The National Gallery was an early adopter of AI tools. Back in 2018 their in-house data analyst developed an AI-based model to predict attendance to the Gallery’s temporary exhibitions, using XGBoost. The Gallery’s considerable database of attendance to past exhibitions was input to the model and, together with data on other factors influencing attendance, used to predict attendance to future exhibitions. The iterative nature of the process then ensured that actual admissions data could be added when the exhibition opened, allowing the model to learn from the latest data and update its predictions as the exhibition progressed. The model proved particularly successful at predicting attendance for exhibitions based on higher profile artists.

andemic on visitor numbers and profile, and variances in the speed with which individual audiences have recovered, has led to the model being put on hold in recent years, and shows the challenges of using machine learning from historic data at a time of cataclysmic change. Now that performance is starting to return to a more normal pattern, the Gallery is considering whether to update the model and start using it again.

Having resource is key to this, and the National Gallery has just recruited a new data analyst with machine learning programming skills to support this work. Going forward, generative AI is seen to be useful for activities such as measuring visitor sentiment, enabling thousands of visitor feedback comments to be quickly categorised, giving the Gallery invaluable understanding of how their visitors view the customer experience and helping identify trends and areas for improvement. Overall, AI is seen as a great opportunity to support, but not replace, human input to creating human experiences.

Mandai Wildlife Group is the steward of the Mandai Wildlife Reserve, a unique wildlife and nature destination in Singapore that is home to world-renowned wildlife parks including Singapore Zoo, Night Safari, Bird Paradise and River Wonders. As part of the Group’s digital transformation towards smart, integrated wildlife park experiences and operations at the Mandai Wildlife Reserve, a dedicated department - the Transformation Office - was set up three years ago. To realise this vision, they receive support from the Singapore government through grants which actively encourage organisations to innovate, test, and collaboratively share insights on the implementation of new technology.

I spoke to Edmas Neo, vice president of the Transformation Office, and his managers Khairulnizam Zulkifle and Moh Ai Wei. AI was initially incorporated into wildlife care such as monitoring habitat perimeters and assessing the body condition of animals in the Group’s care.

More recently, the Transformation Office has adopted generative AI to enhance the guest’s experience. Using Chat GPT via the Microsoft Azure package (which provides an added layer of protection for confidential customer and organisation data) enables them to sift through and categorise thousands of visitor comments in a matter of minutes, a task which previously took days and was open to variances in human interpretation. The categorised comments are then automatically routed to the respective departments for their necessary follow up and action. The increased speed allows them to provide prompt responses to visitors and provides a more nuanced interpretation of the comments, helping better inform future operations decisions.

A key responsibility of the Transformation Team is holding workshops and training sessions for the Group’s workforce to encourage the use of new technology and demonstrate how to get the best out of AI tools. Acknowledging the impracticality of banning the use of generative AI tools, the team has written a policy paper outlining guidelines on how to use such tools responsibly and teaches staff how to disable permissions to share data publicly.

For the future, AI is seen as an important tool to improve the operation of the visitor sites, enabling the Group to drive operational efficiencies.

Katherine Woollard

Like other larger organisations, the National Trust has a dedicated head of innovation and digital, Katherine Woollard. Woollard is excited about the potential for AI to support the National Trust, and is actively investigating ways in which this might happen. Possible areas of interest include adding AI tools to the National Trust’s intranet, testing the use of AI to create guest itineraries, and outcome-focused big data analysis to support operational processes. Some of this is likely to involve creating tailored AI solutions, subject to the potential scale of opportunity being seen as warranting the investment, due to the expense of implementing custom-made systems. Ensuring that this is done in a manner consistent with the National Trust’s values will be key and will require an ethical framework to be defined.

Woollard also sees challenges in using the free, open AI tools available; as a result, the National Trust recommends, and has issued staff with guidelines for the use of, the recently launched Microsoft CoPilot in Bing (previously Bing Chat Enterprise). They’re also exploring the use of CoPilot in Microsoft 365. This is the ‘AI assistant’ which Microsoft has been building into its products and which ensures that confidential data remains part of an organisation’s Microsoft ecosystem and does not become publicly visible. National Trust staff are asked to check with the digital team before using any other form of public AI tool.

rmountable and exceeded by the scale of potential benefits for the Trust.

Lesley Morisetti / photo: Andrea Horth

• Keep an open mind

• Invest in training staff – as with any other new technology, AI requires new skill sets

• Provide organisation-wide guidelines for the use of AI

• Focus initially on how existing AI tools can improve efficiency of tasks, supporting staff and enabling them to work faster

• ganisations with deeper R&D budgets to see how they use AI (Disney is an interesting one to follow on this and it’s likely that a number of your suppliers are already using AI in the services that they provide to you)

• Machine learning can support the delivery of excellent visitor experiences, but we still need people