Float CEO Rob Khazzam on Rethinking Corporate AI Spending Strategies
Canadian businesses must treat AI costs as human capital investments rather than software expenses.
1Uncontrolled AI Spending Risks
Float CEO Rob Khazzam recently discovered his own daily AI usage cost reached $2,364 due to default model settings. This incident highlights a broader trend among Float’s 10,000 customers, where AI spending grew twelve-fold over two years. Seventy percent of these companies experienced a 50 percent monthly spike in AI-related bills.
2Redefining Human Capital Investment
Khazzam suggests moving away from viewing AI as a variable software expense. Instead, he advocates combining token spend with payroll data to evaluate total human capital investment. At Float, this combined cost rose from 0.1 percent to six percent of payroll, reflecting a strategic shift toward measuring long-term organizational productivity gains.
3AI-First Hiring Evaluation
Before approving new hires, Float managers must now explain how AI tools can support existing team workflows. While the company continues to hire, this mandatory pause ensures leaders reconsider resource allocation. For example, internal tools like Surf have fundamentally changed the technical requirements and skill sets needed for their design team.
4Optimizing Model Selection Costs
Not every task requires expensive frontier models, which often provide marginal benefits for routine work like email editing. Khazzam recommends granting model discretion primarily to data scientists and principal engineers. Companies should implement strict usage thresholds and assign specific personnel to monitor spending spikes and investigate the underlying causes of costs.
5Consolidating Fragmented AI Tooling
Initial experimentation led to inefficiencies, with twenty employees using five different tools to solve identical problems. Float eventually consolidated its software stack to improve collaboration and data security. Standardizing tools allows the company to provide shared business context, ensuring AI models understand specific organizational problems rather than just performing generic tasks.
6Measuring Long-Term Learning Returns
Leaders should avoid cutting AI budgets simply because revenue per employee has not immediately increased. Khazzam emphasizes that learning takes time to manifest in financial results. He is comfortable with temporary overspending if it maximizes the total hours employees spend developing their capabilities and mastering new AI-driven workflows across the company.
7Hackathons as Progress Indicators
Float utilizes company-wide hackathons to assess how effectively teams are integrating AI into their daily operations. These events involve non-technical departments like marketing and sales, encouraging the rapid development of new workflows. Observing employees ship functional solutions in short timeframes provides leadership with the confidence to continue supporting ongoing AI experimentation.
8Strategic Financial Outlook
The core challenge for modern businesses is balancing rapid adoption with fiscal responsibility. By treating AI as a core investment rather than a utility, companies can foster faster problem-solving and product development. Organizations that prioritize employee learning today will likely gain a competitive advantage as their internal AI capabilities compound over time.
Original article: BetaKit — Float CEO Rob Khazzam on Rethinking Corporate AI Spending Strategies