Out of all the emerging technologies out there, why are businesses increasingly turning their sights to AI?
Dilpreet Sall: We know that organisations need to invest in AI because so much more data is being generated than ever before. Whether that’s in a structured format or an unstructured format, we know that there’s a lot of it. And businesses struggle to extract the value from all that data.
The businesses that figure out how to leverage artificial intelligence to capture data and process it in a meaningful way first are going to be able to address real business problems efficiently. They are the ones that are going to be the survivors of tomorrow. Most of the clients that we work with have already figured this out.
IT and business leaders often face pressure to ensure initial AI pilot projects succeed to secure further funding. Once an AI pilot project finally launches and the technical feasibility of the technology is proven, many hidden challenges can still emerge such as the integration complexities, high maintenance costs and the need for business process reengineering. How does Nexus FrontierTech’s AI Factory tackle these issues?
Dilpreet Sall: In its simplest form, the factory is a framework to identify, prioritise, and execute pilot projects in a repeatable way continuous business impact is being shown. The real focus of AI Factory is all around just starting. We help organisations identify specific opportunities that exist and take advantage of artificial intelligence, all within a short timescale and with minimal investment. The approach is quite a simple one. The benefit is that here at Nexus we already have the expertise and the people skills needed to carry out something like this.
Nexus’s AI Factory is a production programme inspired by a factory where products are made at speed and scale. Can you elaborate more on how AI Factory works?
Dilpreet Sall: The easiest way to think of the AI Factory is as a funnel: what goes in at the top of the funnel is a list of unpalatable ideas, which have the potential to become a use case. And what comes out is an implementation proposal for a full AI solution.
The first stage of the funnel is all about use case gathering. This is where our consultants conduct a 360 scan on the business and build a holistic view of what the business needs. The result is a prioritised use case list which can get passed on to stage two.
Stage two is all about the process of choosing the right use cases. We try to build a picture of what is the need for the use case and understand if the use case aligns to the overall strategy of the company. We also want to verify that the use case is likely to have a compelling ROI attached to it.
For this reason, we have built an ROI calculator that attempts to put a number on both tangible and intangible benefits that a use case may have. Only the use cases which have a compelling ROI behind them and a compelling need and a compelling alignment make it through to the next stage, which is the technical feasibility assessment.
In this stage, a technical expert gets involved and begins to analyse each use case from a tech perspective and sets out to ideate a solution. They also look out for any implementation complexity, data collection challenges and integration challenges. That concludes the consulting part of the funnel. The result is a report with the recommendations from all the information that is gathered which is presented in a way that shows each use case articulated as watertight as possible. We present the initial business case and also present the feasibility assessment with the predicted technical challenges.
In other words, all of this information helps the client to build a better picture of each use case and make a decision on which use case should go forward for pilot development. And the last part of the framework is about developing the implementation proposal. What is the exact goal of this stage?
Dilpreet Sall: Each proposal that comes out of the factory has already been through rigorous testing of analysis, a feasibility assessment, and has been tested with real client data. And of course, this is a repeatable model. You can have many use cases going through the funnel, depending on the size of the factory that is constructed.
- Does the AI Factory framework include any training or roadmap for future development?
Dilpreet Sall: Yes, the framework also covers the need for managing and executing these projects. The way we do this is by setting up a governance committee within the client’s organisation. The governance committee can oversee all factory progress and has decision-making power and is given updates by the Nexus team and the client that have been allocated to each use case.
Nexus entered a partnership with a UK government agency that needed to scale their onboarding and operational processes in a short time frame. Before COVID-19, the agency ran a bulk of their operations manually. AI and automation technologies were identified as tools to achieve a digital transformation and enable business continuity. Can you give more insights about how AI Factory was implemented in this case?
Dilpreet Sall: The issue was that the agency had so many ideas, but didn’t have a methodology to prioritise any of those ideas and turn them into well-articulated use cases. They had a lot of trouble defining the business value or the business case that they could attach to it. On top of that, they also had a tight timeframe that the government had set out for them.
These conditions made them a perfect candidate for the AI Factory. The AI factory provided the framework that could be used to take hundreds of ideas, get them down into a pilot for testing and build out the proposal for implementation. And this is exactly what the client did.
The agency was able to use Nexus’s tools and frameworks to prioritise more than 100 use cases. They used a holistic approach, which meant that every single use case aligned to the data strategy of the organisation. At the end of the programme, they were able to turn around four proposals within two months, and each proposal had a comprehensive analysis, technical feasibility assessment, and had a pilot with real client data. Using those results to back up each one of the proposals, the client was able to present their plan back to the government on how they were planning to scale their processes. And we also provided them with a repeatable framework for further use cases that they wanted to test.