Why SMBs struggle with AI, but how they can also win

·35 min·3 clips

Ed Kiesling reveals why a $10 million OpenAI deal creates a 'haves and have-nots' divide for small businesses.

The episode begins by framing the disparity in AI adoption between large enterprises and small to mid-sized businesses. Guest Ed Kiesling describes a 'haves and have-nots' scenario, citing an example where a $10 million spend with OpenAI grants access to dedicated engineering support, which is out of reach for most SMBs. He notes that while there is high appetite for AI across all company sizes, many organizations rush to 'AI-ify' everything without first understanding the problem to be solved. A key mistake identified is attempting to automate processes that are not documented and exist only in employees' heads. Kiesling discusses the build versus buy decision, advising SMBs to focus on their core competencies and develop partnerships with trusted vendors rather than trying to do everything in-house. On data challenges, he argues the bigger issue is not data quality but the lack of process documentation before automation. Vendor selection is presented as a balancing act between consolidation benefits and the flexibility of working with smaller, more responsive partners. Risk governance for AI projects is framed as needing the same rigor as traditional software, with additional considerations for increased attack surfaces, bias, and unauthorized access. Kiesling warns against the pressure to cut corners, emphasizing that the final 20% of work—security, observability, scaling—is critical and cannot be ignored. For quick wins, he suggests focusing on enabling employees with research tools like Notebook LM and Deep Research to accelerate learning. He advocates for a 'crawl, walk, run' approach, where employees first use AI tools personally to understand their capabilities and limitations. Agentic RAG solutions are highlighted as a practical starting point to organize unstructured data and build internal knowledge bases. The discussion touches on workplace culture, with Kiesling arguing AI should be used for innovation and expansion, not just efficiency and job replacement. Leadership must clearly articulate the 'burning platform' reason for AI adoption to employees. Looking forward, Kiesling expresses concern about the knowledge gap and the potential widening of the 'haves and have-nots' divide. However, he is optimistic that agentic AI, particularly in coding and business process automation, will mature and deliver real outcomes soon. He concludes that SMBs have a unique advantage due to their agility, personal customer relationships, and blended employee roles, which are well-suited for AI augmentation.

As heard by us

A practical look at how SMBs can test AI while keeping security, access, and worst-case risk in view.

Small and mid-sized businesses face an AI challenge shaped as much by capacity as ambition. Enterprise budgets can put forward-deployed engineers within reach, while smaller teams have to judge what they can attempt with fewer resources.

Read the full review in PlayNext →

Why you'd press play

You need an AI starting point for a smaller business without pretending enterprise budgets or enterprise readiness are already on the table.

Read the full recommendation in PlayNext →

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from Today in Tech