Exploring the Role of Social Support When Integrating Generative AI in Small Business Workflows
Quentin Romero Lauro, Jeffrey P. Bigham, Yasmine Kotturi · 2024 · Companion Publication of the 2024 Conference on Computer-Supported Cooperative Work and Social Computing (CSCW Companion '24) · doi:10.1145/3678884.3681895
Summary
This paper investigates how small business owners from resource-constrained communities leverage their offline social networks to integrate generative AI technologies into their business workflows. The researchers interviewed 11 participants — seven entrepreneurs and four support personnel — recruited from two entrepreneurial hubs in Pittsburgh focused on racial and gender equity. The entrepreneurs ran diverse businesses including a streetwear clothing brand, gift baskets, dance classes, letterpress printing, and art sales. Using semi-structured interviews combined with paper storyboard probes depicting future scenarios, the study examined how entrepreneurs discovered, adopted, maintained, and sometimes refused to use tools like ChatGPT, DALL-E, Canva text-to-image, and Gemini. The research builds on HCI scholarship emphasising the role of small, offline "collectives" — comprising peers, mentors, friends, and loyal customers — in supporting technology adoption for entrepreneurs who lack access to formal infrastructure and training resources.
Key findings
Entrepreneurs used their social networks in four key ways: (1) discovering new use cases by sharing accounts — one entrepreneur shared a ChatGPT Premium account with six business partners, calling it "Group Work for AI" and learning from others' prompt successes and failures; (2) building self-efficacy through side-by-side support — a gift basket entrepreneur described her first AI experience as "horrifying" but overcame this anxiety by working alongside a technical provider; (3) receiving wrap-around support through intermediaries — one entrepreneur's spouse acted as a translator of her needs, brainstorming use cases and creating initial prompts; (4) establishing boundaries of use informed by network attitudes. However, these social configurations also created tensions: social comparison and feelings of being "left behind" when seeing others' AI proficiency, reputational concerns about being seen as needing AI help (one entrepreneur worried peers would think she wasn't "smart enough" to write her own artist statement), and unclear social norms around sharing prompts and accounts. Support personnel felt urgency about the AI "divide," with one stating they tell everyone they work with that they "really need to get on board because they will be left behind."
Relevance
While not directly an accessibility paper, this research has important implications for digital inclusion and the equitable adoption of AI technologies. The entrepreneurs studied — primarily from communities facing racial and gender inequity — mirror many of the barriers that people with disabilities face when adopting new technologies: limited access to training, anxiety about unfamiliar tools, reliance on informal support networks, and the need for intermediaries who can translate between user needs and technology capabilities. The finding that generative AI platforms are designed for single users but are being appropriated for collaborative use highlights a design gap relevant to any marginalised group navigating new technology. For accessibility practitioners, the study reinforces that technology adoption is fundamentally social — people learn from trusted peers, not documentation — and that support programmes should leverage existing community networks rather than assuming individual self-service. The concept of "coopetition" (simultaneous cooperation and competition) among entrepreneurs using shared AI tools has parallels in disability communities where peer support coexists with competition for limited resources.
Tags: generative AI · entrepreneurship · social support · technology adoption · digital divide · digital literacy · lean economies · human-computer interaction