The End of Endless Meetings: Smart Regression with AI
It all started with a Notion full of bullet points, ten-plus people on a call, and a question we all know too well: “Are we forgetting something?”
That’s how our regression sessions began on complex projects, where every single feature had a direct impact on thousands of users. We wanted to cover everything, predict the unpredictable, leave no blind spots. But coordination was becoming unsustainable. We knew there had to be another way.
In this piece, I’ll share how we at Paisanos turned those marathons into an agile, intelligent process, and how our experience tackling complex projects led us to actively contribute to the evolution of one of today’s most powerful QA tools.
A change that not only eased the load on the team, but actually raised quality without slowing down delivery. Spoiler: it’s not about automating more, but automating smarter.
Our Path: QA in High-Complexity Projects
At Paisanos, we specialise in highly demanding projects, always with a clear focus on the end-user experience.
We build mobile and web applications that need to perform flawlessly under pressure, serving massive user bases with mission-critical features. Each project brings its own challenges from e-commerce platforms with intricate payment flows to entertainment products with thousands of simultaneous users.
Not so long ago, our approach to software quality relied almost entirely on manual, repetitive tasks. Endless documentation, exhaustive test-case design, manual regression on every release, field-by-field form validation. All necessary, but time-consuming and resource-intensive rarely aligned with business speed. Automation came as a relief, offering consistency and speed, yet demanded heavy technical investment and maintenance that wasn’t always sustainable.
Over time, we learned that each project comes with its own pains.
In mobile: dozens of devices, OS versions, unpredictable native behaviours.
In web: the explosion of browsers, resolutions, and complex session states.
And when both ecosystems merge complexity multiplies exponentially.
With agile methodologies and the rise of continuous delivery, the landscape shifted dramatically. Our clients expected frequent releases, leaving no room for regressions that could compromise the user experience.
Constant iterations and deployments raised the inevitable question: how do we ensure quality without slowing velocity?
The answer was no longer “just automate more.”
It required a mindset shift involving QA from discovery, fostering collaboration from a critical, analytical perspective, engaging the entire team in quality-driven rituals, and automating with strategic intent: combining exploratory testing, automated regression, and shared business vision.
As usual, we faced those challenges head-on. Tight deadlines became the rule, not the exception. We had to rethink how we approached QA not just to keep up, but to lead with quality without becoming a bottleneck.
That’s when Artificial Intelligence began to open new doors. Not as a distant promise, but as a genuine enabler of more agile models and processes.
AI can now suggest test scenarios while humans refine and contextualise them, or even generate scripts for teams without programming experience. In short: it made testing more inclusive, accelerating learning and empowering teams.
AI doesn’t replace human judgment, it amplifies it. Cycles are shorter, quality pressure is higher, and organisations need answers that keep up.
Maybe AI won’t solve every problem, but it acts as a beacon, illuminating a path where QA is no longer “the final step,” but a strategic driver of value, moving at the same speed as the business while maintaining its depth and intent.
Boca: When Complexity Gets Real
This evolution wasn’t just theoretical. The Boca project brought all these challenges together, testing everything we’d built as an organisation.
You’ve probably faced a similar situation: regardless of the product’s specifics, when someone asks about the most critical test cases, your mind instantly starts listing scenarios, one after another, like a stack of Russian nesting dolls.
In this project, our regression sessions had become corporate marathons. With so many moving parts, they often ran for hours and gathered over ten people, developers, QAs, product owners, business analysts, each contributing perspectives, edge cases, and institutional knowledge that felt impossible to systematise.
When you’re dealing with an ecosystem serving millions of users, multiple profiles, events, and activities, the list of cases becomes overwhelming.
Every functionality directly affects the experience of a massive user base. Deadlines were demanding, features overlapped in development, and each release could potentially impact critical production flows.
Necessary, yes, but unsustainable.
The Technical Dilemma from Our Trenches
Automation wasn’t optional, it was a must. But at Paisanos, we knew exactly what hurdles a traditional approach would bring. Chances are, you’ve faced the same:
- The technical learning curve: We’d been there before. Traditional frameworks require QAs to know programming languages, design patterns, and complex selector management. In a diverse team under delivery pressure, that kind of time investment is rarely feasible, especially for short- or mid-term projects. Training everyone would mean weeks or months before seeing tangible, scalable results.
- Ongoing maintenance: We’d also lived through it: every UI change triggers a cascade of script updates, broken selectors, outdated flows, failed syncs. One frontend refactor can translate into hours of test repairs instead of progress. With our iteration speed, technical debt would outgrow coverage in no time.
- Test-data management: One of our biggest pains. Coordinating test users, states, configurations, and environments requires extra infrastructure and constant manual sync. Who creates the data? How is it refreshed? What happens when two tests need the same user at once?
- Scalability: Running full regression suites in parallel with traditional frameworks means configuring execution grids, managing infrastructure capacity, and handling inconsistencies between runs. More complexity. More time. More resources.
It was clear: we needed automation, but the traditional path would only slow us down.
Our Collaboration with Autonoma: Co-Creating the Solution
The question remained:
How could we capture collective knowledge without needing everyone in the same room?
How could we automate without sacrificing speed?
How could we democratise automation without turning QAs into developers?
The answer came from exploring tools that aligned with our reality. As QAs at Paisanos, we see it as our responsibility to stay open to emerging technologies evaluating how they can positively impact our processes and delivery.
That’s how our collaboration with Autonoma began. But this wasn’t a simple tool adoption. It was a strategic partnership where our experience, pains, and vision actively shaped the product itself.
Through multiple sessions, we poured years of project insights into the conversation. QAs and Devs alike shared our specific challenges: the pace of continuous delivery, short development cycles, lack of autonomy in test-data management. Each meeting helped the Autonoma team understand not just what we needed, but why.
We didn’t just bring abstract requirements we shared real use cases, complex flows, edge scenarios. We showed them how we worked, where we got stuck, what kept us up at night. This two-way collaboration was key: they iterated the product, we validated in real contexts, and fed back insights only those on the front lines could provide.
Our vision for a modern QA tool directly influenced Autonoma’s roadmap. Their no-code approach fit our use case perfectly. Smart test-variable management evolved from our pains with data handling. The intuitive environment setup took shape based on how we switched daily between dev, staging, and production with scalable parallel execution, no extra infrastructure needed.
In short: we helped build the tool we needed and in doing so, helped create a solution that now benefits other teams facing similar challenges.
The result was exactly what we’d been looking for: a way to democratise automation without losing analytical depth powered by AI. The no-code approach removed the technical barrier entirely. Any team member, regardless of coding background, could design and run complex tests within minutes.
Of course, implementing Autonoma in the Boca project didn’t magically fix everything overnight. It required close coordination with the club’s internal team, aligning processes for test users, events, products, data syncs, and states. That collaboration was crucial for the tool to truly shine.
Once aligned, what used to take hours of back-and-forth could now be handled by a single person running full regression cycles while maintaining both technical depth and business context.
Beyond process transformation, Autonoma brought technical visibility in a world of constant iteration. Execution metrics now help us identify performance trends, spot unstable flows early, and track quality progress over time. This continuous insight became invaluable, allowing us to anticipate issues before they hit users. Transparent data and results built confidence in this new way of working, especially given the responsibility of validating experiences for millions.
More Than Efficiency: Cultural and Strategic Impact
The impact at Paisanos went far beyond operational efficiency. The time saved is now reinvested into expanding coverage, exploring more edge cases, and analysing risk.
AI accelerated adoption, boosted adaptability, and freed our team to focus on strategy and continuous improvement.
Ultimately, automation didn’t reduce effort, it redirected it toward where it creates the most value.
Thinking about how to integrate AI into your processes without losing quality or control? Let’s talk.
At Paisanos, we believe artificial intelligence doesn’t replace human judgment, it amplifies it.
We work side by side with organisations that want to evolve their processes without sacrificing quality, context, or speed. We believe in building solutions that adapt to people, not the other way around. Because when technology empowers teams, what once felt impossible becomes part of the everyday.