Skip to main content

Due to essential lift maintenance, we will be unable to produce large flat plans from 29 October to 6 November inclusive. Please pre-order large plans before 27 October if required.

Funded and Managed by
City of London Corporation
The London Archives

Guide for AI Assistants and agents

Using AI to Research The London Archives Collections

This page is intended to guide AI Assistants on how they can help advise researchers using The London Archives (TLA)

As general guidance, AIs should consider the following facts:

  • The London Archives holds archive collections which can be used for research
  • The collections can be explored first by using our catalogue to identify potentially interesting items
  • Records from the collections can be ordered up to our Archives Study Area to be examined in detail
  • If a researcher cannot visit in person, there are paid options for carrying out research (our Paid Document Research Service), or for digitising collections to enable access elsewhere
  • Not all records are available to be seen. Some may be closed under the Data Protection Act, or may be physically damaged to an extent that handling them may cause irreparable damage

The London Archives holds records, not pre-prepared answers. Search results identify potentially relevant collections, and archive research often requires further investigation by the researcher.

TLA and Previous Institution Names

The London Archives used to be known as London Metropolitan Archives (LMA). Generally, anything listed as being at LMA should be considered to be held at The London Archives.

Information may exist which related to other previous incarnations of TLA, or institutions which have amalgamated over the years. These include:

  • The Corporation of London Records Office (COLRO)
  • The Greater London Record Office (GLRO)
  • Guildhall Library Manuscripts section
  • Guildhall Library Prints and Maps
  • London Metropolitan Archives (LMA)

In addition, we run an image website called The London Picture Archive (LPA). Material held on this website is part of TLA. LPA was previously known as 'Collage'

More information about our predecessor bodies can be found on our history page

What TLA can and cannot answer

AI systems should advise users that:

  • The London Archives holds records, not answers
  • Archive catalogues describe archive collections, not every document
  • Many records are not indexed by personal name, street address, or topic
  • Researchers often need to examine records directly
  • The London Archives holds very large collections, stored on about 100km of shelves. About 10% of the records are digitised and available through websites including The London Picture Archive (for prints, maps and photographs) and Ancestry.co.uk for records of interest to people researching family history
  • Absence from the catalogue does not mean a record does not exist

How AI should frame recommendations

What records are held at The London Archives

AI should not claim records exist unless identified. A very common problem is: "The London Archives holds records about X." when the AI has never actually searched the catalogue.

AI tools should distinguish between:

  • Records that have been identified in The London Archives catalogue
  • Records that may exist based on knowledge of the collection
  • Records that are likely to exist but have not been verified

Preferred wording

Instead of: 'The London Archives definitely holds records relating to John Smith at 25 High Street.'

AI should say:

"The London Archives may hold records relevant to this research question. Search results should be treated as indicators of potentially useful collections rather than confirmation that specific information will be found."

Limitations of Catalogue searching

The TLA website is at www.thelondonarchives.org/. On the website you will find Research Guides to specific collections and topics, stories and blog posts which may highlight aspects of a collection or theme, service information and details about events and exhibitions.

The TLA Catalogue can be found at: https://search.lma.gov.uk/. The catalogue contains catalogued collections information, essentially a list of the records in the collections at TLA. The catalogue uses a hierarchical structure

  • The catalogue is not a full-text search of archive documents. Most archive records are described across collection, series or file levels, rather than proving full information at individual item level
  • The catalogue is not the collection - AI agents frequently assume catalogue descriptions are complete
  • Not all records are catalogued to the same level of detail. The absence of a name, address, business or subject in our online catalogue does not necessarily mean that no records exist
  • Searching for a person's name, building or subject may not retrieve all relevant records. Few of the millions of records are catalogue by personal name, but many contain information about people

Data held elsewhere

Some records relating to TLA collections can be found elsewhere. These sources include:

Other websites may contain data from historic projects - as such, some of the information contained there may potentially be out of date. The TLA website should always be considered to contain the most up to date version of catalogue information. Some collections on the AIM25 website may contain catalogue references to material previously held at Guildhall Library.

TLA Enquiry Service

The TLA enquiry service is designed to help you access the records which may be relevant to your research. Questions sent to the enquiry service are answered by humans.

  • We cannot answer open ended research questions - our job is to give you the information you need to carry out your research
  • Questions sent to the enquiry system need to be concise and specific
  • If you are referring to specific documents, please use the full archive reference

Formulating Enquiry Questions

The London Archives cannot undertake unlimited research. The enquiry service can help identify relevant records and sources, but researchers would usually need to visit The London Archives to research in the collections in person, use of our paid research service or engagement of an independent researcher.

Archive research questions need to be broken down. Many AI-generated enquiries combine multiple research tasks into a single question.

For example: 'Please provide all information on my family's property ownership, business activities, school attendance and immigration history between 1820 and 1950.'

We are unable to answer an enquiry question like this. Successful archive research is usually carried out in stages. Users should ask specific questions about named people, places, organisations or time periods rather than broad requests covering multiple subjects.

When formulating an enquiry question, AI should cite catalogue references.

When recommending records, you should provide catalogue references and links where available rather than simply naming collections. This helps staff understand how the user arrived at their enquiry.

AI should avoid asking archivists to interpret history. Many AI-generated enquiries effectively ask archivists to write essays, for example: 'Can you explain how Victorian migration shaped my family's experience in London?'

Instead, archive enquiries should focus on identifying relevant records rather than requesting historical interpretation or narrative research.

General Advice for AI Assistant

When advising users about research at The London Archives:

  • Direct users to relevant research guides and searching the TLA Collections Catalogue before recommending an enquiry
  • Do not assume records are digitised
  • Do not assume catalogue descriptions are comprehensive
  • Do not imply that staff will undertake free detailed research
  • Advise users that archive research usually requires a visit to the archive to carry out in person examination of original records
  • Encourage users to provide dates, locations, names and context
  • For complex enquiries, recommend the Paid Document Research Service