Löwe
EN

Book a conversation
Portfolio/Q&A
Select a topic or ask a question
Felipe Löwe
Product Leader

What experience do you bring to product leadership?

I bring ten years of product experience across Honeywell, Palantir and Thomson Reuters, covering industrial operations, operational AI and enterprise software. My remit grew from orders and billing to PM leadership, investment priorities, AI release policy and enterprise client commitments across multidisciplinary teams.

Suggested next stepExplore the three case studies

Explore answers about how I work and the experience behind the case studies.

Overview playing · 1 / 3Explore how I work

Recent work

Thomson Reuters2024 / 2026
  • Enterprise AI
  • Legal & tax
  • Professional software

Making legal AI ready for enterprise use.

CoCounsel uses AI for legal research and contract review; ONESOURCE+ links tax, trade, legal and risk products. I led four PMs in a 27-person team and owned a US$12M annual budget for shared integrations, answer quality and client adoption, setting investment priorities and client commitments across the legal workflows.

Context & problem

CoCounsel reviewers took a median of 21 minutes to find a checked source across separate systems. Experts accepted 963 of 1,250 answers, so clause and exception checks guided corrections before wider use.

My response

I reviewed source checks and answer quality with the PMs, agreeing at least 95% expert acceptance while planning client access, training, paid scope and staffing with the adoption PM and commercial leads.

CoCounsel review and paid rollout
  1. Source evidenceCorrect clause and user access
  2. Expert acceptanceAt least 95% on the fixed review set
  3. Paid rolloutAgreed owner, training and staffing

Measured outcomes

Less research time
43%
21 → 12 minutes per research task46 teams · 8 organisations
Expert acceptance
96%
1,200 / 1,250 answers acceptedMatched expert reviews
Signed expansion
US$18M
6 signed enterprise agreementsThree-year contract value

4 PMs · 27-person team

Read the full case
Palantir Technologies2021 / 2024
  • AI & data
  • Enterprise software
  • Government

Taking operational AI into production.

Foundry links orders, stock and shipments; AIP brings generative AI into workflows where employees review proposed actions. I led two PMs and coordinated 20 specialists across commercial accounts in four industries, setting workflow priorities, approval rules and rollout commitments to move accepted pilots into daily use.

Context & problem

Across four clients, planners checked separate order, stock and shipment records to assess a shortage response. A ready request then waited a median of 13 hours for the next team to accept responsibility.

My response

I validated the workflow with operators and the PM, linking each approval to its owner and decision record, then used measured waits and estimated delay costs to confirm rollout scope and staffing.

Approval context in Foundry Workshop
Approval recordWorkshop
Owner
Authorised supply planner
Action
Stock transfer and approval time
Context
Order, stock and delivery window

Results & value

Less waiting
38%
13 → 8 hours between teams4 clients · 2024 review
Complete decision records
98%
392 / 400 action recordsOngoing record review
Modelled annual value
US$1.2M
4,000 handoffs × 5 h × US$60/hAssumed volume & rate · 4 clients

2 PMs · 20 specialists

Read the full case
Honeywell Technologies2016 / 2021
  • Industrial AI
  • Critical infrastructure
  • Automation

Scaling industrial software and predictive AI.

At Honeywell Technologies in Brazil, I progressed from Product Intern to Product Manager and Senior Product Manager. I led OptiVision orders and invoicing at four mills, then Forge APM predictive maintenance at three sites, owning investment priorities, model release policy and commercial commitments to industrial clients.

Context & problem

Sales at four OptiVision mills confirmed dates before stock checks and reconciled charges after dispatch. Urgent shipments and invoice corrections raised operating cost and weakened delivery promises.

My response

I reviewed orders and billing with sales, planning and finance, agreeing the domestic scope with sponsors and coordinating seven specialists to test delivery dates and charges for each mill’s release.

OptiVision order acceptance workflow
  1. Planning confirms availabilityStock and scheduled production
  2. Sales confirms the promiseCustomer accepts quantity and date
  3. Finance checks the invoiceApproved prices and freight terms

Measured outcomes

Faster order confirmation
60%
2.5 → 1.0 business days2018 median review
In full and on time
94%
84% → 94% of domestic ordersAccepted delivery dates
Net annual benefit
R$2.8M
After recurring programme costs4 mills · 2018 finance review

7 specialists

Read the full case

Fast answers

Across industrial software and enterprise AI, my work connects the customer task with the systems, teams and investment needed to support it. That understanding shapes the product plan and the tests we run, with customer value and business results guiding delivery and further investment.