Experts

Assembled per problem, not maintained as a bench.

Frontier models compress the analytical work. Judgement still has to come from people who have done the thing. So the model is a small network of senior specialists brought in when a problem needs them, you pay for the expertise the question requires, not for a standing organisation.

Featured
Michael Beutler, climate strategy and carbon reporting leader

Michael Beutler

Climate strategy, carbon reporting & decarbonisation

Michael has spent fifteen years building the systems companies use to measure, report and reduce climate impact, and to understand what climate change does to their business in return.

As Chief Sustainability Officer of Moncler Group he led the Group's climate strategy: SBTi-validated emissions targets (70% absolute Scope 1 and 2 cut by 2030, 30% Scope 3 by 2032, net zero by 2050), the transition plan underneath them, decarbonisation across a supply-chain-dominated footprint, and the reporting architecture that has to withstand assurance and regulatory scrutiny.

Before that, as Director of Sustainability Operations at Kering, he led the rollout of the world's first Environmental Profit & Loss (EP&L) account across a global luxury group, brand by brand, consolidated at Group level, tracing supply chains tier by tier back to raw material production. He drove the open-sourcing of the methodology in 2015; it became an international reference well beyond fashion and revealed that more than 66% of impact sat in raw materials.

Earlier roles include six years at SAP (latterly Global Director of Sustainability and Strategy) plus finance and operations positions at DHL, PwC and Ford. He was previously CEO of Pebble.ai, focused on taking frontier AI out of the black box and making it operable inside sustainability functions.

Credentials

  • Former CEO, Pebble.ai
  • Responsible AI Advisory Board, AI 2030
  • ex-CSO, Moncler Group
  • ex-Director of Sustainability Operations, Kering
  • BA Finance (Michigan State), MBA Operations, JD Environmental Law (Colorado)
Panu Kärävä, sustainability data and energy leader

Panu Kärävä

Data, energy & circularity

Panu is a business and technology leader who has spent two decades at the intersection of sustainability, energy and industrial data. He spent nearly eight years at Siemens in global portfolio roles spanning asset performance, energy and sustainability, followed by a senior sustainability consulting role at AWS, designing cloud-based approaches for carbon accounting, ESG reporting, traceability and IoT-driven energy optimisation across buildings and industry.

He is now a Partner at E2 Management Consulting in Zürich, working across data-driven sustainability, digitalisation and sustainable IT, circular business models, and the application of AI to sustainability challenges. His work runs from board-level strategy to proof-of-concept code. As interim Sustainability Director at augment.eco, he translated the ESG requirements of eighteen impact investors into a practical programme ahead of a financing round, addressing priority gaps and leading Scope 1, 2 and 3 GHG accounting alongside supplier monitoring criteria and a code of ethics.

Panu is also active as a founder and advisor in utility-scale battery energy storage, combining energy strategy, project development and digital modelling, including work on PV-BESS simulation. He advises Finnish risk management software company Inclus, bringing a sustainability and ESG perspective to risk analytics. In circularity, he is an investor and Chairman of Ninyes, a turnkey resale platform enabling fashion brands to integrate recommerce into their business models. He is based in Zug, Switzerland.

Credentials

  • Partner, E2 Management Consulting (Zürich)
  • ex-Siemens · ex-AWS
  • Founder and advisor, utility-scale battery energy storage
  • Advisor to Inclus (Finnish risk management software)
  • Chairman, Ninyes

Neil Brown

Capital markets & investor view

Neil has spent more than twenty years inside sustainable and responsible investment. Most recently he was Head of Equities at GIB Asset Management, where he launched and ran the firm's global sustainable equity strategies and built out its emerging-market and European teams.

Before that he was a partner and fund manager at Liontrust, and an SRI fund manager at Alliance Trust Investments and Aviva Investors, running pan-European and ethical equity mandates. His career began in governance research, senior researcher at PIRC, then Pan-European analyst and Head of Governance and Responsible Investment at Threadneedle.

He has also sat on the market-shaping side of the table: member of the UN-supported Principles for Responsible Investment (PRI) listed equity steering committee, natural capital forum speaker, advisory board member at Impactive Tech, and judge for impact investing competitions. On AI he is deliberately augmentation-first, having completed Stanford's Leadership in the Age of Generative AI programme, the interest is in sharpening human investment judgement rather than automating it.

What he brings to a piece of work is the allocator's read: how a transition plan or disclosure actually lands with investors, and which claims move a rating rather than a page count.

Credentials

  • Ex-Head of Equities, GIB Asset Management
  • ex-Liontrust partner and fund manager
  • UN PRI listed equity steering committee
  • Stanford Leadership in the Age of Generative AI
  • MSc Development Economics (SOAS), BA Economics (Manchester), CFA UK IMC
Kyle Cheng, applied AI engineer and platform lead

Kyle Cheng

Frontier AI engineering & agentic systems

Kyle is an applied AI engineer and platform lead who builds secure AI agents for complex enterprise workflows. He currently leads the build-out of a production AI-agent platform at a financial-services firm, directing its architecture, security and engineering team.

Previously, Kyle was a Member of Technical Staff at Anthropic, where he built applied AI systems for internal business teams, including secure data access, high-throughput data pipelines and semantic search over large unstructured datasets. Earlier, he built Amazon CodeWhisperer’s Chat integration for JetBrains at AWS and an internal data platform at Oracle used by more than 60,000 employees, credited with approximately $7 million in annual savings.

His work focuses on the production layer that determines whether an AI system can be trusted: permissions, data access, agent security, evaluation, traceability and human oversight. On sustainability engagements, he brings that technical discipline to messy supplier, disclosure and operational data, working alongside domain specialists who own the climate judgement.

Credentials

  • Leads development of a production AI-agent platform for a financial-services firm
  • Former Member of Technical Staff, Anthropic
  • Built Amazon CodeWhisperer’s Chat integration for JetBrains
  • Built an Oracle platform used by 60,000+ employees, saving approximately $7M annually
  • Created and taught AI-agent development courses at Interview Kickstart
  • Open source: Seren AI (Flutter and LangGraph)
Domains covered

Climate, nature and impact accounting

Valuation methodology, impact accounting, and the upstream lens that surfaces exposure a standard risk review scopes out.

Supply chain and operations

Production planning, inventory and replenishment discipline, supplier data quality, where Scope 3 leverage actually sits.

Disclosure and assurance

CSRD, ISSB and sector regimes: boundary choices, restatements, and what survives an auditor or a hostile reader.

Frontier AI engineering

Model selection, evaluation, retrieval over messy document estates, and deployment inside enterprise governance constraints.

Geospatial and climate science

Satellite and hazard data, scenario framing, and translating physical risk into operating and procurement decisions.

Capital and investor view

How disclosure and transition plans read from the allocator side, and which claims move a rating or a price.

How it works

  1. Scope the question first. The problem statement determines which specialisms are genuinely needed.
  2. Name the people. You see who is on the work and why, before it starts.
  3. Keep the team small. No pyramid, no juniors learning on your budget.
Book the Sprint