FTL Survey of 181 HVAC Technicians: Willingness Is High, Net Return Is Zero — the Tools Are the Problem

A survey of 181 commercial HVAC service technicians conducted by FTL (Field Technologies) between 6 and 20 April 2026 found that the widely held assumption of technician resistance to digital technology is not supported by the data. Willingness is high and uniform across experience levels, but the typical technician reports that technology has helped as much as it has hurt — a net return of zero.

The adoption gap

The study was designed to separate four questions that industry discussion routinely conflates: whether technicians are willing to use technology, whether the conditions for use exist, whether the tools help or hurt productivity, and which kind of technology works best. The results point to what FTL calls an "adoption gap" — the distance between demonstrated willingness and realized benefit.

On the willingness side, the numbers are unambiguous: 78.5% say digital tools make their job easier, 96.1% want more training (73.9% unconditionally, 22.2% only on paid work time), and fewer than 3% say digital tools do not help. Comfort with new technology stands at 66.9%, with only 3.9% describing themselves as uncomfortable. Technicians also report that technology has reduced troubleshooting time (81.7%), improved service quality (68.0%), improved job safety (80.7%), and made communication easier (64.1%).

On the benefit side, technicians credit 41% of the technology they have encountered with helping and blame 31% with hurting. When the two are subtracted for each individual, a third come out ahead, fewer than one in ten come out behind, and the remaining 58% see no net benefit or harm. For the typical technician in this survey, technology has given back precisely as much as it has taken away.

What actually helps: problem-solving, not coordination

The study tested four hypotheses. The tenure hypothesis (H1) — that willingness declines with experience — was rejected. No downward trend appears in either willingness or comfort across the four experience groups. Three quarters of technicians with more than ten years in the trade say digital tools make their job easier, against 85% of those with one to three years; the 3-to-7-year group is the least enthusiastic at 73%. Formally, there is no association between experience and willingness (chi-square = 4.06, 6 d.f., p = 0.67). Comfort runs the other way: 68% of the most experienced group considers themselves comfortable or very comfortable, against 55% of the newest (chi-square = 9.59, 9 d.f., p = 0.38). Nearly half of technicians with more than 10 years say technology now demands more technical and digital skill from them, against a quarter of the newest group.

The information hypothesis (H2) was supported. Technicians who arrive informed on three quarters or more of calls judge technology as helpful 55% of the time, versus 30% among those who arrive informed on a quarter or less. This is the most statistically robust relationship in the dataset (rho = 0.361, p < 0.000001, n = 179). In a logistic regression (n = 179, 2,000 bootstrap resamples), arriving with job information multiplies the odds that more than half of the technology used is useful by 2.08 (95% CI: 1.46–3.48), and technology that cuts diagnostic time multiplies those odds by 1.90 (95% CI: 1.21–3.45). Both effects hold with a response-style control in place (odds × 2.38, 95% CI: 1.53–4.27), so neither is an artifact of scale use.

The connectivity hypothesis (H3) was rejected as stated. Network availability has essentially no relationship with willingness to adopt (rho = 0.014, p = 0.85) and only a borderline one with realized benefit (rho = 0.142, p = 0.058). The study reformulates the finding: what matters is not how often a technician has a connection, but what the tool does during the time when they do not. Inaccessibility without a connection is the most-cited failure mode at 53.7%.

The problem-solving hypothesis (H4) was partially supported. Arriving informed and shortened diagnostic time are both valuable outcomes. Within the model, the effect of arriving informed is 2.13 times the effect of easier coordination, and that difference is statistically significant (p = 0.018). However, the evidence on coordination technology is inconclusive; the study calls for further research with a larger sample. An optimism test confirmed that shortened diagnostic time is the only factor that moves the helpful and harmful sides in opposite directions, making it the clearest evidence that the effect is real rather than sentiment.

Field conditions: information is the scarcer resource

Only 23% of technicians arrive knowing the information required for the job on three quarters or more of their calls; 38% have that information on half their calls or fewer. The sample average is 55%. Connectivity is better but uneven: 46% have service on three quarters or more of calls, with a sample average of 67%. The two conditions fail largely independently (rho = 0.245, p < 0.001), and a technician needs both for their technology to work properly. Only 26 of 180 technicians (about one in seven) reliably have both information and connectivity.

The "undecided" group is not resistant

Thirty-four technicians (18.8% of the sample) answered "not sure" when asked whether digital tools make their job easier. None of the 34 describes themselves as uncomfortable with technology, and more than nine in ten want additional training. They are marginally more senior than the affirmative group (56% over 10 years versus 46%) but not significantly so, and their access to connectivity is statistically indistinguishable (p = 0.59). What separates them is what they have been given: they report technology as useful 31% of the time against 44% for the affirmative group (p = 0.002), and their information availability is 9% lower (p = 0.055). The study notes that treating this group as resistant writes off the most persuadable population in the sample.

Industry context

Charles Mosley, an HVAC field technician at BME Services (which provides mechanical, electrical, plumbing, and HVAC services for commercial and industrial buildings across the Midwest), stated that he embraces all new technology and is open to learning new things. Ben Armstrong, Executive Director and Research Scientist at MIT's Industrial Performance Center, where he co-leads the Work of the Future initiative, drew a distinction: in his experience, workers are not scared of technology but are skeptical of management, and there is a difference. Jeff Hubler, President of BME Services, acknowledged that technology rolled out in the industry that has not worked has made technicians' jobs harder, so they immediately resist.

Recommendations from FTL

Based on the findings, FTL recommends the following to service providers and field technology developers:

• Prioritize technology that helps technicians solve problems. Arriving informed is the strongest predictor of realized benefit, and only 23% of technicians have it on three quarters or more of calls. Technology that cuts diagnostic time nearly doubles the odds a technician reports benefit.
• Use diagnostic time as the acceptance test for new tools rather than login counts, which measure compliance.
• Ensure tools remain functional when network connectivity is lost, either through improved offline capability or by moving diagnostic assessments forward so insights reach the technician before arrival on site.
• Design for the technician, not around them. Only 21% of respondents think the tools they use feel built for them.
• Fund training on paid time. Demand is 96.1% in total but 73.9% unconditional; training on paid time makes acceptance virtually unanimous.

Methodology and limitations

The survey instrument comprised 17 closed-form questions administered to working HVAC technicians. Responses were anonymous, with a median completion time of 4.1 minutes. The four field-condition and outcome questions used ordered bands (25% or less, 25–50%, 50–75%, 75–100%) rather than open percentages, on the judgment that a technician can place a frequency in a quartile more reliably than name a specific percentage. The sample of 181 respondents intentionally skews experienced and commercial: 47.5% have more than 10 years in the trade, 70.2% service industrial properties, and 70.2% use mobile work order apps most often. Property percentages exceed 100 because the question permitted multiple answers; 44% of respondents selected more than one.

Statistical methods included Spearman rank correlation, chi-square tests, Mann-Whitney U tests, and logistic regression with confidence intervals from 2,000 bootstrap resamples. A result was treated as significant at less than a 5% chance of occurring by chance. No answers were imputed; the base varies between 162 and 181 by question. The share of technology that helped and the share that hurt are positively correlated (rho = 0.398), not negatively as one might expect, so a control for response style was carried in the model.

The study identifies four limitations: the sample is a convenience sample of commercial technicians not weighted to the national HVAC workforce; all measures are self-reported, including the field conditions; the banded questions sacrifice precision for accuracy; and this is a single snapshot rather than a before-and-after study. FTL suggests four directions for further work: a second survey around a specific technology deployment, tracking measurable data instead of subjective survey results, weighting to HVAC workforce demographics, and using a larger sample with three or four questions per hypothesis.

The paper was authored by Clayton K. Callander, President of FTL (Field Technologies).