Rate how reliably the joint quality assurance loop catches agent errors before they reach live operations

The mean rating was 3.7 and the median 4.0 (n = 787 respondents). Ratings of 3 or below, the distressed range, made up 47.90% of the answers. The rating was associated with all four study measures. Higher ratings came with less cancellation (r = −0.61, 95% CI −0.70 to −0.50, 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 = 0.85, 95% CI 0.79 to 0.89, p < 0.001, n = 150 engagements) and lower intent to leave (r = −0.66, cluster bootstrap 95% CI −0.72 to −0.59, p < 0.001 with standard errors clustered by engagement, n = 742 respondents). After Holm correction, ratings differed by engagement status and lifecycle stage (Welch F 38.66, Holm p < 0.001, omega squared 0.371) and respondent role (Welch F 7.28, Holm p = 0.008, omega squared 0.015).

Few written responses addressed the indicator: 29, nearly all from C156, the comment field on joint management, of which 22 were negative and 7 positive. Negative responses described quality checks that caught typing and formatting errors and missed the errors agents make, or that ran after the output was already live. Several teams ran their own checks outside the joint loop. On engagements where the written responses were more negative, cancellation was more common, and respondents whose written responses were more negative reported higher intent to leave.

Ratings of 3 or below, the distressed range, accounted for 377 of the 787 answers (47.90%). The engagement mean fell in the distressed range in 60 of 150 engagements.

The primary responses numbered 26, 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 3 responses, 10.34% of the total, were secondary and came from 3 other questions. Respondents who wrote both a primary and a secondary response numbered 0. Negative responses made up 73.08% of primary responses (19 of 26) and 3 of the 3 secondary responses.

Of the 29 responses, eleven stated that quality assurance missed the agent's errors or found them too late: a daily check of twenty documents that caught the easy cases and missed misfiled ones, a gate of one senior reading a diff that did not catch configuration errors, a check that caught typos and missed rating rules, a wrong tier rate that passed both firms and surfaced a month later when an adviser complained, and an assistant that told an employee the parental leave allowance from an old policy document with nothing to catch it. Four responses described checks the team ran outside the joint loop: a senior reading each synthesis against the source, a private index of every document the agents touched, a ten-minute sheet of row counts and empty columns that caught two problems in a month, and an analyst checking every rerun because the reruns did not match.

Against these, four positive responses said the joint quality loop caught the errors the agents actually made before they went live, and placed the remaining weakness in ownership, in stopping the agent outside office hours, or in drift between releases.

Of the 29 responses, 22 (75.86%, 95% CI 58.62% to 90.00%) were negative and 5 (17.24%) carried the strongest negative score of −10. Positive responses numbered 7 (24.14%), and no response was neutral. The mean written score was −3.4 (95% CI −5.2 to −1.6, n = 29 responses), and the median was −5.0.

Of the 29 responses, 29 came from respondents who also rated the indicator. Among them, the written score rose with the rating (Spearman rho 0.68, 95% CI 0.39 to 0.86, p < 0.001, n = 29 responses). The written responses corroborated the rating. Of the 22 negative responses with a rating, 10 (45.45%) came from respondents whose rating was in the distressed range.

Respondents who wrote about the indicator and gave a rating numbered 29, with a mean rating of 4.3. The 758 respondents who rated the indicator without writing about it had a mean of 3.7. The difference, 0.7 (95% CI −0.1 to 1.4), included zero. Ratings in the distressed range made up 34.48% of the ratings from respondents who wrote (10 of 29) and 48.42% of the ratings from respondents who did not (367 of 758).

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 38.66, Holm p < 0.001, omega squared 0.371). Active engagements in Build-Out rated highest (mean 4.7, n = 158 respondents) and canceled engagements lowest (mean 2.3, n = 297 respondents). Ratings also differed by respondent role (Welch F 7.28, Holm p = 0.008, omega squared 0.015). Portfolio directors rated highest (mean 4.8, n = 30 respondents) and individual contributors lowest (mean 3.6, n = 605 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 27 engagements with a written response (Kruskal-Wallis H 8.54, df 2, p = 0.014). Fewer than two groups held 10 or more responses, so the groups are reported as counts only: active engagements in Build-Out (8), active engagements in Transition/Cut-Over (9), active engagements in Production (1) and canceled engagements (11).

No test across respondent roles was run, because fewer than two roles had 30 or more written responses. Individual contributors wrote 19, engagement managers wrote 9 and portfolio directors wrote 1 responses, reported as counts only.

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 = −0.61, 95% CI −0.70 to −0.50, Holm p < 0.001, n = 150 engagements). The association held after Holm correction in fourteen 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 = 0.85, 95% CI 0.79 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 = −0.66, 95% CI −0.70 to −0.62, Holm p < 0.001, n = 742 respondents). The association held after Holm correction in thirty-five of the 37 segment levels tested.

Two comparisons set the engagements with at least one negative response against all other engagements. Of the 129 other engagements, 122 had no written response addressing the indicator. These two comparisons therefore do not isolate written sentiment. Of the 21 engagements with a negative response, 10 (47.62%) were canceled, against 46 of the 129 other engagements (35.66%). The difference did not reach significance (Fisher exact p = 0.335). Customer satisfaction among active engagements did not differ between the two groups (mean 64.45% on 11 engagements against 64.05% on 83, Welch t 0.11, p = 0.911).

Two correlations test written sentiment directly. Across the 28 engagements with a written response, a more negative engagement mean written score came with cancellation (point-biserial r −0.52, p = 0.005). Across 28 respondents, a more negative respondent mean written score came with higher intent to leave (Spearman rho −0.53, p = 0.004). Of the four comparisons, two 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 = 0.407).

Answers clustered within engagements (ICC 0.68). Clustering reduced the 787 respondent answers to the equivalent of about 203 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 0.85 to 0.85.

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, 75.86%, ranked 49th from the highest, against a median of 80.82%. Its mean written score, −3.4, ranked 55th 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 8 of the 29 responses (27.59%). C157, “Rate the factual reliability of the AI tools' outputs on this engagement before any human review.” (7 responses), C138, “Rate how clearly ownership of each workflow step is assigned when agents, vendor staff, and customer staff work in the same process.” (6 responses), C147, “Rate how well human review points are placed across the two firms at the moments where agent errors would matter most.” (6 responses) and C141, “Rate how effectively the two firms manage issues arising from agent work as one team rather than through two separate processes.” (5 responses) followed.

The most frequent theme was “Raised it, nothing came back”, in 4 responses (13.79%). The second was “Root cause reordered”, in 4 responses.

Nearly half of all ratings of how reliably the joint quality assurance loop catches agent errors fell in the distressed range (377 of 787 answers, 47.90%). 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 2.3, n = 297 respondents).

The written evidence is small. Its main point is that the checks were designed for the errors people make, such as typing and formatting, while agents make different errors, such as a plausible answer drawn from an outdated policy document or a file placed under the wrong category. Where the checks did not look for those errors, they reached live operations.

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 (29 responses on 28 engagements). Intervals for written sentiment therefore used a bootstrap by engagement.

The written responses number fewer than 30, so their shares and the written-sentiment tests rest on a small base.