What Held Back AI Adoption at 16 Companies

Employee resistance to AI was the main barrier at 5 of 16 companies, more than any other. Poor data quality came next at 4, followed by gaps in workflow ownership at 3.

Study methodology: Experts responded via a media response platform, and 16 companies comprised the final sample. Each company named a single main barrier, and the counts below apply the published shares to 16 companies.

What held back AI adoption at 16 companies

Essential statistics: employee resistance to AI and other barriers

• Low employee trust or fear of being replaced was the main barrier in 5 of 16 companies (31%).
• Poor data quality and weak source material came second, holding back 4 companies (25%).
• 3 companies (19%) named gaps in workflow redesign and ownership.
• Legacy integration and scattered systems were the main barrier for 2 (13%).
• Brand or customer-experience risk held back 1 company, and security or compliance risk held back 1 more.
• Barriers rooted in people and ownership, low trust plus workflow gaps, account for 8 of the 16 companies (50%).
• Data and systems barriers account for the next 6 (38%).

Key takeaways

• Employee resistance to AI showed up where the tool looked like a stand-in for staff judgment. It outranked every technical barrier in the study.
• Half of the main barriers came down to people and ownership, with low trust and unclear ownership holding back 8 of 16 companies.
• Data quality ranked second at 4 of 16, which puts content upkeep, from help-center articles to ticket tags, on the critical path for adoption.
• Only 2 of 16 companies named a risk concern, brand or security, as the main barrier. Employee trust held back 5, so internal acceptance was a bigger hurdle than external risk did.
• Unclear workflow ownership held back 3 companies, compared with 2 for legacy integration, so process gaps outweighed old systems in this sample.

Actionable insights

• Run new AI in shadow mode beside the team before go-live, since employee resistance to AI was the main barrier at 5 of 16 companies. The study pairs that barrier with side-by-side testing, which gives staff evidence to weigh.
• Give staff override rights from the first day. Companies in this study that left the final call with people built trust faster.
• Audit the knowledge base and ticket tags before rollout, since poor data quality and weak source material held back 4 of 16 companies.
• Name a single owner for each AI workflow, as ownership gaps held back 3 companies, more than legacy integration did.
• Connect legacy systems through middleware or API bridges before the rollout, since they were the main barrier at 2 of 16 companies.

“Teams pushed back when AI looked like a verdict on their own judgment. Adoption picked up once people could compare its output with their work and overrule it when they disagreed.” — Katerina Merzlova, Digital Transformation Consultant, SumatoSoft