligent Automation now comes with built-in AI models or businesses can choose a model of choice from a variety available. What's required for implementation is vetting solutions, sometimes integration work and adjusting processes & policies. Implementation is faster due to leveraging pre-built technology but still needs integration. Benefits include speed and lower costs, whereas downsides appear in the form of less customisation and a reliance on vendors.
Next-gen technology platforms now include AI, ML, Data and Large Language Models and can be enhanced within the guardrails of automation tools, as part of the AI & Data Fabric Layer. The layer is further boosted by containing tools for Process Discovery (to surface transformation and optimisation candidates), Robotic Process Automation (RPA), Automation Co-Pilots, a combination of RPA and generative AI, Intelligent Document Processing and advanced data analytics.
Businesses implementing an AI & Data Fabric with Automation, can increase productivity through understanding processes for improvement and automation, relieving the workforce from repetitive tasks by implementing Robotic Process Automation, and infusing AI, ML, LLM and Data models within RPA to create faster processes and improved decision-making at all levels. Intelligent Document Processing is enabling businesses to now rapidly ingest large swathes of data from structured & unstructured documents into existing systems. This technology revolution is creating a Human + Digital workforce that is focused on creating better experiences for customers. All of this is deployable within weeks, scalable and easy to implement within the organisation. Expect over 5x typical ROI, with significant productivity gains. Also expect a bonus in retaining talent, with the elevation of workforce from repetitive tasks to champions focussing on solving complex issues, resulting in better business outcomes,
Use cases are endless and include data-intensive customer services processes, operational workflows, supply chain and vendor management, traceability and sustainability, scheduling, yield management, and many more.
Option 3: The Hybrid: An approach with a mix of internal talent and external vendor solutions.
A combination approach. Hire dedicated AI developers, data scientists, and other roles to build custom AI solutions but also utilise AI/data platforms and services. Requires strategy for the hybrid model, coordination between teams, and platform evaluation. Implementation incorporates hiring of AI talent procurement and integration of platforms. Benefits realised through this approach combine custom work with speed and lower costs of platforms.
Conclusion:
AI is becoming table stakes, so companies must start operationalising it somehow, even if it is one team inside one business unit, before rolling-out company-wide. Each approach has its pros, cons, costs, and organisational implications. Unfortunately, there is no one-size-fits-all answer. Companies must carefully weigh these options against their specific business context, resources and objectives. With the right strategy and planning, AI and data science can successfully move from theoretical boardroom conversations to practical results improving operational performance. The key is to move beyond theoretical talk and take tangible steps like process analysis, skills development, and selecting partners. With the right strategy tailored to your business context, AI can transform operations through automation, data insights and elevated workforces. But you must act decisively to operationalise AI before competitors capitalise on the advantage. With thoughtful strategy and planning, AI can move from boardroom discussions to operations.
References
1. Hooson M. UK Artificial Intelligence (AI) Statistics And Trends In 2023. Forbes.
https://www.forbes.com/uk/advisor/business/software/uk-artificial-intelligence-ai-statistics-2023/
2. AI Activity in UK Businesses Report, UK Government.
3. AI Adoption in the UK: Putting AI into Action. Techuk.org.
https://www.techuk.org/resource/ai-adoption-in-the-uk-putting-ai-into-action.html
4. Business Insights and Conditions Survey (BICS) QMI.
