My Coworkers Don't Want AI. They Want Macros
My coworkers don’t want AI. They want macros.
Let me back up a little. I spent April gathering and May refining and organizing requirements for a system to replace our current ILS. This meant asking a lot of people about how they use our current system, taking notes, and turning those notes into requirements. 372 requirements.1
Going into this, I knew that some coworkers used macros to streamline tasks. I came out of it with a deeper appreciation of the different ways they’ve done so.
It made me think about the various ways vendors are pitching “AI” for their systems and the disconnect between these pitches and the needs people expressed. Because library workers do want more from these systems. We just want something a bit different.
A Few Ways We’re Using Macros
Some environmental factors: We use Sirsi Symphony and the Workflows desktop client in Windows with MacroExpress.
Canned Notes
One of the simplest things we do with macros is add canned notes. These are most often action notes, which have to be added to existing records. For example, this is a standard deacidification note:
Deacidified.|c20030107.|iliquid.|xBookkeeper.|jCranberry Township, PA 16066.|kPreservation Technologies Inc.
The text string is the same every time except for the subfield c, which grabs today’s date, e.g. 2003-01-07 in the example above.
Spine Label Marking
Let’s look at something more complicated. Our marking team has a set of tasks they need to perform when setting an item’s permanent location and preparing a label. They use a macro that does the following:
- If it’s in the respective library’s main stacks: update Home Location field to the location which corresponds with the Library field (e.g. if it’s at Altoona, it would be STACKS-AA).
- Copy the call number to the clipboard, with some line breaks.
- If the Home Location has a special marking prefix: prepend the marking prefix, again with line breaks. This data isn’t in Symphony at all, it comes from a separate marking table.
They then can paste the result into the label software, ensure the line breaks are correct, and print a label.
This saves them:
- Clicking in the Home Location field and typing or scrolling to find the correct Home Location (when it’s in the main stacks, otherwise they still have to do this part).
- Clicking in the call number field and copying the data.
- Opening our locally-maintained marking documentation, checking whether the Home Location is in the prefix table, and copy-pasting that separately into the labeling software.
The marking table used by MacroExpress requires regular maintenance, but it’s done in parallel with updates to our department’s documentation.
One macro.
Annex Location Management
We also use Symphony to manage our Annex data. There are pros and cons to this, but I’ll be focusing on how macros makes the work much easier.
When an item moves to an annex, our team needs to update the following information:
- The item’s Library and Home Location fields
- The item’s very specific position within that annex
The second piece of information is encoded in a new barcode. We want to record this data in our Extended Item Information fields (locally mapped) which we use for everything from retrieval to reporting. We need to be able to identify everything in a row, a shelving unit, box, even position in box. So we need it broken apart in a way that can be easily queried or used to generate pull slips. We also want to record information about the size of the box (A3, A4, etc.).
To use the macro, our Annex workers click in the barcode field, press a key on their numpad, and scan in the new barcode. This performs the following steps:
- Update barcode with the new barcode
- Update the Library field to UP-ANNEX
- Update Home Location to the specific annex at which the work is being performed
- Fill in locally-mapped Box Size field using a lookup table and the value entered from the numpad.
- Fill in a set of locally-mapped Extended Item Information fields with the values of Row, Shelf, Box, and Position within the box based on the value of the barcode.
They use a special numpad which is just an ordinary external numpad that’s had labels applied to its keys. These indicate which key maps to A3, which to A4, etc. Users don’t have to memorize “Press 7 for A3,” they just have to press the A3 key.
Once again, the macro eliminates the human error factor that would come with manually typing or even cutting and pasting in any of this data. It transforms a task that would take a good minute of clicking, tabbing, typing/copy-pasting into one that takes maybe 10 seconds.
One macro.
Why Macros, Not “AI”
To summarize at a more abstract level, we are using macros to:
- Enter consistent data
- Avoid typos
- Improve system design2
- Add conditional logic
- Save the user repetitive, mindless work
At its best (pattern-recognition), “AI” is overengineered for what we need: logic and lookups.
At its worst (predictive text), it’s the opposite of the very concrete and repeated things we want to be able to do.
It doesn’t have to be macros. My coworkers would welcome a system which let them create custom workscreens with only the few fields they need for a regular task and let them add their own conditional logic. 3 Or they could collect barcodes and run a batch process that inserts those deacidification notes with canned text and a variable date field vs. running the macro on each record.
But the direction is one of control, of refinement, of engaging with the system on one’s own terms. That is in contrast to the black box we’re being presented by both “AI” companies and library system vendors.4
As for me? I’m excited about APIs.
-
Some of those requirements are: “Staff users can check in and check out library materials.” On the one hand, yes of course. On the other hand, a friend’s university just implemented a system in which you can only hire people on the first day of a pay period. No, you can’t hire them on the second day and make it retroactive. Sometimes you have to ask about really, really basic stuff. ↩︎
-
I’m thinking here in particular of the ways that our use of macros significantly reduces the need to tab between fields or scroll through controlled values in Workflows. ↩︎
-
I have seen some good ones developed at individual institutions using APIs. ↩︎
-
As I’ve said elsewhere, I think OCR and machine-learning capabilities which can support catalog record creation could be helpful! I’ve also seen great uses of these tools to transform a scanned page into a tab-separated file or reformat existing data quickly. The problem is that we’re being offered tools that also generate non-existent page counts we have to clean up or ones that propose taking over course design for faculty (ELUNA 2025 plenaries). ↩︎