How to take a sustainable and scalable approach to AI

London Tech Week blog by Modular Data

Every man and his dog wants to do AI right now. And it’s not hard to see why. The wide-ranging applications of AI, including Large Language Models (LLMs), means that it can be used by nearly every industry to innovate, improve efficiency, and provide a competitive advantage. But moving from concept to solution isn’t straightforward, and if AI projects aren’t aligned with business needs, it can be hard to scale value.

Turning experimental AI models into reliable and scalable business solutions can be a significant challenge, but it’s much easier if they’re implemented with a holistic data strategy and an eye on future use. Operationalising AI transforms its value from single use to multi-faceted, which improves its ROI. But how can we operationalise AI from the outset?

Operationalising AI

Operationalising AI involves transforming experimental or ad hoc models into reliable, scalable, and maintainable systems that deliver consistent and long-term value. This transformation can be achieved systematically; for instance, by establishing governance frameworks and ensuring data readiness.

Reengineering business processes is essential for properly integrated AI. This includes:

Identifying key processes that can benefit from AI
Analysing and redesigning these processes to incorporate AI technologies
Ensuring data quality and relevance.

Taking these steps prior to implemention ensurse that AI optimises processes to improves efficiency and effectiveness, rather than amplifying existing inefficiencies.

Data Governance supports Operationalised AI

Proper data governance is also crucial for operationalising AI. Establishing a Design Authority helps to set AI standards and review and approve AI designs. This can help with consistency, quality, and adherence to best practices; but comes with its own challenges.

A top-down governance approach can run the risk of being too vague. This leaves business domain experts in the stick situation of being responsible for the governance transformation plan without the authority to execute change. They may not have been involved in developing the plan, and their objectives may not align with it, which renders it useless. A top-down approach can also stifle innovation and slow down AI adoption.

What’s the alternative?

It's crucial to make sure that subject matter experts with strong business domain knowledge are given data stewardship and custodianship. They should be able to set governance policies relevant to their business domain. While a design authority can provide guidance and advice, it's important to have the right level of authority to avoid unnecessary bureaucracy.

How to scale AI sustainably

Sustainable scaling of AI involves growing and developing AI capabilities while maintaining ecological integrity and social responsibility. This requires both sides of the coin: implementing AI for the purpose of improving sustainability, and ensuring any AI implemented is both ethical and sustainable.

Tracking carbon footprint: It’s important to consider the environmental impact of AI development to scale AI sustainably. This includes managing the energy consumption and carbon footprint of AI training and tuning processes. Fortunately, there are tools available help, such as carbon trackers to monitor and reduce the environmental impact.

Sustainable evolution: AI systems should be designed with long-term sustainability built-in, so that they can adapt to evolving business needs without excessive resource consumption.

Building an AI democracy: A sustainable approach can focus on empowering non-technical users to leverage AI insights through user-friendly interfaces and self-service tools. This democratises access to AI, reduces dependency on technical teams, and ensures that AI initiatives are aligned with business goals.

Building sustainable AI with data products

Data products offer a structured and scalable way to implement sustainable AI. Data products provide bespoke value that a business needs to achieve their goals. They are designed to deliver specific business value by packaging up data up with access rules, policies, security, and governance before applying AI. This means governance and sustainability can be directly packaged up directly within the data product, essentially baking it in from the outset.

This approach presents several benefits:

Accelerated Time to Value: Data products eliminate unnecessary activities by focusing on business objectives, ensuring efficient resource use. Focusing on valuable data also reduces unnecessary overheads.
Scalability: Once a data product has been created, it’s easily shared and scaled by simply adding in new data feeds. This has the added benefit of standardising information, making it easier to compare between departments.
Accessibility: User-friendly dashboards allow non-technical users to interact directly with data products, promoting broader adoption and impact. Users can also easily find and leverage valuable insights by providing catalogues of data products, promoting a self-service model for data-driven decision-making.

Why create a sustainable and scalable AI with Modular Data?

Data transparency

All our data products are open source, with all policies, rules, and users visible to ensure transparency and accountability around responsible data use.

We’re committed to a fair data democracy

Data can be an asset for everyone across an organisation by applying data product thinking. Our data products are low code, making them easy for non-technical team members to use. And our tools, templates, playbooks and approach empower your team to confidently navigate their data journey, developing their own insights, safely, securely, and at scale.

Access the leading experts in the field

We’re pioneers of the data product approach, and our team of experts have experience of delivering across Central Government - including HMRC, MOJ, and DWP - and private sector - including EDF and Hargreaves Lansdown.

Conclusion

Organisations can effectively integrate AI into their operations and scale it sustainably by reengineering processes, establishing robust governance frameworks, and adopting a data product approach. This enhances business efficiency and innovation and ensures that AI development and deployment are aligned with ecological and social responsibilities. Embracing this strategy will enable organisations to harness the full potential of AI and drive meaningful and sustainable transformation.

Author: Finbarr Murphy, CEO, Modular Data

About Modular Data

Modular Data make it quicker and easier to access, use, and share data to achieve immediate business value that’s easily scaled. With a data product approach, we close the gap between business users and their data and enable ethical and responsible innovation with emerging technologies, such as AI.

From improving interoperability, to migrating from legacy systems, to delivering insights that power service improvements – we are trusted by organisations such as the Ministry of Justice to drive meaningful change.

techUK and TechSkills at London Tech Week

techUK is proud to once again be a strategic partner of London Tech Week. The event is a fantastic showcase for UK tech, and we’re pleased that techUK members and colleagues are represented on the agenda.

We are pleased to be working with 20 techUK members at London Tech Week to highlight the positive impact of the tech industry in the UK.

You can find more details on the author of this blog, below:

Modular Data

The data product people

London Tech Week: Fringe events

techUK is leading several fringe events across the week on topics as varied as Justice and Emergency Services, Local Public Services, Digital ID, and international trade. Visit the pages below to register: