Everyone right now is talking about what AI can do. At Mailbox Merchants, we’ve been more interested in finding out what it can actually do for us.
We rolled out AI tools across our team in early 2026, putting them directly in the hands of people in sales, client services, and administrative roles. Since then, our team has been experimenting with ways to use AI to solve real problems in their day-to-day work.
Some of those uses are pretty simple. Others have grown into tools and processes that can help us make smarter decisions for our business and our clients.
We wanted to share some of the coolest things we’ve come up with so far. We hope this is valuable information for our customers to see how we’re using AI to improve our processes, as well as potential inspiration for how your organization might use AI.
Sometimes the best use of AI isn't particularly flashy. It just takes something that used to be a pain and makes it much easier.
Our Client Services team found a great example in the carrier route and market count data we use for planning. Some of that information comes to us in PDFs, which makes getting the data into a format we can actually map and analyze a manual process.
Now, a team member can use AI to convert those PDFs into mappable spreadsheets. A task that could previously take a chunk of an afternoon can be completed in minutes.
As one Client Services team member put it:
“Some of this was almost impossible to do without it.”
The same idea has spread to recurring reports. Our Administrative Services Manager has used AI to rebuild several reports that previously had to be created manually every week or month. Rather than repeating the same process at the next deadline, she's looking for opportunities to build it once and make the work easier going forward.
AI has given us new ways to work with information we already have. Our sales reps regularly receive reports about the performance of their markets. The information was there, but what each rep did with it could vary considerably.
So, our Chief Sales Officer created a standardized AI prompt that every rep could use with their market report.
The prompt asks AI to help identify markets that may need attention, review account activity, develop prospect lists, think through product and rate opportunities, anticipate objections and turn all of that into a six-week call plan. It also includes guardrails telling AI to rely only on the numbers provided and not invent figures or businesses.
Now, every rep can start from the same framework while still bringing their own knowledge of their territory and customers to the plan.
And sometimes the insights aren't what you'd expect. After going through the exercise, one rep's biggest takeaway was:
“My biggest insight was that I need to update the CRM more regularly.”
That's useful, too. AI can only work with the information it has, and sometimes the process exposes where we need to give it better information.
As our team has gotten more comfortable with AI, we've started moving beyond prompts in a chat interface and building reusable tools.
One example is an internal profitability tool that can analyze market performance and give our sales team a simple way to identify areas that may deserve a closer look. The model is directional rather than absolute, but it gives reps another source of information to use when deciding where to focus their attention.
We've also built shared resources for other recurring work. Our brand guidelines have been loaded into reusable AI tools, helping employees create materials that follow the same colors, typography, logo usage and even Mailbox Merchants-specific terminology.
We also have a shared case study project. Instead of every sales rep starting from a blank chat, they can provide the initial information and work from the same structure and guardrails. Among those rules: AI can't invent quotes or metrics, and anything it creates is still a draft that needs approval.
The goal isn't just to have everyone “just use AI to do stuff faster.” It's to figure out where something useful can be built once and then shared across the team. This helps keep everything consistent without limiting human creativity.
Some of our experiments have also opened up new ways to analyze marketing decisions. We're exploring and using AI to support work such as:
That can help our team work through questions that go well beyond simply getting a mail piece out the door.
One sales rep, for example, had long suspected that the inherited targeting for a national franchise client wasn't capturing enough of its actual trade area. AI helped her turn that analysis and rationale into something the client could evaluate and act on.
The client used those insights to build its own ROI tracker and eventually committed to a new targeting profile.
That's where we're especially excited about AI. Saving time is great: but using extra analytical capacity to help our clients make better-informed marketing decisions is even more interesting.
Of course, “AI gave us an answer” and “AI gave us the right answer” aren't the same thing.
We learned that firsthand while using AI to analyze whether reopening two dormant markets made sense.
The analysis came back with a convincing argument. It correctly recognized that fixed costs were a major factor in the decision.
There was just one problem: it was using cost data from 2018 and 2019.
Our CEO caught the outdated data during review. We updated the model to reflect current conditions and ran the analysis again using our own profitability tools.
It's a good example of both what makes AI useful and where its limits show up. It can process information, structure an analysis and help us see things differently. But it doesn't know our business the way our people do.
That’s an important reality check we’ve kept in mind when playing with AI: it will confidently give you the wrong answer. That’s why we also built out organization-wide guidelines around sources of truth, confidential information and checking work that informs operational decisions. We're also intentionally moving more slowly when it comes to giving AI access to production systems for that reason.
We're still figuring that out, which is part of the fun!
The way our team uses AI today already looks different than it did when we first rolled it out. ool. Sometimes it's an entirely new way to look at a client's market. And as our team keeps experimenting, we're expecting that list of cool things to keep growing.
Mailbox Merchants is the Northwest’s largest independently owned direct mail, multi-channel and digital marketing provider for local businesses. We marry the old and the new by tapping into the perpetual power of print media and using technology to strategically segment, target behaviors, and optimize performance through measurable marketing results.
Image created using ChatGPT and then edited in Photoshop, 10/1/26
©2026 Mailbox Merchants | All Rights Reserved | Privacy Policy | Terms of Use | Disclaimer | Cookie Policy | Minneapolis Web Design by BizzyWeb
