In B2B products, users often deal with abstract concepts, technical terminology, and high cognitive load. That’s why mascots became more than just a branding element. They turned into a tool that helped make complex systems feel more human, approachable, and memorable.

Long before design systems, brand guidelines, or UX patterns existed, mascots were already helping people form emotional connections with products and services.

Nobody calls Mickey Mouse a “design asset.” But that’s exactly what he is and has been since 1928. Same with Ronald McDonald or the Michelin Man. They became cultural phenomena that made a brand feel like something you could actually like, not just buy from.

Today, the context is different, but the role of mascots remains surprisingly similar. The Duolingo Owl became a popular meme because it never stops reminding users about lessons. Octocat lives on every developer’s laptop sticker collection, and the Android robot is… Well, everywhere.

These aren’t decorations. Mascots make complex ideas feel human and soften products that would otherwise feel cold, and hold a brand together across touchpoints.

In this article, I’ll share how we built mascots for two B2B SaaS products at Railsware (TitanApps and Coupler.io)—from early research and rough sketches to guidelines, AI experiments, and the moment these digital characters stepped off the screen to become physical objects.

Researching and positioning: Why TitanApps needed mascots

It all started with a TitanApps blog design. (TitanApps is a set of productivity tools for Jira and monday.com, designed for professional teams that want to get more done without the manual overhead).

The product deals with workflows, hierarchies, automations, and all the messy logic behind complex systems. It is quite powerful and tech-heavy, so explaining it conventionally gets dry and boring fast. We needed a way to make the blog more visual, more memorable, and easier to connect with without dumbing anything down.

Together with the Head of Marketing, we looked at what others in the space were doing. Most brands stick to safe, functional visuals. That’s when we decided to try creating our own characters for blog illustrations. We’ve noticed for a long time that when a brand uses a mascot, it immediately feels more alive, becomes more recognizable, and is easier to remember. So we moved forward with the understanding that if this approach succeeded, it could eventually scale.

Creating a mascot system

We ran a bunch of brainstorming sessions with marketing and content teams and quickly fell into a rabbit hole of cartoon nostalgia, reviewing references, sketching, and debating different directions.

Since TitanApps develops apps for Jira, much of our content revolves around helping users navigate Jira concepts and workflows. As we explored mascot ideas, we realized that the characters needed to reflect the environment our audience interacted with every day.

At first, we searched for one perfect hero to explain everything. But Jira itself doesn’t work that way. It’s a layered system where multiple entities interact with one another.

Trying to compress that complexity into a single character felt unnatural. Instead, we designed a small cast of mascots, each representing concepts people encounter in their everyday work.

That’s how five characters appeared, not random designs, but a system that mirrors Jira itself:

  • Epic: The big entity. Large, important, and the gentle giant of the group
  • Task: Smaller, courageous, and capable of getting things done
  • Subtask: Tiny, comical, the micro-helper (the cute kid of this “family”)
  • Story: The nerd/engineer archetype, contrast, and texture for storytelling
  • Bug: Not evil, more a comedic source of trouble you can laugh at

We created all five at once so they could work as a family, mix and match, and give flexibility. Their size, personality, and behavior directly reflect Jira concepts, down to small visual cues like Epic badges, Task marks, Subtask details, and Story icons.

The final result did exactly what we hoped: It made workflows easier to visualize and share. But it also sparked a new question: Could we use the same approach to TitanApps products themselves?

Testing early via blog as a safe sandbox

Our team didn’t overthink at this stage. I took a few of our popular articles and re-illustrated them with the new characters. Nothing fancy, just trying various visuals and analyzing them.

In about three weeks, we got 3–4 cover directions and enough real examples to confirm the direction is working.

However, during early testing, we realized we couldn’t introduce a single mascot for the TitanApps brand. Like the complexity in Jira itself, TitanApps spans several products, and one character couldn’t stretch across them all or translate each product’s concepts into simple illustrations. So we decided to give every product its own mascot.

We started with our most popular product, Smart Checklist. We needed a character that fit our existing mascot family and could sit alongside the others on blog covers. Since most Jira mascots were humanoid, we wanted something different but still tied to tech. So we landed on a sleek robot with the product logo as a sticker. It gave us the contrast we wanted while still fitting the overall style.

Smart Templates was where we hit a wall. Another robot felt too cliché, and we wanted each mascot to have its own personality. So we mixed technology with something more emotional. Not a human, not another machine, but an animal with a special power or futuristic gadget. Since Smart Templates is all about duplication, the idea came naturally: Two little chicks flying in a UFO with the uncanny ability to turn almost anything into a template. This alien spaceship became their vehicle to travel across our blog covers, completing one mission at a time.

Keeping everything consistent in the guidelines

Once the system started working, the next challenge was consistency.

We brought in several freelancers, and it quickly became obvious that everyone has their own style and way of interpreting the same brief. Great for exploration, but tricky when you want one unified world.

So we went deeper than just visual rules. We built a detailed guideline that covered not only how things should look but also how to think when creating covers and illustrations. At the same time, I naturally stepped into a lead designer role. I helped to keep everyone aligned, reviewed early outputs, and gently steered things back when they drifted off track.

The aim was simple: No matter who created the piece, it should feel as if it came from the same world, without any visible authorial fingerprints.

Building a hybrid AI workflow that actually scales

To keep up with the workload, AI became our favorite sidekick for quick ideas. Here’s how we

set it up to run on its own:

Quick ideas and AI drafts

I used a prompt to generate 10 visual concepts for any new article. I’d pick the best one, create a quick draft using AI (GPT, Midjourney, and NanoBanana), and pass it to our designers to finish up.

However, anyone who has worked with generative AI knows the ultimate pain point: Character consistency. To be honest, we couldn’t achieve 100% consistency purely through neural networks, as these tools still struggle to replicate identical characters across different scenes. To crack this, we developed a simple, three-step hybrid workflow:

  • First, we intentionally chose very simple, geometric shapes for the mascots from the start, making them much easier to replicate across different scenes
  • Second, we manually created a proper character sheet—a turnaround with front and profile views—so the designers could clearly understand the character’s anatomy and how it should look in any setup
  • Third, we shifted the AI’s role: We used it only to generate the core concepts and basic variations, and then our design team took over, fine-tuning and polishing the final vector assets in a tight feedback loop

Teaching the designers

At first, I left detailed comments on their work so they could really understand the vibe. Once it clicked, I just handed them my AI prompt so they could brainstorm ideas on their own.

Looping in the content team

I also showed the content managers how the whole system works, including how to use the guidelines to choose the most suitable cover direction for each article, so everyone was on the same page. Now, the designers and content managers work together completely on their own. The process runs itself, and the graphics still look exactly the way they did on day one.

Scaling the mascots’ usage: Product, merch, and physical objects

The blog was the first place the mascots showed they could be more than just nice visuals. Once we added them to our blog redesign, we started noticing small but steady shifts in how people used the site. Time on page grew by 18% over six months. Key conversion events went up by almost 18%, too, meaning more readers ended up signing up. Engagement rate and session time also improved, which felt like people were actually staying with the content, not just scrolling past it. Surely, these improvements were the result of the broader blog redesign rather than the mascots alone. However, the characters became a meaningful part of the new visual system and played an important role in making the content feel more approachable and memorable.

Still, the most convincing feedback didn’t come from numbers. It came from people in real life. The mascots felt strong enough that we wanted to see how they’d work outside the blog. So we brought them to Atlassian conferences in the US as stickers and small swag items. The reaction was immediate and very human. People picked them up, kept them, and started talking about them. And guests who saw them one year came back the next year asking for the mascots again, this time looking specifically for our Epic toy.

From there, it just naturally grew into more formats. We made T-shirts, banners for events, plush toys, and other physical pieces.

We never treated it like “apply design everywhere.” Every format had its own reality. A sticker is not a banner, and a plush toy is not a blog illustration. So each one needed a bit of rethinking and small adjustments to make sure the mascots still felt right in that new form.

Creating a new mascot with a new approach: Coupler.io and an AI-first mascot

After TitanApps, I was asked to create a mascot for another Railsware product: Coupler.io. It’s a data automation platform that syncs data between different sources. The team was building a new AI agent feature and needed a character to make the technology feel fun and approachable.

I treated it as a fresh experiment: A mascot built mostly with AI. The approach was based on a simple formula: Combining two stylistic references and a prompt describing the intended final result. The idea centered on a cloud, since the product lives in the data world. The first version wore a helmet. Then we started iterating: A Sherlock Holmes–style deerstalker hat, then headphones, a cap, a beret, and many more.

All of this took just 5–6 hours using AI. It was my first experience working on a character concept this quickly. We landed on the cloud with the deerstalker hat. It fits the AI agent perfectly: Always searching for important data like a detective, while still feeling cute and approachable.

What I’d do again and what I’d do differently

Looking back, building a mascot system is less about art and more about managing teamwork, tools, and expectations. Here is the down-to-earth advice I’d give to any designer tackling this:

1. Use AI as your junior assistant

Don’t overthink the AI debate; just use it to move faster. When you need to explore ideas, AI is the best tool for quick, messy sketching. It lets you test 20 different vibes in an hour instead of spending days drawing them out by hand.

Use AI prompts as a handoff tool. In some cases, they can communicate the intended direction more efficiently than long, detailed briefs.

2. Stop designing in a vacuum

A mascot isn’t just a random drawing. It has to do a job for the business. That means you can’t build it alone in your design corner. You need the marketing, content, and product teams in the room from day one.

Never accept or write a brief that just says, “We need a mascot.” A real brief needs to answer: What problem are we solving? Where will this character live? What should people feel when they see it?

3. Build a massive reference library

Before you start designing, go on a massive hoarding session. Spend a few hours saving screenshots of cartoons, vintage logos, Pinterest boards, anything that catches your eye.

For our second mascot, I spent about five hours just collecting random images and two hours organizing them. This library won’t give you the final answer right away, but it gives you a solid foundation. When you’re ready, you can feed these specific visual ideas into AI to start generating variations.

4. Over-communicate the raw drafts

The biggest trap is spending two weeks polishing a character, only to show the team and realize you completely missed the mark.

Share your rough, ugly sketches early. Talk about the personality, test out different expressions, and explain your logic as you go. When the team is part of the messy middle, the final mascot actually feels right to everyone.

Mascots can deliver great results, even for products that aren’t widely known. The key is to focus on expertise, thoughtful approaches, and creative execution. The most important thing to remember is that designers aren’t a Swiss Army knife. They need context, experimentation, and a clear vision.