AI is Already in the Hall. It's Time to Claim It.

A communications director at a mid-size local has been using an AI writing tool for about three months. She uses it to draft member notices, generate first drafts for a meeting agenda, and summarize long documents before she reads them. It saves her close to two hours a week. She's sheepish about it though. It feels like cheating in some way. Would others think she is lazy or trying to cut corners? Her local has not adopted AI, whatever that really means.

That is a consistent picture of where most union organizations stand right now: individuals quietly using tools that make their own work faster, while the union as a whole has not yet figured out what it means collectively.

The conversation worth having is not "should we adopt AI." It is a much simpler one: how do we put this new technology to work for us in a way that can free up real time to focus on the things that automation or technology can't solve?

AI is not a product. It is a tool, not a budget line.

The vendor pitches have made AI sound like a feature you add to your stack. A platform upgrade. Something you purchase and roll out. That framing turns the question into a procurement decision when it should be an operational one.

Your local has a set of things it needs to do every week: communicate with members, manage training records, process new member intake, follow up on dues, prepare for meetings. Much of that work is repetitive, rules-based, and time-consuming. AI is useful not because it is innovative but because it can absorb a large share of that repetitive load, so that the people doing the work have more time for the parts only they can do.

Where most locals actually are

Individual staff members and officers have figured out on their own that tools like ChatGPT help them work faster. They use them personally, informally, often without a policy covering it. A steward drafting a grievance summary. A training director generating a first pass at a renewal notice. A business agent using AI to summarize a long contract section before a meeting.

What most locals have not done is ask what changes when that capability moves from one person's workflow to the organization's operations. When the time savings are shared. When the benefit compounds instead of staying with whoever figured it out first.

That gap, between individual informal use and benefit to the whole local, is where the real opportunity is for unions right now.

What "more with less" looks like in practice

Most local teams we work with lose an hour a day per person to work that is real but should not require them. Member communications drafted from scratch every time. Training records compiled manually from multiple sources. New member intake handled by whoever picks up the phone. Meeting documentation that takes longer to write than the meeting took to run.

AI does not replace any of that work. It compresses the time it takes to do a first pass, so the person doing it can spend their time on the judgment calls and the edits and the context that actually requires them. A staff member who used to spend ninety minutes drafting a member communications sequence can generate a first draft in ten minutes and spend the rest of the time making it right. That is not a small change. Across a week, it is half a day back.

Three categories worth starting with

Not every task is a good candidate. Many require a human touch. The rule is the same one that applies to any administrative decision: if the work is repetitive and rules-based, it is worth examining. If it requires reading a room, building trust, or knowing the member on the other end, it belongs to a person.

Three categories consistently return the most time:

1. Member communications. Drafts, follow-up sequences, segmented notices for different member roles. AI writing tools do not decide what to say or when to say it. They eliminate the blank page and the slow first draft, and they do it in a fraction of the time.

2. Training and credential records. Tracking completions, surfacing expirations, generating status summaries across a cohort. The data already exists in most locals. The question is whether a person is compiling it manually every week or whether a process is doing it automatically.

3. New member intake and onboarding. Intake forms, welcome communications, routing to the right steward or program. Getting this right in the first 90 days matters more than most locals treat it. Good process design makes sure the right handoffs happen without someone holding all the steps in their head.

What changes when it becomes organizational

The business case for one person using AI is obvious once they try it. Time saved is time saved.

The business case for building it into how the organization operates is different, and it is more important. When a local builds these tools into how it actually operates, the benefit does not stay with the one staff member who figured it out. It compounds across everyone doing that kind of work. It persists when that person leaves. It becomes something the local can improve over time rather than something that disappears when a good employee moves on.

That is the conversation worth having internally, not just checking the AI box.

The work that still belongs to people

Nothing in this changes what union leadership actually is.

The grievance follow-up belongs to a person. The member who needs to understand what the contract allows belongs to a person. The steward building trust at a new job site belongs to a person. The local president holding a difficult membership meeting together belongs to a person.

What changes is how much of the day gets used up before those conversations happen. A staff member running on administrative debt does not bring her best to the member who needs her. A business agent who spent the morning compiling records brings a different quality of attention to the afternoon grievance than one whose morning was clear.

Start with one question, not a policy

The locals that will move well on this are not the ones that start with an AI policy. They are the ones that start with an honest accounting of where their people's time is going.

Spend an hour with your staff mapping the top ten administrative tasks by hours per week. Separate the rules-based from the judgment-based. Then pick one task from the rules-based column and find out whether a tool already exists to help with it. Most of the time, one does, and most of the time it takes less than a week to try.

AI is genuinely capable of more than most locals are currently asking of it. The point is not to chase the capability. The point is to put it to work on the right problems: the ones costing your people the most time, the ones getting in the way of the work that matters. Start there, and real, lasting impact follows.