Goldman Sachs survey shows 86 per cent of small firms use AI, but deep use lags

A study of 2,473 business owners by the Goldman Sachs 10,000 Small Businesses program shows that artificial intelligence adoption among smaller enterprises reached 86 per cent in 2026. That is an increase from 56 per cent two years ago, according to the research conducted with Babson College.

While use has spread, the research found a distinct gap between trying software and running a company on it. Only 15 per cent of surveyed business owners report that AI is fully integrated into their core daily operations. Most owners still treat the technology as a series of isolated experiments or single-task aids.

Shifts toward core operations

The way owners apply these tools is changing. Instead of focusing mostly on copy generation, operators are directing software toward basic business tasks.

Key changes recorded in the survey include:

• The share of AI users applying tools to financial analysis and business planning grew from 18 per cent in 2024 to 42 per cent in 2026.
• Use for operational data analysis reached 42 per cent, up from 26 per cent two years prior.
• The proportion of businesses using AI to write marketing copy fell from 74 per cent down to 65 per cent.
• Three-quarters of owners reported that their companies still lack formal internal rules governing workplace AI use.

To address this gap between surface trials and deeper adoption, Goldman Sachs committed 500 million dollars toward expanding its entrepreneur training, setting up a dedicated Small Business AI Institute alongside Anthropic, Google, Meta and OpenAI. The bank said smaller firms do not need the largest frontier computing models to see real benefits, but need practical guidance on daily workflows.

What this means for Australian businesses

For Australian small and medium enterprises, the study points to where commercial value is moving.

• Drafting social media posts is an easy entry point, but bigger savings come from feeding bookkeeping records, supplier spreadsheets and sales figures into analytical tools to spot cash flow trends.
• Rather than asking staff to use chatbots haphazardly, identify a single repeated manual job, such as invoice reconciliation or quoting, and connect automation directly into existing software.
• Because three in four small firms operate without guidelines, setting clear standards on customer privacy and sensitive data entry protects your business while giving staff confidence to experiment.