The AI Implementation Friction Field Study
A field study of what stops companies from getting AI out of the test phase and into daily use: technical, cost, people, and process problems.
- First published round
- 60
- Target number of interviews in the first round of published results Evidence class: Proposed.
- Adoption stages
- 6
- Six planned stages, from first interest to daily use with someone watching it Evidence class: Proposed.
- Problem categories
- 12+1
- Twelve planned categories plus one for anything else Evidence class: Proposed.
- Record-review target
- 10%
- Target share of interview records checked for quality Evidence class: Proposed.
Interviews are underway. These figures describe the study design; results are not published yet.
What the study is designed to answer
Where do established companies lose confidence, momentum, or a clear view of the cost when they try to move AI from a test into work people rely on every day?
- 01
What goes wrong before an AI project reaches daily use?
- 02
How do owners and the people doing the work describe the same problem differently?
- 03
Which workflows are worth money but lack an owner, clean data, or connected systems?
- 04
What does it take for people to trust an AI-assisted workflow enough to keep using it?
- 05
Which 90-day measures matter to the people paying for the project?
How we plan to collect data
- 01
Interviews with people doing the work and the managers who own it, using the same question guide each time.
- 02
Research notes are kept separate from any sales follow-up.
- 03
Numbers people give us are labeled as reported, not as something we measured.
- 04
Each interview sorted by adoption stage and problem type, with totals shown next to percentages.
- 05
Results are anonymous by default; naming anyone needs a separate written release.
Checks before analysis
- 01
Interviewers practice with an observer before interviews start
- 02
Early records reviewed before more interviews count toward the target
- 03
Two people review the same sample separately and settle differences
- 04
Each record notes where the information came from
- 05
We publish the limits, how people were recruited, and the totals
What we will publish if the study goes ahead
01
Summary of who we talked to and how
02
How many companies are at each adoption stage
03
Problem counts by type of operation
04
Types of work with the most reported problems
05
What makes people trust it, and 90-day success measures
06
Anonymous notes from the interviews and limits of the sample
Taking part in research is separate from paid work.
Reading a study does not put you on a sales list. If a problem is costing your shop real money, you can request a free consultation directly.
Free consultation