Updates on the latest developments in AI as part of the SMB strategy.
The Point
Agent deployment is advancing faster than testing, security, inventory, rollback, and cost controls.
Why It Matters
Before giving an agent access to business systems, require an inventory, permission limits, release evaluations, spending thresholds, audit logs, and a tested shutdown process.
The Point
An AI agent needs its own identity and limited permissions rather than unrestricted access through an employee’s account.
Why It Matters
Shared credentials make it difficult to determine whether a person or agent accessed data, changed a record, sent a message, or approved a transaction.
What You Can Do
Require every production agent to use a separate account with least-privilege access, activity logs, and human approval for sensitive or irreversible actions.
The Point
Companies are separating into shallow adopters and firms that connect agents to tools, context, and repeatable workflows.
Why It Matters
Measure whether employees are turning useful individual experiments into governed, reusable workflows instead of treating licenses or logins as evidence of adoption.
The Point
The flagship report combines surveys of more than 34,000 small and midsize business owners with anonymized data from more than 5.3 million QuickBooks businesses. It found that 77% of U.S. businesses used AI regularly, 78% reported improved productivity and 43% reported increased revenue, while roughly eight in ten businesses that paid for AI in 2024 were still paying in 2025.
The Point
Effective AI use is associated with explicit task allocation, human quality control, and deliberate preservation of critical-thinking skills.
Why It Matters
Train employees to classify tasks before using AI: delegate routine execution and synthesis, but keep intent, quality standards, exceptions, judgment, and outcome accountability with people.
The Point
Goldman Sachs Research's AI adoption tracker put U.S. business AI adoption at 19.8% as of April 2026, alongside continued growth in AI-related investment now standing at roughly $360bn (1.1% of GDP) in the national accounts.