Staff product designer · Data, insights and AI products
Based in Barcelona, Spain · Open to EMEA roles

I find the assumption that would
sink an AI or data product, and test it
before the team builds around it.

15+ years, from a startup’s first UX team to AWS CloudWatch and HubSpot’s AI products.
I facilitate teams through ambiguity: I turn complex system data into decisions people can act on,
and bring everyone along with the evidence.

  • ~1 quarter

    of build saved: I tested AI output on customers’ own data, it missed the bar, and we moved the investment to a higher-impact bet.

  • 346k

    monthly views kept on an AWS page one org wanted to kill. My research showed it carried 27% of the console’s traffic.

  • Funded

    HubSpot’s agentic pages private beta, on the evidence of my customer quality test (content rated 4–4.5 / 5). Live since September 2026.

Laura Montefusco presenting at UXDX Dublin 2026, in front of her slide of questions to diagnose your current or next team

Public speaking

UX is a Team Sport · UXDX Dublin 2026

How design teams raise UX maturity without authority: three stories from a startup, a big tech company and a scale-up.

About the talk and slides (30 minutes)

Selected work

AI template setup: choose the page structure and audience, start from an account-based preset or write instructions, with a live preview of the personalized page

Agentic AI · 0→1 · Marketing + Sales · HubSpot

From 200+ manual steps to one chat: agentic, personalized pages for every target account

Customers were taking account-based marketing to competitors: 50 personalized pages took 200+ manual steps. I tested AI output with customers first: content passed, design didn’t, so I made design deterministic and kept AI to the content. That evidence funded a private beta, live since September 2026.

Reach: set the UX leadership review agenda; mapped the vision across Marketing, Sales, CRM and Data Hub; wrote the handover the receiving teams now build on.

Read the case study

The V2 service map vision: services grouped by application, health on each group and the failing path highlighted in red

Observability · Migration · AWS CloudWatch

One APM console from two competing products

The only designer merging two AWS orgs’ tracing products. Research challenged both orgs: the page one wanted to kill carried 27% of traffic, and the celebrated service map broke at scale. We moved 1.3M+ monthly views from one console to the other, and users moved across, not away.

Reach: aligned two orgs on one priority list; my V2 vision secured GM investment; five UI teams adopted my feedback triage process.

Read the case study

The marketing email list with red and green performance indicators and an insights panel with a recommended action

Insights · AI agent · HubSpot

Turning a low-adoption beta into a multi-year strategy

An Inbound 2024 launch reached only 7.5% of customer portals in beta. I split the problem into reach and impact and took on impact: happy CSAT rose from 38% to 48%. Then I laid the foundations of HubSpot’s QA agent and a multi-year vision for messaging reporting.

Reach: leadership adopted my vision; it moved insights into the email editor, beyond my triad’s scope, and became the foundation of the QA agent, launched at Unbound 2026.

Read the case study

Calls I made to stop or redirect

01

Stopped an AI experiment before it burned a quarter

I named AI output quality as the riskiest assumption and tested it on customers’ own data. It missed the bar, so we stopped about a quarter’s build and moved the investment to a higher-impact bet.

02

Redirected a roadmap with customer evidence

30+ EMEA customers ranked their priorities in a panel I ran. Their evidence moved CloudWatch’s 2022 investment from aspirational features to cross-account and cross-region, and went into annual planning.

03

Never built: the estimate moved the bet

I mapped a variants scope across six product surfaces and two editors, and recommended designing it once instead of rebuilding it twice. The estimates put the real cost on the table before any of it was built, and the team moved to a higher-priority bet.

Earlier work

The same skill at growing scale: building a practice from zero, then owning AWS’s hardest observability data, then calling the AI bets at HubSpot.

  • 2026Human–AI interaction audit · Claude skillBuilt to audit AI and agentic features from the user’s side: agency, trust calibration, automation fit and problem framing. Grounded in guidance from Microsoft, Google PAIR, Apple, IBM and Anthropic.
  • 2025Email Delivery Dashboard · HubSpotAligned four teams to a private-beta MVP in about two months. 34% of eligible portals adopted it, 61% return weekly. HubSpot docs
  • 2022–23Service Level Objectives · AWSOwned native SLOs from first concept to private beta; shipped in CloudWatch Application Signals. AWS blog
  • 2022Dashboards: raising the data-ink ratio · AWSUsers said the new UI took “5x more space”. The audit found ~40% of the screen carried no data, so I cut the chrome.
  • 2020Metric Explorer · AWSTag-based metric queries. The Washington Post’s Arc XP used it to cut Lambda costs by 20%. AWS blog
  • 2014–19Built the UX practice · ForceManagerFrom zero to a team of three. Voice AI research for sales reps shaped a two-year roadmap.

How I work

De-risk first

Find the assumption that would sink the product and test it before anyone builds around it.

Discover with customers, iterate fast

I start in discovery with real customers on their own data, then put working prototypes with realistic data in front of them within days, so each round answers a question instead of opening one.

Stop what shouldn’t ship

Saying no early is cheaper than saying sorry after launch.

Bring the teams along

Most of my products span several teams, so alignment is part of the design work, not a meeting after it.

“She has a real talent for understanding both the needs of users and the technical realities of the systems she’s designing for.”
Mike Meaney, Sr Principal Software Engineer, Amazon
Worked with Laura at AWS CloudWatch on the same team
“[Quote about one specific quality, from someone who saw it in a big room]”
[Name], [Title], [Company]