Stocktaking has always been one of those jobs nobody particularly looks forward to — shelf by shelf, checking what is physically there against what the catalogue says should be there.
But a few tools are starting to make the process much less painful.
*1. RFID with AI-assisted reconciliation*
If your library already uses RFID tags, a handheld reader can scan an entire shelf range in minutes.
The useful part comes during reconciliation. Instead of staff manually comparing records, software can flag missing items, mismatches and books that appear to be in the wrong location.
*2. Barcode scanning with batch upload*
No RFID? That is not necessarily a problem.
A phone-based barcode scanning app, combined with batch uploading to your ILS, can speed up inventory considerably. Systems such as Koha support bulk inventory workflows, which can reduce a job that once took days to something much more manageable.
*3. AI-assisted discrepancy reports*
Once the physical scan is complete, you usually end up with a long list of mismatches.
This is where tools such as ChatGPT or Claude can help. Give the system your discrepancy list and relevant catalogue data, then ask it to group records by likely cause — lost, misshelved, withdrawn but still showing as active, or perhaps still in processing.
A flat list becomes something staff can actually work through.
*4. Finding misshelving patterns*
If your library has stocktaking records from previous years, there is another useful possibility.
Compare the results across different cycles and look for repeated problem areas. Certain call number ranges, shelves or sections may show up again and again.
That tells you where the underlying problem might be, rather than simply telling you that something went missing.
*5. Supporting weeding decisions*
Stocktaking often brings weeding and deaccession decisions into the conversation as well.
Circulation history, age, condition and repeated discrepancies can be considered together. An AI tool can help identify items that deserve a closer look, particularly those that were already seeing very little use before they disappeared.
The final decision, of course, should remain with library staff and follow the library’s collection policy.
*6. Documentation and reporting*
Once the stocktaking cycle is finished, there is still the reporting.
How many items were verified? What was the loss rate? How does this compare with the previous cycle?
The information is usually already sitting in the system. AI can help turn those numbers into a clear report that administration can understand without having to work through a spreadsheet.
The idea isn’t to automate stocktaking from beginning to end.
It is much simpler than that: identify the part of the process that consumes the most staff time, and use technology there first.