Boosted.ai

Boosted.ai brings the longest machine-learning heritage among buy-side agent vendors. Founded in 2017 by a co-founder of Maluuba and former Bloomberg ML engineer, it moved from quantitative tools to Alfa, an agent platform that monitors prices, filings, news and earnings, and Alfa Prime, a multi-model investment committee. Liontrust, ChinaAMC and SS&C Eze are among named users, and Questrade runs its AI assistant on Boosted.

Certification claims are strong; capital and data disclosure are not. Boosted.ai states SOC 2 Type II and ISO 27001, 27701 and 42001, a no-training commitment and an option to deploy its model in the client's AWS environment. Its last equity was a $15m bridge in November 2024, it had no open roles at our review, and its data vendors and model providers are not disclosed.

We initiate at 6.69, Established, with a Stable outlook. Boosted.ai sits in the Emerging quadrant, just above the Emerging band. Workflow Fit, at 7.17, is its strongest dimension; Data & Content, at 5.42, its weakest. Disclosure of data licences and new capital would move the score most.

Strengths

• Domain tenure. Machine learning for investment managers since 2017; the CEO co-founded Maluuba, later acquired by Microsoft, and spent four years in ML at Bloomberg.
• Monitoring agents. Price, filing, news and earnings monitors with minute-level updates, plus email inbox monitoring.
• Channel breadth. Self-serve Individual and Pro tiers, institutional Prime, white-label Alfa Powered at Questrade and BX Partners, and an API with MCP.
• Certification claims. SOC 2 Type II and ISO 27001, 27701 and 42001 stated, with a no-training commitment.
• Cost-efficient AI. A fine-tuned capital-markets model that cut AI cost by 90% and can be deployed in the client's AWS VPC.
• Named clients. Liontrust, ChinaAMC, SS&C Eze and Questrade, with RBC as investor and client.

Cautions

• Funding runway. No equity since a $15m bridge in November 2024, alongside growth capital from CIBC Innovation Banking; no open roles at our review.
• Traction figures. Client counts of 300+ and "180 of the largest asset managers" and AUM figures of $3trn and $5trn appear within months of each other; ARR is not disclosed.
• Performance evidence. The Alfa Prime result is an internal hypothetical backtest; no independent evaluation was located.
• Data provenance. Data licences are not disclosed, and broker and expert content were not located.
• Security paperwork. The security page says SOC 2 Type II audited and the trust centre says SOC 2 compliant; SSO/SCIM, zero data retention and MNPI controls were not located.
• Model supply chain. Model providers are not named.
• Office workflow. Native Excel, Office or Teams integration was not located.