BestEx Research

BestEx Research is a research-led specialist in the cost of trading. Its algorithms cover US, Canadian, 15 European and 4 Asia-Pacific equity markets and futures on 21 exchanges, and it reports $893bn traded in 2024, its first profitable year. Scotiabank runs a branded Canadian algorithm platform on it. In 2026 it launched a dark algorithm (Curator), AMS One for banks and brokers, Pulse AI through clients' own assistants, and a US equities cost model validated out of sample.

The evidence is unusually quantitative but company-run, and enterprise documentation is thin. IS Zero cut implementation shortfall 30% against VWAP in a 140,000-order A/B test and one asset manager saved $21m (both company-reported). We did not locate SOC 2 or ISO 27001, a data-handling or no-training position for Pulse AI, published prices, or the firm's regulatory registrations. Disclosed funding is $15m; no round since March 2023.

We initiate at 6.51, Established, with an Improving outlook. Product Capability (7.25) and AI & Technology (7.17) lead, with Data & Content at 6.75 on execution data depth. Workflow Fit (6.58) reflects an execution rather than portfolio-construction focus, and Enterprise Readiness (5.07) and Commercial Value (5.33) are the gaps.

Strengths

• Research depth. Execution models designed and documented in white papers, with out-of-sample validation on a full year of executions.
• Live A/B evidence. IS Zero tested against VWAP across more than 140,000 parent orders (company).
• Configurability. Strategy Studio exposes 150+ parameters without code; banks can build branded algorithms on AMS One.
• Asset-class reach. Global equities and futures on 21 exchanges, with 24×5 futures trading since April 2025.
• Shipping velocity. Four launches in 2026, from dark execution to AI-assistant access.
• Profitability. First profitable year in 2024 and volume up 87% (company).

Cautions

• Assurance. SOC 2, ISO 27001 and SSO documentation not located.
• AI data handling. Pulse AI runs through third-party assistants such as Claude and ChatGPT; no data-handling or no-training terms published.
• Evidence. Cost-saving and A/B results are company-run; no independent evaluation located.
• Scope. Execution and cost analytics only; no optimiser, factor model or portfolio risk.
• Pricing. No prices or pricing model published.
• Capital. Funding last disclosed in 2023; $15m in total.
• Regulatory status. The firm's own registrations were not located.