1189: Why Finance Should Lead with “Yes, And” | Bruno Annicq, CFO, Wellhub
Bruno Annicq, CFO of WellHub, discusses his non-traditional path from engineering to McKinsey to becoming an 'accidental CFO,' his philosophy of finance as a business enabler with a 'yes, and' approach, and how AI-powered multi-model forecasting reduced their prediction error rate from 10% to under 2%, enabling faster reinvestment and growth.
Summary
Bruno Annicq, CFO of WellHub, shares his unconventional career trajectory that began with an engineering degree chosen as a 'diversification strategy,' followed by an MBA, strategy consulting at McKinsey, and operational roles at AOL during its transformation era. He credits AOL's chaotic, fast-changing environment for teaching him adaptability and the critical distinction between urgent and important tasks — a framework he still applies daily by reserving the first part of his day for non-urgent but important work.
At AOL, Holly Hess identified leadership potential in Bruno that he hadn't seen in himself, offering him his first CFO role over the platforms business. This pivot shaped his view that finance's core purpose is using data and analytics to tell the business story — to employees, investors, clients, and partners alike. He eventually joined WellHub after meeting co-founder Cesar and being drawn to the mission, people, and rapid growth trajectory.
WellHub operates as a B2B wellness infrastructure platform serving 50,000 corporate clients across 18 countries with 100,000 wellness partners, growing 40% year-over-year. The company operates a three-sided ecosystem: corporate clients (HR/CFO buyers), employee subscribers who pay for tiered plans, and wellness partners (gyms, studios, apps) who receive incremental new customers with zero CAC. Bruno argues this is an investment, not an expense, noting that 75% of healthcare costs are preventable but only 4% of healthcare spending goes toward prevention.
Bruno's 'yes, and' philosophy for finance centers on being a business enabler rather than a gatekeeper. He pushes his team to find ways to accommodate business investment requests rather than defaulting to budget constraints. On storytelling, he emphasizes radical simplification — critiquing 150-page board decks as noise rather than signal — and advocates distilling business performance to three core metrics: cohort performance/net revenue retention, a P×Q (price times volume) framework, and gross profit dollar growth relative to expense growth.
The most technically detailed section covers WellHub's AI-powered forecasting transformation. Their partner utilization expense — a major cost driver tied to nearly 6 million employee subscribers — was previously forecast with ~10% MAPE error. Using a multi-model approach inspired by the Windy weather app (which shows divergence across meteorological models as a proxy for forecast uncertainty), they built four competing models and reduced error to 2%. Subsequently, using Anthropic's Claude Code, a single engineer in two weeks added six new models across every country, segment, and cohort — work that previously required a large team, external vendors, and nearly a year. Error rates halved again. Bruno ran a live company-wide demo of Claude Code building a monthly reporting automation in real time, which he credits with breaking through the learning curve barrier for his team.
Bruno also discusses leadership lessons around time management (reviewing weekly calendar against stated priorities), the value of long-term thinking in founder-led businesses, and WellHub's collaborative culture. He recommends the book 'Algorithms to Live By' by Christian and Griffiths for its application of computer science concepts — like the 37% stopping rule for hiring and the explore/exploit tradeoff — to real-world business decisions.
Key Insights
- Bruno argues that forecasting accuracy directly determines how fast a company can grow — a 10% MAPE error forced conservative cash allocation, while reducing it to under 2% via AI enabled WellHub to reinvest more aggressively and accelerate from 40% to faster year-over-year growth.
- Bruno claims that a single engineer using Claude Code in two weeks replicated and exceeded work that previously required a large team, external vendors, hundreds of thousands of dollars, and nearly a year — adding six new forecasting models across every country, segment, and cohort.
- Bruno describes using the Windy weather app's multi-model divergence display as the conceptual framework for WellHub's forecasting: when models converge, he banks the number; when models diverge, he treats the spread as a signal of uncertainty warranting conservative planning.
- Bruno argues that the most underappreciated capability of AI in finance is not speed but resiliency — AI-assisted pipelines can detect when upstream inputs change and self-adjust, whereas manually built workflows break silently and require human intervention.
- Bruno contends that 96% of healthcare spend goes toward treatment while only 4% goes toward prevention, despite 75% of costs being preventable — and that CFOs who implement WellHub consistently have an 'aha moment' realizing they should have treated wellness as an investment, not an expense.
- Bruno describes WellHub's partner value proposition as pure incrementality: nine out of ten people sent to gym/studio partners are net-new customers who had never visited before, meaning partners receive zero-CAC revenue that flows almost entirely to EBITDA.
- Bruno argues that the most powerful single metric for evaluating a business is the ratio of gross profit dollars added versus expense dollars added — because it strips away revenue structure noise and reveals true operating leverage, which he considers rarer and harder to fake than top-line growth.
- Bruno claims the '37% stopping rule' from computer science — which states you should hire the next best candidate after evaluating 37% of the available pool — provides an algorithmically optimal answer to the hiring timing problem that most business leaders never consider.
Topics
Full transcript available for MurmurCast members
Sign Up to Access