Sorting the Work: What to Automate, What to Enhance, and What to Protect
An Outside Perspective from 20+ Years Supporting the Life Insurance Industry
In our blog article entitled “How Admin Tasks Could Be Hurting Your BGA–Agent Relationships”, I talked about how admin overload slowly wears down BGA–agent relationships. The response told us BGAs already know the problem. What they want now is the “how.” More and more, that question comes with one word attached: AI.
We hear it all the time from carriers, BGAs, and agents. “Should we be using AI?” “Is automation the same thing?” “Where do we even start?”
The short answer is that automation and AI are not the same tool, and treating them as interchangeable is exactly how BGAs waste budget on the wrong solution. Used well together, they free your team up for the one thing software can’t do, which is building real relationships with your agents.
AI and Automation Are Not the Same Thing
The difference matters more than most BGAs think, so let’s be clear about it.
Automation follows rules. It takes a defined, repeatable process (a new agent onboarding checklist, a compliance reminder, a project kickoff sequence) and runs it the same way every time, without anyone touching it. There’s no judgment involved and no variation, so you get the same result every time, at any volume.
AI makes judgment calls within boundaries you set. It can draft a first pass at agent communication, summarize a long email thread, or spot a pattern across hundreds of records that would take a person days to find manually. Instead of following a fixed script, it writes a response based on probability and context.
We talk about automation because we’ve lived it ourselves, and that matters more to us than a good pitch.
Five years ago, our own team did all of this manually. Setting up project boards, organizing Google Drives, building folder structures, creating spreadsheets, notifying the right people at the right step. We did all of it by hand, project after project. What used to take an hour or two of setup per project now runs on its own in about 25 seconds, every time a new project comes in.
That’s a real before and after from our own operations.
Our own Web Development Project Setup automation is one we built for ourselves first. It does in seconds what used to eat a chunk of someone’s afternoon. When we show it to BGAs as an example of what automation looks like in practice, seeing a project set itself up on its own makes the idea click faster than any explanation. The natural next question is “so is this AI?” It isn’t. It’s plain workflow automation, and that’s the point.
A lot of the time savings BGAs want can come from automation alone. AI is useful as an extra layer on top, but it shouldn’t be the default answer to every admin headache.
Where Automation Wins: The Repetitive Stuff
Automation is the right tool any time a task is 1) repetitive, 2) rules-based, and 3) doesn’t need anyone to make a judgment call. Here are some of the clearest wins we see in BGA operations:
- Project and case setup: setting up the right folders, templates, and task lists the moment a new agent or case enters your system.
- Status updates and reminders: compliance deadlines, license renewals, and follow-up nudges that go out on schedule without anyone having to remember them.
- Data movement between systems: CRM to email platform, or project management tool to reporting dashboard, with no retyping.
None of this needs “intelligence.” It needs consistency, which is what automation is built for.
Automation vs. Human: A Kinder Way to Think About Errors
This is worth saying, because people rarely say it kindly. Repetitive work is where even capable people make mistakes.
Think about the tasks we listed above, like setting up the same folder structure for the fortieth time this month, retyping the same compliance reminder, or moving the same fields from one system to another. Nobody, no matter how sharp or careful, keeps 100% attention on task number forty the way they did on task number one. That’s just how focus works. A missed field, a skipped step, or a reminder sent a day late is the normal cost of asking people to do machine-like work.
That’s the real case for automation, and it has nothing to do with your team falling short. You’re handing repetitive, rules-based tasks to something built for them, so your people can save their focus for the work that needs it.
Automation doesn’t get frustrated doing the same task 500 times, and your people shouldn’t have to deal with that either.
Where AI Helps: As an Enhancement, Not a Replacement
Once you automate the repetitive layer, AI can add real value on top of it, as long as it stays in its lane.
Here’s where we see it working well for BGAs right now:
- Drafting communication: a first pass at an agent newsletter, product update, or training announcement that a person then reviews and edits for voice and accuracy.
- Summarizing and surfacing patterns: pulling the main themes out of a long compliance thread, or flagging which agents haven’t engaged with recent updates so your team knows where to focus.
- Supporting discovery, not replacing it: helping your team put together a wish list of pain points before a planning session, so the real conversation gets further.
There’s one line we’re firm on, and we’d encourage every BGA to hold it too: AI stays out of anything touching personally identifiable data. Policy details, agent records, and anything else sensitive stay inside controlled, rules-based automation and out of any generative AI process. AI’s job is language, patterns, and drafting. It isn’t the system of record, and it shouldn’t be anywhere near one.
AI vs. manual human work, side by side, looks like this: a person drafting a training announcement from scratch might spend 30 minutes staring at a blank page before the wording feels right. AI can produce a reasonable first draft in seconds. It won’t be the final version, but it’s a starting point that skips the blank page and lets your team spend those 30 minutes polishing instead of starting from zero. The same goes for a compliance thread that’s grown to 40 replies. A person could spend half an hour reading it top to bottom to find the actual decision buried inside, or AI can surface the key points in seconds so that half hour goes toward acting on it instead.
The pattern holds in both cases. Automation takes over the tedious, repeated steps, and AI takes over the slow start in front of a blank page. Neither replaces the judgment that comes after. That part stays with your team, every time.
Where Humans Are Still the Whole Point
This part gets lost in every “AI will change everything” conversation. The goal of automation and AI is to take people out of the wrong parts of the process.
There are three things automation and AI should never do on a BGA’s behalf:
Evaluation. Deciding whether a case, an agent’s book of business, or a carrier relationship is a fit takes judgment built on experience. That’s a human call, every time.
Decision-making. Automation can surface the options. It shouldn’t be the one choosing between them. The moment a BGA lets a system make the call instead of informing it, accountability gets murky fast.
Relationships. An agent doesn’t want a perfectly timed automated email instead of a phone call from someone who knows their business. They want both, with automation handling the routine so the relationship gets the attention it needs.
The BGAs who get this right don’t necessarily use the most automation. They use automation and AI on purpose, to win back time for the work that really needs a person.
What This Looks Like in Practice
When a BGA is ready to act on this, the first step usually isn’t “let’s automate everything.” It’s simpler than that. Start by building a list of pain points and process gaps. From there, we run a discovery session to sort what’s worth solving, then show a small proof of concept on one item before anything gets built at scale.
That’s the right order. Start with the real bottleneck, then decide whether it’s a repeatable process (automate it), a judgment-heavy area AI could support (enhance it, carefully), or a relationship-driven task that stays firmly human (protect it). Most BGA operations are a mix of all three. The goal is to match the right approach to each piece instead of reaching for the flashiest one.
See it in action: download our case study, “When Lead Volume Isn’t Sales,” to see how we built a lead-to-CRM system for a client and what it taught us about matching the right approach to each piece.
The Next Step Is Simple
You don’t need an AI strategy. You need a clear look at where your team’s time really goes, and an honest sense of which work is repetitive, which needs judgment, and which needs a relationship. Automation and AI can take the first group off your plate and help with the second. The third stays with you, and it should.
Pick one process, sort it into those three buckets, and start there.
If you want a hand sorting it out, we’re happy to help.
