AB Tasty and VWO have united, because half the loop was never enough.

Learn about the merger

Blog

When New Ideas Take Flight: How PagerDuty Put Wingz to Work on Experimentation Reporting

5 Min Read
When New Ideas Take Flight: How PagerDuty Put Wingz to Work on Experimentation Reporting

Back to Blog

For some people, the sunshine follows them wherever they go – and the same can certainly be said for Laura Freire, Senior Growth Marketing Manager, from PagerDuty,

When Laura arrived at Experience Next Amsterdam from Lisbon – originally being from the “Hawaii of Europe”, or Madeira – she opened her session by marvelling in a surprisingly  sunny Amsterdam – originally think that copious amount of sunshine were a Lisbon-only phenomenon. 

Having grown up with nature and the outdoors, Laura shared how she surfs when she can, skis whenever winter allows, and even founded a community called Motion that brings people together through sport. 

Similar to her adventurous spirit outside of the office, she also had a confession for the room: she’s always looking for a way to be outside – and, apparently, a way to get her reporting to run itself.

That’s exactly why PagerDuty and Wingify have made the perfect match to soar above the guesswork – specifically, by using Wingz as their primary automation tool. 

Meet PagerDuty (and its small, mighty team)

PagerDuty is an American, publicly traded company headquartered in San Francisco – with an office in Lisbon and teams across other countries. 

PagerDuty’s  main product is a reliability platform that helps companies detect, prevent, and fix incidents in real time – all before  their customers even notice. 

Laura has been with the company for four years, and within a marketing team of more than 100 people, she leads web experimentation and personalization with a “small, mighty” team of four: 

  • Erica (Data Analyst)
  • Joao (Web Developer)
  • Sir John (UX Designer)

Having started this journey alone, she initially onboarded with Wingify three years ago to establish an experimentation framework – and then the team came along to help her skyrocket it even further.

The challenge: great insights, terrible timing

PagerDuty is a product-led growth company, which means everything starts with a trial. 

That trial flow has several steps, meaning that every experiment on it is essentially its own multivariate test. The challenge with this is that each individual multivariate test can take what feels like forever to analyze, and then when you add behavioural data, heat maps, and recording in the mix – and you have a mountain as high as the ones in Madeira of that can make the data genuinely valuable, but very slow to process.

The bigger problem was what happened next: those insights got stuck on a slide deck. Since access to the experimentation platform isn’t available to all stakeholders, Laura maintained a deck that documented every experiment since 2023.

This back-up of previous experiments included a visual of the control and variant, the metrics, whether it won, and what was recommended in the end. As her CMO didn’t want to log into an app and hunt for data, Laura built those slides by hand – every single time. 

In turn, she kept fielding the same questions: have we done this? Have we done that? What were the results?

Luckily, Wingify had the answers to help Laura from asking her the same questions on repeat – and instead, allow her to better feed the relevance loop. 

How they overcame it: getting the tools to talk

The turning point came from a conversation with PagerDuty’s customer success manager at Wingify, who introduced her to Wingz, which is Wingify’s AI workflows – and the MCP connector that lets Wingz and Claude stay in communication with one another. 

Her first use case was to ease the reporting. With Wingz, Laura could create a workflow triggered when an experiment ends, and instead of dragging and dropping boxes – she simply chatted with the agents, which built the flow for her. 

Her prompt asked for a summarized version of recordings and heat maps for a campaign once the experiment was paused, delivered by email in the format of her slide deck. Two or three iterations later, it worked.

She then connected Claude through the MCP connector, tweaking the same prompt to send the report weekly over Slack, where PagerDuty’s teams already live – which allowed the whole growth pod to see it, as opposed to staying buried in her inbox.

The results: an hour becomes ten minutes

The numbers speak for themselves. 

A trial-flow multivariate test that once took Laura an hour or more to analyze, suddenly only takes a mere ten minutes on average — as she can simply check for mismatches and hallucinations and cross-reference with a statistical calculator. Writing the summary went from a chore to zero, as she instructed the format once that’s been used ever since. In total, building both flows only took about two days. 

The team kept what matters: its biweekly sync – where the insights become decisions about what to double down on, prioritize, or drop.

Wingz overall made it much easier for PagerDuty to maximize their time and dedicate their efforts to the growth where it mattered most. 

Psgerduty Wingz Workflows

Use case two: an experimentation assistant for everyone

Laura’s second use case tackled the recurring questions directly. 

She built a Claude project called “Experimentation AI Assistance,” grounded in the slide deck as the source of truth – with one non-negotiable instruction: don’t fabricate the data. 

Automated tools like these usually refresh weekly and monthly to ensure anyone asking about experiments always gets the most updated version. If you ask whether the web team tested reducing trial-flow steps, it’ll answer with the slides, the takeaways, and even a roadmap with recommendations.

A second version is already in progress, allowing the team to ask directly in Slack. Today, five or six people continue to test this tool – including PagerDuty’s directors of commercial and enterprise and its CMO.

The takeaway

Laura’s message under the Amsterdam sun was refreshingly simple: most of us already use these tools – the main trick is getting them to talk to each other. 

Connect the platform where experiments run to the assistant your team already trusts, and the reporting that used to eat your week will start to write itself – which leaves more time for the things that require a human’s expertise – and, if you’re Laura, a little more time outdoors.

Stephanie Safdie

Stephanie Safdie holds a bachelor's degree in English Language and Literature from the University of Maryland, specializing in multimedia studies. She has worked as a social media video creator, freelance copywriter, SEO copywriter at Greenly climate tech, and runs a travel blog Destination Dreamer Diaries.

You might also love to read these

From Experiments to Better Decisions: How DNZ and Studio Anneloes Are Redesigning Experimentation
5 Min Read

From Experiments to Better Decisions: How DNZ and Studio Anneloes Are Redesigning Experimentation

How a City and a Sports Media Giant Work with Wingify to Test What Works: Their Passport to Success
4 Min Read

How a City and a Sports Media Giant Work with Wingify to Test What Works: Their Passport to Success

Hitting the Ground Running: How ASICS Turned Testing Speed Into Experimentation Velocity
5 Min Read

Hitting the Ground Running: How ASICS Turned Testing Speed Into Experimentation Velocity

The better way to test,
learn and grow

Get a Demo Start Trial