Get tomorrow's brief in your inbox
Today: Solo business owners are reporting meaningful revenue and avoided outsourcing costs from AI, but those results are not a guarantee of profit. Dropbox is rolling out approval-based file automation, while recruiters report faster screening and face questions about their fees. Security incidents underline the need to control what connected AI tools can access.
AI's strongest small-business pitch is doing work you otherwise could not afford.
The more useful example is an underdeveloped sales channel. Willen says he used ChatGPT and Claude to rebuild hunting-gear brand Lone Trail's Shopify site and AI to help run Meta advertising; Shopify's share of the brand's revenue rose from roughly 1% to 30% in a recent 30-day period. He also reports avoiding traditional photography shoots costing $2,000 to $5,000 and listing-image design costing $300 to $1,000, although those are his quoted outside-service costs, not universal savings.
What to do: Pick one neglected revenue channel or recurring outsourced task, then test AI against your current approach. Record software spending, advertising costs, your review time, errors, and resulting sales before calling it a win. Preliminary research cited by CNBC links rising business applications without near-term hiring indicators to AI-exposed industries, but it does not establish that AI caused the increase or that those businesses stayed solo.
Who should try it: Agencies, consultants, and other teams repeatedly creating similar project structures. Rollout has begun for more than 70,000 teams on Standard and Advanced plans in the US, UK, and Australia, with broader availability planned for 2027; check your account's eligibility and confirm current pricing before buying or upgrading. Start with a copied, low-risk folder and review the proposed changes, because the opportunity is less administrative setup, not permission to reorganize client records unattended.
What to do: Before your next agency engagement, ask which tasks are automated and what the fee buys beyond candidate discovery, such as assessment, specialist knowledge, and relationship management. Recruiters disagree about whether automation will reduce fees, so treat this as a negotiation opportunity rather than evidence that prices have already fallen; the business benefit is potentially faster hiring with a clearer account of the value you purchase.
What to do: Write down the input, steps, owner, approval point, and expected output for one repetitive job, then measure turnaround and rework before and after the pilot. This makes it easier to identify a genuine improvement and avoids speeding up one task while leaving the wider bottleneck untouched.
What to do: If you use these services or connected developer tools, review provider incident notices and establish whether your accounts, credentials, or plugins were implicated before deciding on remediation. These accounts describe different scopes and do not establish that every customer was compromised; the practical concern is how one connected service could expose another.
What to do: If you run tutoring, training, or other services for minors, verify age eligibility, administrative controls, data handling, and human escalation procedures before introducing a chatbot. Ask vendors to demonstrate the safeguards rather than relying on broad safety assurances; preventing an unsuitable deployment protects students and reduces operational and reputational risk.
What to do: If your business depends on substantial cloud or AI capacity, ask suppliers about regional availability and contingency arrangements before a major commitment. For ordinary small-business subscriptions, this is something to monitor, not a reason by itself to switch providers.
What to do: Evaluate available tools against your own work rather than waiting for internal research systems. The opportunity is to keep improving your workflow as usable capabilities arrive, without committing to benefits you cannot yet test.
Create an approval-first automation specification. Try: "Map this task into inputs, steps, responsible person, approval gates, and expected outputs. Identify one low-risk step to automate. Flag exceptions requiring human review, and propose a way to measure time saved and errors." Use a sanitized task description and treat the result as a draft, following the workflow-first approach.
Run an avoided-cost test. Choose one recurring task you currently outsource, time an AI-assisted attempt, and include checking and correction time in the comparison. Use the solo-business examples as inspiration, not a savings forecast; continue only if the output meets your quality standard.
The strongest opportunities here are narrow and measurable: organizing project files, reducing recruitment administration, and testing neglected sales channels. This week, watch actual account-level feature availability and relevant security notices, while judging every pilot on net savings or additional margin after human review.