Rate how quickly the two firms together can stop an agent that is behaving badly in live operations

The mean rating was 3.6 and the median 3.0 (n = 804 respondents). Ratings of 3 or below, the distressed range, made up 50.50% of the answers. The rating was associated with all four study measures. Higher ratings came with less cancellation (r = −0.63, 95% CI −0.72 to −0.52, p < 0.001, n = 150 engagements), higher customer satisfaction among active engagements (r = 0.60, 95% CI 0.45 to 0.71, p < 0.001, n = 94 engagements), a higher engagement indicator mean computed without the indicator (r = 0.86, 95% CI 0.81 to 0.89, p < 0.001, n = 150 engagements) and lower intent to leave (r = −0.67, cluster bootstrap 95% CI −0.73 to −0.61, p < 0.001 with standard errors clustered by engagement, n = 753 respondents). After Holm correction, ratings differed by engagement status and lifecycle stage (Welch F 41.68, Holm p < 0.001, omega squared 0.389) and respondent role (Welch F 8.28, Holm p = 0.003, omega squared 0.016).

Written responses addressing the indicator numbered 73, from 71 respondents on 51 engagements. Of these, 59 were negative and 13 positive. Nearly half stated that no one knew who could stop a misbehaving agent, or that no one could, and described agents that ran on for hours or days while the two firms worked out who held the switch. Several proposed the same remedy: one named person with authority to pause the agent. Where the firms could stop an agent within minutes, respondents placed the remaining difficulty in deciding afterward who owned the harm. Respondents whose written responses were more negative reported higher intent to leave.

Ratings of 3 or below, the distressed range, accounted for 406 of the 804 answers (50.50%). The engagement mean fell in the distressed range in 62 of 150 engagements.

The primary responses numbered 60, all written in answer to C156, “Add any comments on the joint management of this engagement that the ratings in this section do not capture.” The remaining 13 responses, 17.81% of the total, were secondary and came from 3 other questions. That secondary share ranked 37th from the highest among the 71 Part 2 indicators with 30 or more written responses, against a median of 18.92%. Respondents who wrote both a primary and a secondary response numbered 1. Negative responses made up 80.00% of primary responses (48 of 60) and 84.62% of secondary responses (11 of 13).

All 10 responses to C029 named kill-switch authority among the challenges that reduce job satisfaction, as never assigned, argued over in the middle of an incident, or unresolved after several incidents.

Among the 60 responses to C156, the comment field on joint management, 34 negative responses stated that no one knew who could stop the agent, that no one could, or that stopping it took a ticket, several calls or days. The incidents were specific: a sector flagged in hundreds of alerts that ran the full cycle and was then cleared by hand, a vendor exception posted forty times before the call reached anyone, a password reset agent that closed tickets for four days before anyone at the institution could switch it off, a stale balance rule reported on a Monday with no decision on who would stop it by Friday. Five described the gap outside office hours: no one at the institution able to authorise a stop after six in the evening, overnight, or on a Friday night. Seven asked for the same remedy, one named person on each side or each shift with authority to pause the agent without a meeting. In answer to C055, the question on how the vendor manipulates performance reports, one respondent described escalation response times reported without the incidents delayed by unclear kill-switch authority.

Against these, eleven positive responses said the firms could stop an agent within minutes or within the hour, and that both sides knew whom to call. Several of these named the next gap: deciding whose fault the incident was, or which firm would pay to remediate the two hundred wrong replies sent before the stop.

Of the 73 responses, 59 (80.82%, 95% CI 71.43% to 89.19%) were negative and 40 (54.79%) carried the strongest negative score of −10. Positive responses numbered 13 (17.81%), and 1 was neutral. The mean written score was −5.8 (95% CI −7.1 to −4.3, n = 73 responses), and the median was −10.0.

Of the 73 responses, 71 came from respondents who also rated the indicator. Among them, the written score rose with the rating (Spearman rho 0.59, 95% CI 0.41 to 0.74, p < 0.001, n = 71 responses). The written responses corroborated the rating. Of the 57 negative responses with a rating, 34 (59.65%) came from respondents whose rating was in the distressed range.

Respondents who wrote about the indicator and gave a rating numbered 70, with a mean rating of 3.2. The 734 respondents who rated the indicator without writing about it had a mean of 3.7. The difference, −0.4 (95% CI −1.0 to 0.2), included zero. Ratings in the distressed range made up 55.71% of the ratings from respondents who wrote (39 of 70) and 50.00% of the ratings from respondents who did not (367 of 734).

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 41.68, Holm p < 0.001, omega squared 0.389). Active engagements in Transition/Cut-Over rated highest (mean 4.6, n = 236 respondents) and canceled engagements lowest (mean 2.2, n = 312 respondents). Ratings also differed by respondent role (Welch F 8.28, Holm p = 0.003, omega squared 0.016). Portfolio directors rated highest (mean 4.7, n = 31 respondents) and individual contributors lowest (mean 3.5, n = 621 respondents). Outsourcing model, institution type, process family, class of work and vendor type showed no difference.

Written sentiment differed by engagement status and lifecycle stage across the 51 engagements with a written response (Kruskal-Wallis H 10.59, df 3, p = 0.014). Of the groups with 10 or more responses, responses from canceled engagements were the most negative (mean −7.7, 24 responses) and responses from active engagements in Production the least negative (mean −2.5, 10 responses).

No test across respondent roles was run, because fewer than two roles had 30 or more written responses. Individual contributors wrote 49, engagement managers wrote 20 and portfolio directors wrote 4 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 came with less cancellation (r = −0.63, 95% CI −0.72 to −0.52, Holm p < 0.001, n = 150 engagements). The association held after Holm correction in fifteen 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 came with higher customer satisfaction (r = 0.60, 95% CI 0.45 to 0.71, Holm p < 0.001, n = 94 engagements). The association held after Holm correction in ten 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 is one of the rated indicators. Each comparison therefore uses the mean recomputed without it.

Higher ratings came with a higher engagement indicator mean (r = 0.86, 95% CI 0.81 to 0.89, Holm p < 0.001, n = 150 engagements). The association held after Holm correction in twenty-four 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 = −0.67, 95% CI −0.71 to −0.63, Holm p < 0.001, n = 753 respondents). The association held after Holm correction in thirty-seven of the 37 segment levels tested.

Two comparisons set the engagements with at least one negative response against all other engagements. Of the 107 other engagements, 99 had no written response addressing the indicator. These two comparisons therefore do not isolate written sentiment. Of the 43 engagements with a negative response, 15 (34.88%) were canceled, against 41 of the 107 other engagements (38.32%). The difference did not reach significance (Fisher exact p = 0.714). Customer satisfaction among active engagements did not differ between the two groups (mean 63.61% on 28 engagements against 64.30% on 66, Welch t −0.27, p = 0.791).

Two correlations test written sentiment directly. Across the 51 engagements with a written response, the engagement mean written score was not associated with cancellation (point-biserial r −0.22, p = 0.121). Across 65 respondents, a more negative respondent mean written score came with higher intent to leave (Spearman rho −0.28, p = 0.025). 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, five of the six pairs differed. The association with the engagement indicator mean was the strongest and the association with customer satisfaction the weakest. The associations with customer satisfaction and intent to leave did not differ (Holm p = 0.373).

Answers clustered within engagements (ICC 0.67). Clustering reduced the 804 respondent answers to the equivalent of about 205 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. The indicator is one of the rated indicators, and the engagement indicator mean was recomputed without it. The correction changed r by 0.00, from 0.86 to 0.86.

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 followed the study's typical pattern on all four measures, and no deviation held after Holm correction.

Among the 71 Part 2 indicators with 30 or more written responses, the indicator's share of negative responses, 80.82%, ranked 36th from the highest, against a median of 80.82%. Its mean written score, −5.8, ranked 20th from the most negative, against a median of −4.7.

The indicator most often present in the same responses was C136, “Rate how effectively the customer and vendor jointly manage this engagement now that agents perform part of the work.”, in 19 of the 73 responses (26.03%). C157, “Rate the factual reliability of the AI tools' outputs on this engagement before any human review.” (13 responses), C139, “Rate how clearly accountability lands on one firm when an agent's action causes a problem in this engagement.” (8 responses), C142, “Rate how well the speed of joint issue resolution matches the speed at which agent-caused problems can spread.” (7 responses) and C144, “Rate how well the joint change control process keeps pace with the speed of agent behavior changes.” (7 responses) followed.

The most frequent theme was “Respondent proposes a remedy”, in 10 responses (13.70%). The second was “Raised it, nothing came back”, in 8 responses.

Just over half of all ratings of how quickly the two firms together can stop an agent behaving badly in live operations fell in the distressed range (406 of 804 answers, 50.50%). 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 2.2, n = 312 respondents).

The written responses treat the stop as a matter of named authority. Where a person on each side held the authority and could be reached, the agent was stopped in minutes. Where the authority was unassigned, or rested with the institution's staff who kept office hours, the agent ran on until the question was settled, and the delivery team cleared the output by hand afterward.

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 (73 responses on 51 engagements). Intervals for written sentiment therefore used a bootstrap by engagement.