Imagine a woman entrepreneur running a small restaurant. She may be running more than one small business, or balancing the business alongside another job. Like most women entrepreneurs in low- and middle-income countries, she is also carrying a substantial share of caregiving responsibilities – for children, older relatives, or both.
She began using AI around two years ago, mainly to write Instagram posts, respond to Google reviews and design the menus. It saves her around three hours a week – time she values enormously. Yet her food waste is still unpredictable, her supplier costs keep quietly creeping up, and she has no clear picture of whether she’ll make payroll next month.
AI is saving time — a precious commodity — however much of that time is absorbed by childcare and admin. What it is not doing is helping her grow the business or make it more resilient.
This story echoes findings from our new research: Adopted not embedded: AI, productivity and uneven gains for women entrepreneurs, produced by the Cherie Blair Foundation for Women with Intuit and the World Bank’s Women, Business and the Law project. Across more than 3,000 women entrepreneurs in 66 countries, we found the same pattern: broad adoption, uneven returns depending on the depth of that adoption.
AI at the edges — and what’s being missed
Women entrepreneurs are enthusiastic AI adopters – 82% already use it. They’re using it in consumer-facing functions such as communications, marketing and design. That matters. However, it represents only part of the opportunity. The real growth potential lies in employing AI in the engine room – the back-end functions that determine whether a business can actually scale.
For our restaurant owner, that could mean using AI to analyse long-term sales data, days of week, seasons, weather and local events to predict covers, forecast food prep to cut waste or schedule shifts for staff to cut the staffing costs. It could also mean AI to track supplier invoices, flag when a price has quietly crept up, or to map and plan cash flow to prevent gaps when suppliers or staff need to be paid and avoid costly borrowing.
This is where AI stops being a time-saver and starts being a growth driver. Yet most women entrepreneurs we surveyed are not yet using AI in these operational functions.
In reality, many of small businesses are only partially digitised. Records may be spread across notebooks, WhatsApp messages, paper invoices and spreadsheets. AI tools designed for clean, structured enterprise data often struggle in these environments. For many women entrepreneurs, useful AI needs to work with messy, incomplete and unstructured information rather than assuming fully digitised systems.
The paradox at the heart of the data
Here is the uncomfortable finding: the women who stand to gain the most from deeper AI integration are precisely those with the least capacity to pursue it.
Time savings are the most widely reported benefit of AI use — 69% of respondents say it saves them time. This matters because many women entrepreneurs operate under constant time pressure. Care responsibilities are nearly universal among the women in our survey; only two out of more than 3,000 respondents reported having none.
However, the findings suggest that AI is helping women keep heads above water rather than being a real game changer for their businesses. Integrating AI into operations takes time to experiment, learn and embed new workflows.
What holds them back is time, skill and confidence constraints, alongside the reality that many tools were not designed for a sole trader managing both a business and significant caregiving responsibilities.
Growing concerns about AI bias and safety are compounding the problem. Women entrepreneurs perceive higher downside risk when using AI for higher-stakes functions, so they confine their use to functions where risks are low.
The result is a two-tier AI economy taking shape. Those with capacity to find time, resources and confidence within their businesses are integrating AI deeply and seeing returns. Everyone else is just staying afloat, often using AI primarily to relieve day-to-day operational pressure.
How can we close the gap?
This is not a story about women falling behind. It is a story about a solvable problem if businesses, policymakers and development organisations act with intention. Several practical actions could help close the gap.
Tech companies should design integrated tools that cover the full range of small business operations — not dozens of separate products. Onboarding and in-product guidance must be built for users with limited time, not assumed.
Policymakers should invest in training that goes beyond basic adoption – covering bookkeeping, procurement, cash flow – alongside policies that address the care burdens constraining women’s time.
Funders and development organisations should invest in structured, practical support that helps women entrepreneurs move from consumer-facing functions to operational integration. The confidence gap is real, and it will not close by itself.
AI has the potential to be a genuine leveller for women entrepreneurs in low- and middle-income countries – helping them leapfrog traditional barriers to growth. Women-led SMEs are critical to economic growth; when they are held back by an unequal AI landscape, the result is lost productivity across entire economies.
