AtkinsRéalis AI Pilot | Telefónica Tech UK&I

AtkinsRéalis unlocks AI’s revenue protection potential

Artificial intelligence enables project margin interventions

Case Studies | 3 July 2024

About the customer

Engineering a better future

Global engineering business AtkinsRéalis has proven that AI can identify project risks and protect the profit margins of the business in a partnership with Telefónica Tech. The artificial intelligence (AI) Pilot gives business leaders at AtkinsRéalis the ability to identify and intervene in major projects before the profit margin of the project goes into decline.

AtkinsRéalis is one of the world’s largest and most significant engineering consulting organisations, employing 37,000 staff in 50 countries. It is the organisation that delivered the famous Burj Al Arab sail shaped building in Dubai, and is currently developing Net Zero structures, public transport upgrades and nuclear power stations around the world. “We’re engineering a Better Future for our Planet and its People,” says Ross Offer, AtkinsRéalis VP of Data & AI.

The Challenge

Uncovering Potential Profit

The operational finance teams at AtkinsRéalis required data-driven insights to anticipate potential declines in project profit margins. With these insights, AtkinsRéalis would be able to intervene on a project before a decrease in margin became a risk to the business. Offer says: “We wanted to use data and statistical analysis to show where we had challenges, because as an organisation, we typically have small margins on the projects, so any erosion can have a massive impact.”

Artificial Intelligence was seen as a tool for uncovering potential profit margin risks during the lifecycle of a project, which were previously undetected by the operational teams. This insight is advantageous for both project management and overall business performance.

The Solution

Proof of Concept through to Deployment

Having realised the potential of AI to identify interventions, Offer and his team approached their data technology supplier Microsoft, who recommended Telefónica Tech, a certified Data & AI Microsoft Solution Partner to jointly develop a proof of concept (POC) for AtkinsRéalis.

An important consideration was that the AI didn’t inherit any bias. If bias existed, the AI would likely give AtkinsRéalis information the business already knew. Offer says: “We didn’t want a model that looked for specific outcomes, because we can identify those outcomes already with Excel spreadsheets and that was already being done.”

Offer tasked Telefónica Tech to develop an AI that would: “tells us what we don’t know”.

To train the AI model, the team at AtkinsRéalis provided Telefónica Tech with data samples from a period of over eight years. These samples included data from more than 1500 different projects that were known to have experienced a decrease in margin.

Additionally, a two-day workshop session was held with the Telefónica Tech team to ensure a clear understanding of the business context, rather than simply relying on facts and figures. Offer believes that “context is key”.

During a span of three months, the team created a collection of AI models that could detect project risks, forecasting the ultimate profit margin of a project, and identifying the causes of this margin decline. This was accomplished through the utilisation of machine learning models that examined all aspects of a project, such as project category, client, unbilled expenses, and cost escalations. The team utilised Microsoft Azure Machine Learning in a Software-as-a-Service (SaaS) instance for AtkinsRéalis.

The Outcomes

Scaling AI to boost skills and abilities

“This POC has shone a light on the challenges that can occur,” Offer says of the proof of concept. Analysing the results of the project, AtkinsRéalis believe they will achieve significant savings in lost revenue from being able to intervene and protect the margin of projects, in real time.

AI has proven its case for AtkinsRéalis, Offer says: “This shows the possibilities are endless if you expand this technology across the whole organisation.” AtkinsRéalis now has the opportunity to scale AI to boost the skills and abilities of its front-line project managers, operations directors and finance teams with data driven insight. This will reduce the level of financial risk in the organisation and ensure that projects achieve their target profit margins – something the CEO has highlighted as vital to the business strategy.

The rising public awareness of AI following the release of ChatGPT has increased interest in the potential of AI across AtkinsRéalis, from the CEO to project managers, Offer says. AtkinsRéalis is now considering how to adopt the Pilot into production across the business.

Telefónica Tech’s collaboration with AtkinsRéalis also provided unforeseen insights for the customer by determining a set of new indicators that were not previously known. Looking ahead, Offer sees this AI concept as a technology that will not replace the skills or roles of AtkinsRéalis staff, but boost their abilities and improve their working lives.

“Their flexibility is commendable, especially when project scopes evolve, showcasing their commitment to delivering for their clients. A collaborative approach has made them an invaluable partner in our deliveries.”

VP of Data & AI, AtkinsRéalis

Ross Offer

"Working with Telefónica Tech has been an incredibly positive experience. Their team's expertise shines through in every engagement, demonstrating a deep understanding of our needs and the technology in question. They listen attentively, ensuring that every requirement is truly understood."