CILIP member, Josiah Richardson, is a Senior Library Assistant at An NHS trust in the Knowledge and Library Service based on the hospital site, whilst also doing the Level 3 Library, Information and Archive Services Assistant apprenticeship, for which CILIP Pathways is the end point assessment organisation. In this case study, Josiah discusses how AI has simplified and sped up reporting and increased his knowledge of Excel

This story is featured on the AI hub

• AI is a useful collaborative tool which may create extra capacity and encourage review and modification of existing job roles.
• AI may be used to increase knowledge of Microsoft systems, and others, used across the organisation, which may increase productivity.
• Some uses of AI are universally applicable across the whole organisation in a variety of roles.
• AI acts as a virtual assistant and does not replace roles but augments existing knowledge and expertise.
• Limitations include a limited ability to provide context or discern the implications of conceptual links.
• Risks should be addressed by enhancing internal governance structures and procedures to ensure confidential organisational information is not shared.

Upskilling Excel Skills with AI Assistance

I was tasked with a couple Excel projects that required skills he (nor anyone else in the library team) possessed. Like many colleagues I can confidently use Excel basics but have never received formal training on advanced formulas and coding. I asked colleagues for help, but no experts were available. I tried online tutorials but struggled to apply them to his specific data and tasks.

With tight deadlines and limited budget for formal training, I opted to use AI as a virtual assistant to expand my Excel skills. I had a workbook of multiple data sheets and needed to build formulas for counting, percentages, and filtering between the sheets.

The AI application guided me through functions such as COUNTIF, COUNTIFS, and referencing cells across multiple sheets. I described the worksheet layouts and terminology without sharing sensitive data. When initial formulas failed, the AI application provided troubleshooting tips on checking parts step-by-step, verifying date formats, and fixing typos.

In a similar project, I had raw data on book and journal usage that required analysis. The Excel formatting was disorganised made it difficult to extrapolate the required data. I asked the AI, through a series of informal prompts, to help reformat it into a structured table and summarise statistics by subject area. Through these collaborative efforts, I transformed the data into actionable insights.

Interacting with the AI boosted my Excel skills. I am now more confidence with complex functions, cross-sheet references, pivot tables, and percentages. The AI supplemented my knowledge and taught effective problem-solving. The conversational approach provided personalised guidance to strengthen his data analysis abilities.

With the AI's assistance, I achieved higher quality data insights far more efficiently, saving the organisation time and money. While an experienced professional remains essential for full context, I found the AI most valuable when combined with my specialised expertise. As a collaborative tool, not a replacement, the AI reached its highest potential.

Website Editing and Search Engine Optimisation (SEO)

I have used AI to edit and optimise content for the library website. An AI assistant can analyse library website content and suggest adding keywords in natural language to help search engines understand what each page is about. The AI may also recommend rewriting or restructuring content to focus on what users are searching for. This helps search engines display the most relevant pages for each search query both within the site and on Google and other search engines.

AI can help with SEO for a library website by generating optimised meta descriptions. The meta description is the short summary of a web page that shows up in search results. AI tools like Claude can take website content as input and automatically generate meta descriptions that include relevant keywords and are the ideal length to get clicks. This makes it easier to have unique, appealing meta descriptions across all pages rather than writing them manually, saving large amounts of time. This learning is universally applicable across organisations.

Literature Searching and Synthesising

The extra capacity generated by using AI tools has enabled a review of Senior Library Assistant roles and an expansion into conducting basic literature searches. I have utilised AI literature search tools for basic keyword searches. I found that across large databases, AI can compile results much faster than I could manually. This allowed me to conduct wider initial searches to get an overview of the available literature on a given clinical or non-clinical topic.

It has been useful to ask the AI to generate search terms for a query, as this expands on existing knowledge of the topic. One other promising capability of the AI tool is uncovering connections between concepts across literature. This has the potential to find papers that likely would have been missed in manual searching, and this is an aspect which requires further exploration.

Limitations have been identified, namely the AI has limited ability to provide context or discern the implications of conceptual links. Also, the AI currently only has access to public databases, not proprietary NHS resources. This means the AI cannot fully replace knowledge of NHS-specific databases and resources.

AI text summarisation also provides helpful high-level insights from articles, supporting any reviews made of the search results.

While the AI lacks the nuance to fully evaluate relevance or scientific rigor as a trained library professional can, it may provide a starting framework and offer insights that may otherwise be missed. Full text results still require manual review and the use of expertise to assess quality.

While the AI augments certain search tasks, it remains simply a virtual assistant. It does not replace the need for an experienced information specialist to conduct literature reviews.

Writing and Reporting

AI may be significantly useful for writing reports such as this, which is a use that may be universally applied across a range of roles in the organisation, saving time and freeing up capacity. For this study, AI was used to provide an initial framework and to help summarise conversations with it about Excel formulas. Substantial editing the draft was undertaken to ensure accuracy, relevance, and appropriate tone.

AI is also useful for correcting grammar, assessing readability, and adjusting the material for differing audiences which may assist colleagues with learning difficulties. However, specialised knowledge is still required to oversee the AI and refine the final output.

Translation of written material for research purposes

Working as part of an international community, with colleagues from over 80 countries, AI has been used to identify the language of a requested journal article and then produce an English summary to confirm the correct paper has been sourced. This saves time compared to manual translation.

Image and Video Creation for a website

AI has been used to generate or edit images for the library website and promotional materials. It allows customisation of graphics with photographs produced by the organisation without needing to source royalty-free images. Further investigation by the Library team is ongoing to use AI for basic instructional video editing and creation.

Idea Generation and Innovation

Asking an AI to provide ideas on a subject has been a useful brainstorming input. I have used AI to generate ideas for team building activities, promotions, marketing slogans, and finding new information sources on a topic.

Conclusion

AI has been utilised as a collaborative tool to enhance productivity on various tasks. However, it does not replace the contextual and specialised knowledge of an experienced information professional. The AI is most effective when combined with personal expertise, not as an independent solution. With this integrated approach, AI reaches its highest potential as an assistant rather than a replacement.