How likely are members of your current team to leave their jobs in the next 12 months?

Every rating is read on the goodness view, the reversed scale on which a higher figure is the better outcome. The mean rating was 4.3 and the median 4.0 (n = 799 respondents). Ratings of 3 or below on the goodness view, the distressed range, made up 46.06% of the answers. The rating was associated with all four study measures. Higher ratings on the goodness view came with less cancellation (r = −0.81, 95% CI −0.86 to −0.75, p < 0.001, n = 150 engagements), higher customer satisfaction among active engagements (r = 0.69, 95% CI 0.56 to 0.78, p < 0.001, n = 94 engagements), a higher engagement indicator mean (r = 0.86, 95% CI 0.81 to 0.90, p < 0.001, n = 150 engagements) and lower intent to leave (r = −0.76, cluster bootstrap 95% CI −0.81 to −0.71, p < 0.001 with standard errors clustered by engagement, n = 749 respondents). After Holm correction, ratings differed by engagement status and lifecycle stage (Welch F 150.35, Holm p < 0.001, omega squared 0.656) and vendor type (Welch F 8.19, Holm p < 0.001, omega squared 0.039).

Every one of the 50 written responses outside C030 described team members who had left, were looking for other work, had asked to move, or a queue that had turned over. In 13 responses, the knowledge of the manual work left with them and was mostly never written down. The five responses to C030 named a team that held together to the end. One of the four written outcome tests reached significance.

Ratings of 3 or below on the goodness view, the distressed range, accounted for 368 of the 799 answers (46.06%). The engagement mean fell in the distressed range in 63 of 150 engagements.

The primary responses numbered 5, all written in answer to C030, “What are the top 3 things that you love most about your job?” The remaining 50 responses, 90.91% of the total, were secondary and came from 5 other questions. That secondary share ranked 9th from the highest among the 11 Part 4 indicators with 30 or more written responses, against a median of 100.00%. Respondents who wrote both a primary and a secondary response numbered 1. Negative responses made up 0 of the 5 primary responses and 100.00% of secondary responses (50 of 50).

Every one of the 50 responses outside C030 described departures: team members who had left, were looking for other work or had asked to move, often two of a team of four. The respondents gave reasons. Seven tied departures to evenings spent checking or re-keying what the AI tool had produced, so that people did the same job twice. Seven tied them to days idle waiting on the financial institution or a provider followed by late nights, night calls and throttled month-end runs. Six tied them to the absence of training, and six to decisions that no one on either side would make or own. Three described replacements that took weeks to arrive and had not done the work before.

Across 13 responses, most from C198, on the third-party ecosystem, respondents said the people who left took with them the knowledge of the manual work, such as the quirks of each custodian file or the corrections made by hand before the model saw a file, and that most of it had never been written down. Several said no one left could do the work by hand if the agent or a provider failed. The five responses to C030, on what the respondent loves most about the job, named a team that stuck it out or held together to the end, several on engagements that ended.

Of the 55 responses, 50 (90.91%, 95% CI 82.69% to 98.11%) were negative and 3 (5.45%) carried the strongest negative score of −10. Positive responses numbered 5 (9.09%), and no response was neutral. The mean written score was −4.4 (95% CI −5.2 to −3.5, n = 55 responses), and the median was −5.0.

Of the 55 responses, 54 came from respondents who also rated the indicator. Among them, the written score did not track the rating (Spearman rho −0.07, 95% CI −0.26 to 0.11, p = 0.600, n = 54 responses). Of the 49 negative responses with a rating, 39 (79.59%) came from respondents whose rating was in the distressed range.

Respondents who wrote about the indicator and gave a rating numbered 51, with a mean rating of 2.6. The 748 respondents who rated the indicator without writing about it had a mean of 4.4. The difference, −1.8 (95% CI −2.3 to −1.2), excluded zero. Ratings in the distressed range made up 80.39% of the ratings from respondents who wrote (41 of 51) and 43.72% of the ratings from respondents who did not (327 of 748).

The respondents who wrote came disproportionately from the distressed end of the scale. The written evidence therefore gives more weight to the respondents who rated the indicator worst.

Every figure in the segment tables is at respondent level. Segments with fewer than 30 respondents carry no interval. Segments with fewer than 10 report counts only.

Segment families recorded at engagement level were tested on engagement means. Respondent role and vendor type were tested on respondents. Where an expected count fell below 5, the test of the distressed share used Fisher's exact test for a two-by-two table, marked exact, or a Monte Carlo test, marked simulated. After Holm correction, ratings differed by engagement status and lifecycle stage (Welch F 150.35, Holm p < 0.001, omega squared 0.656). Active engagements in Production rated highest (mean 6.3, n = 98 respondents) and canceled engagements lowest (mean 1.8, n = 308 respondents). Ratings also differed by vendor type (Welch F 8.19, Holm p < 0.001, omega squared 0.039). Pure-play agentic AI firms rated highest (mean 4.9, n = 70 respondents) and BPO specialists lowest (mean 3.4, n = 186 respondents). Outsourcing model, institution type, process family, class of work and respondent role showed no difference.

Written sentiment did not differ by engagement status and lifecycle stage across the 44 engagements with a written response (Kruskal-Wallis H 1.83, df 2, p = 0.400). Fewer than two groups held 10 or more responses, so the groups are reported as counts only: active engagements in Build-Out (6), active engagements in Transition/Cut-Over (7), active engagements in Production (1) and canceled engagements (41).

No test across respondent roles was run, because fewer than two roles had 30 or more written responses. Individual contributors wrote 30, engagement managers wrote 21, portfolio directors wrote 3 and customer management executives wrote 1 responses.

Cancellation rate is the share of engagements canceled by the study date, measured at engagement level (150 engagements, 56 canceled). Canceled engagements were rated after cancellation. The association is therefore a descriptive contrast and not a forecast of cancellation.

Across the study set, higher ratings on the goodness view came with less cancellation (r = −0.81, 95% CI −0.86 to −0.75, Holm p < 0.001, n = 150 engagements). The association held after Holm correction in twenty of the 20 segment levels tested.

Customer satisfaction is the percentage recorded against each engagement. The primary estimate uses the 94 active engagements. A descriptive contrast uses all 150 engagements.

Across the active engagements, higher ratings on the goodness view came with higher customer satisfaction (r = 0.69, 95% CI 0.56 to 0.78, Holm p < 0.001, n = 94 engagements). The association held after Holm correction in thirteen of the 16 segment levels tested.

The engagement indicator mean is the mean of an engagement's ratings across the rated indicators, measured at engagement level (150 engagements). The indicator does not enter the mean.

Higher ratings on the goodness view came with a higher engagement indicator mean (r = 0.86, 95% CI 0.81 to 0.90, Holm p < 0.001, n = 150 engagements). The association held after Holm correction in twenty-three of the 24 segment levels tested.

Intent to leave is the answer to C019, “How likely are you to leave your current job in the next 12 months?”, rated 1 to 10. It is measured at respondent level on up to 798 respondents, with the engagement-level form reported alongside.

Respondents who gave higher ratings reported lower intent to leave (r = −0.76, 95% CI −0.79 to −0.73, Holm p < 0.001, n = 749 respondents). The association held after Holm correction in thirty-six of the 37 segment levels tested.

Two comparisons set the engagements with at least one negative response against all other engagements. Of the 109 other engagements, 105 had no written response addressing the indicator. These two comparisons therefore do not isolate written sentiment. Of the 41 engagements with a negative response, 28 (68.29%) were canceled, against 28 of the 109 other engagements (25.69%). The difference reached significance (Fisher exact p < 0.001). Customer satisfaction among active engagements did not differ between the two groups (mean 66.23% on 13 engagements against 63.75% on 81, Welch t 0.69, p = 0.500).

Two correlations test written sentiment directly. Across the 45 engagements with a written response, the engagement mean written score was not associated with cancellation (point-biserial r 0.20, p = 0.187). Across 48 respondents, the respondent mean written score was not associated with intent to leave (Spearman rho 0.19, p = 0.194). Of the four comparisons, one reached significance at 0.05.

The four associations were compared in pairs on their engagement-level forms, each oriented so that a higher value is the better outcome, across the 150 engagements with all four measures. After Holm correction, two of the six pairs differed. The association with the engagement indicator mean was the strongest and the association with customer satisfaction the weakest.

Answers clustered within engagements (ICC 0.84). Clustering reduced the 799 respondent answers to the equivalent of about 172 independent answers. Respondent-level intervals therefore used the cluster bootstrap. Respondent-level tests of intent to leave used standard errors that allow for clustering by engagement.

No association was equivalent to zero. Each of the four differed from zero. The detection floor shown is the study floor at 150 engagements. Intent to leave was tested on its effective sample size.

Each change was computed before rounding and may differ by 0.01 from the difference between the two printed values.

All 22 checks held. A check holds when its test falls on the same side of 0.05 as the primary test. Where the primary test falls below 0.05, the check must also keep the sign of the primary correlation. The thin-base check removes the engagements in any level of outsourcing model or process family with fewer than 30 engagements. Institution type is not used for that check, because every one of its levels has fewer than 30 engagements.

Each of the four measures was computed from data the indicator does not contain. Each result is therefore a finding and not an arithmetic consequence.

Each deviation was computed before rounding and may differ by 0.01 from the difference between the two printed values.

The comparison covered 150 engagements. The indicator departed from the study's typical pattern on cancellation, customer satisfaction and intent to leave. Its association was stronger than the norm on cancellation, customer satisfaction and intent to leave.

Among the 11 Part 4 indicators with 30 or more written responses, the indicator's share of negative responses, 90.91%, ranked 1st from the highest, against a median of 33.33%. Its mean written score, −4.4, ranked 1st from the most negative, against a median of 0.8.

The indicator most often present in the same responses was C018, “How happy are you with your current job?”, in 10 of the 55 responses (18.18%). C020, “How happy is your extended team with the current engagement?” (10 responses), C123, “Rate the adequacy of your employer's investment in training its workforce.” (8 responses), C157, “Rate the factual reliability of the AI tools' outputs on this engagement before any human review.” (6 responses) and C178, “Rate how well the agents on this engagement are protected from breaking silently when a third party changes something without warning.” (6 responses) followed.

The most frequent theme was “Occurs every cycle or most days”, in 6 responses (10.91%). The second was “Workaround is a habit, not a control”, in 6 responses.

Nearly half of all ratings of how likely team members were to leave fell in the distressed range on the goodness view (368 of 799 answers, 46.06%). The rating was associated with all four study measures, and every association held under each of the 22 robustness checks. Canceled engagements rated far below the three active stages (mean 1.8 on the goodness view, n = 308 respondents).

The written evidence describes departures already under way, and it adds a consequence the ratings do not measure: the knowledge of the manual work left with the people, so that the remaining staff could not fall back on it when an agent or a provider failed. Engagements with a negative written response were canceled more often. The other three written outcome tests did not reach significance.

The evidence is the reported experience of the vendor's delivery teams, collected in a single cross-section. The associations reported here are not causal.

Tests by respondent role and by vendor type treated respondents as independent. The section on clustering reports how far clustering within engagements reduced the information in respondent-level figures.

Customer management executives gave 2 answers, a base of anecdote strength.

Shares of written sentiment are shares of responses and not of the study set of respondents. The written score takes five fixed values, so its mean and median are coarse. Responses clustered within engagements (55 responses on 45 engagements). Intervals for written sentiment therefore used a bootstrap by engagement.