IN CONFIDENCE
Contents
Purpose
3
Executive Summary
3
Recommendations
4
Appendix I
Objectives and Milestones
5
Programme Delivery
7
Application Processing Systems
10
Immigration Risk
12
Visa Decision Quality
15
Customer Experience
19
Page
2 of
25
IN CONFIDENCE
Purpose
1. The purpose of this report is to review the implementation and delivery of the 2021 Resident Visa
programme, including factors that contributed to its success. The report also offers some
recommendations for future programmes of a similar nature.
Executive Summary
2. INZ received 106,582 RV21 applications in total. The RV21 programme’s key objective was to
process 80% of applications by June 2023. This was achieved in early February 2023. As of 31 July
2023, 96.5% of RV21 applications had been decided.
3. The deadline for the 80% objective was initially set to be completed by 30 November. The
completion date was later amended to June 2023, along with changes around how applications
were to be prioritised for assessment. Customers were dissatisfied with these changes and
resulted in a large number of customer correspondence; which may have been avoided if external
communication was simpler and more transparent.
4. The average number of applications decided across all platforms per Full Time Equivalent per
week, increased from 6.1 to 23.2 with the introduction of ADEPT.
5. INZ may not have achieved the key RV21 objective if it relied on other platforms, namely AMS,
IGMS, and Plone. There were significant limitations of using these legacy platforms including an
inability to handle peak application volumes and the reliance on manual effort. As such, the ADEPT
platform was critical in delivering RV21.
6. To further assist meeting the key objective, INZ developed a streamlined approach for the RV21
programme. The streamlined approach included:
a. Reducing the number of risk triage rules from 41 to 33.
b. Accepting certain documents at face value.
c. Not requiring offshore police certificates and full medical certificates.
d. Reducing the number of questions in the quality control (QC) check question set from
60 to 22.
e. Conducting QCs on RV21 applications that met the ‘Skilled’ and ‘Scarce’ criteria only,
thus reducing the QC rate below 100%.
7. The streamlined approach to risk management did not produce any significant gaps or unintended
consequences, however not enough time has elapsed to gather sufficient data on adverse
outcomes.
8. Visa decision quality was sustained, and quality scores were not negatively impacted by the RV21
specific streamlined approach.
Page
3 of
25
IN CONFIDENCE
9. Proportionately, fewer RV21 customers submitted formal complaints, and no significant
differences between RV21 customers and customers of other visa products were found using
Customer Experience Survey data.
10. There were, however, a number of issues that affected the delivery of RV21. Examples include:
a. The ADEPT platform was not fully developed and tested in time for RV21 processing.
b. Payment failure errors when submitting applications with more than six applicants.
c. RV21 customers being recorded as unlawful when they were in fact lawful, due to
interim visas not generating.
d. Poor ‘technical support’ processes.
e. A lack of transparent communication about the limitations of ADEPT.
f. Staff finding it challenging with the amount of change occurring at all levels.
g. There was a significant cost to staff wellbeing throughout the delivery of the
programme.
11. Further information of the above findings is detailed in Appendix I.
Recommendations
12. Delivery of the programme felt rushed by multiple individuals involved in the implementation of
the programme. To ensure a more successful programme, more realistic timeliness windows
should be agreed upon.
13. It was reported by programme leaders that the delivery of the RV21 programme may have been
better in terms of control, oversight and consistency, if it was restricted to one visa processing
office rather than spread over multiple sites. Future one-off programmes may benefit from
concentrating visa processing to one office, if possible.
14. A large number of RV21 specific correspondence may have been avoided if communication to
customers were more transparent. INZ should take a closer look at its communication methods
used when there are changes made to announced programme objectives, including informing
customers as to how their specific application would be affected.
15. Technology support for ADEPT related issues were reported as being difficult to manage, as the
escalation process was not well defined. Suggested improvements for managing issues included
developing better communication channels between processing staff and product owners.
16. It was reported that sufficient consultation with Technical Advisors regarding the quality control
check process did not occur, which consequentially led to a decreased capability to manage quality
at the visa processing office level, especially in ADEPT. It is therefore recommended that any future
designs of the quality check process are developed in partnership with Technical Advisor teams.
Page
4 of
25
IN CONFIDENCE
Appendix I
Objectives and Milestones
17. The 2021 Resident Visa (RV21) was announced
in September 2021 to provide a one-off pathway
to residence for migrants, partly as a result of the effects of COVID-19 on normal migration
patterns. The objectives of the policy were to:
a. Provide certainty for employers that their existing migrant workforce would be able
to remain in New Zealand permanently.
b. Enable migrants who were already in New Zealand for the purpose of work to remain
permanently.
c. Attract health workers and other highly skilled critical workers by providing them a
pathway to stay in New Zealand permanently.
18. Eligibility was primarily based on applicants meeting at least one of the following three criteria:
d. Settled
– lived in New Zealand holding an eligible visa for at least three years.
e. Skilled
– in employment where they have been paid at least $27.00 an hour.
f. Scarce – in employment where their role has been deemed by INZ as scarce.
19. The programme’s key objective was to complete at least 80% of applications within 12 months of
delivery, from December 1, 2021 to November 30, 20221.
20. RV21 was implemented in two phases, with Phase 1 split into two sub-phases: Phase 1 and Phase
1.5 (see Figure 1).
1 This was later amended to 30 June 2023, as discussed in ‘Programme Delivery’ section
Page
5 of
25
Phase 1
21. Applicants who had the following characteristics on 29 September 2021, were able to submit an
RV21 from the opening date, 1 December 2021.
g. An open application for residence under the Skilled Migrant Category (SMC).
h. An open application under the Residence from Work Category (RfW).
i. An open Expression of Interest (EOI) which included a dependent child aged 17 years
or older.
22. Prioritising these customers aimed to ease the ongoing delays of the backlogged SMC and RfW
queues, as well as allow working age dependents to enter the New Zealand workforce.
Phase 1.5
23. Phase 1.5 opened on 21 February 2022, allowing all other INZ customers who held an EOI on 29
September 2021.
Phase 2
24. Phase 2 opened to all other eligible applicants under the Settled, Skilled, or Scarce criteria.
25. On 31 July 2022, INZ closed the RV21 category and customers could no longer submit an RV21
application. 28,964 applications were lodged under Phases 1 and 1.5, and 77,618 were lodged
under Phase 2, resulting in 106,582 RV21 applications in total.
26. As of 31 July 2023, 96.5% of applications have been processed.
Programme Delivery
27. Before the announcement of the programme, senior leaders of the Manukau office were
designated as Project Leads, with ownership of the delivery of RV21. The Manukau Project Leads
reported that they embraced this decision (made in partnership with INZ Operational Policy team).
They believed it was logical to design a programme that they would be responsible for and are
subject matter experts on. Part of the programme design included developing a streamlined
approach in order to meet the key objective.
28. Although the RV21 programme was led by Manukau, it was delivered over three processing
offices: Manukau, Hamilton, and Christchurch. A large majority was processed by the Manukau
processing office. In hindsight, it may have been better to restrict delivery to one processing office.
Having one office responsible for the delivery of the programme would have ensured
consistencies, with control and oversight of delivery being easier to maintain.
29. In June 2022, a decision was made to amend the key objective. As a result, INZ was to now
complete 80% of RV21 applications by
30 June 2023. Changes were also implemented around how
applications were to be prioritised for assessment.
30. The Manukau processing office reported that a large number of customers were dissatisfied with
both the above changes, as customers were unsure how the changes would affect processing of
their application. Manukau reported a large influx of complaints, and thus additional work. It was
Page
7 of
25
IN CONFIDENCE
37. Taking into account only the quantitative evidence summarised above, the delivery of the RV21
programme was a success. However, if INZ is to present itself as a people-centric organisation,
then any future programmes should consider the wellbeing of its staff during periods of transition,
especially processing office staff.
Application Processing Systems
38. Historically, most Residence visa applications have been paper based. As a result of the COVID-19
related lockdowns, staff were not able to enter physical processing offices, This was a significant
disruption to the processing of Residence applications. To mitigate this issue INZ decided to house
RV21 processing on online platforms only, especially so to further guarantee meeting the key
objective.
39. At the time of the announcement of RV21, INZ was already using AMS, and the Immigration Global
Management System (IGMS) for online submissions, however it was mainly being used for
submitting temporary visa applications. It was decided that RV21 Phase 1 applications would be
received via IGMS. The Enhanced Immigration Online Platform, otherwise known as ADEPT, was
still being built.
40. When the programme entered Phase 1 on 1 December 2021, INZ received a large initial influx of
applications, which affected the performance of the IGMS platform. Subsequent performance
testing indicated that IGMS would struggle with peak loads as observed in the opening phase of
the programme. It was reported that processing during Phase 1 was straightforward as AMS was
familiar, but also that the platform connected with 3rd party stakeholders such as the s 6(a)
and the Immigration Health System (IHS). To cope with peak loads,
INZ made the decision to use the Plone platform. Plone was historically deployed at times of high
demand and outages. It was also used to manage border exceptions during the COVID-19 border
closure. Although Plone was largely able to handle the influx of applications, it did require
significant manual processing, affecting all parts of the visa assessment process. Given the amount
of effort required to lodge applications using Plone and the expected number of applications to
be received in Phase 2, Plone was not marked as the platform to be used as a permanent solution
if INZ was committed to achieving the Key objective.
41. Due to the issues identified during Phase 1, some changes were made for Phase 1.5. To alleviate
the difficulties with peak loads, submitted applications were staggered. INZ accepted 1,500
applications per day, with a maximum of 200 per hour. This process was supported by regular
monitoring. As a result, there were no additional disruptions reported during Phase 1.5. Given the
expected application volumes for RV21 Phase 2 customers, AMS, IGMS, and Plone were not
marked as the technology solution.
42. INZ decided that Phase 2 would be processed using the ADEPT platform. As ADEPT is a cloud-based
software service, it’s capacity was scalable to accommodate much larger volumes of applications.
However, the service was not built with RV21 in mind. ADEPT was developed to accommodate
processing of approximately 1,000,000 applications across all categories per year, which equates
to 4,000 per day. During Phase 2, it was estimated that INZ would receive around 120,000 RV21
applications within the first few days, well beyond ‘normal’ visa volumes. With the support of
Microsoft, the INZ ADEPT Operations Team (AOT) implemented a solution, using an iteration
Page
10 of
25
IN CONFIDENCE
process with a front- and back-end throttling capability. This capability enabled applications to be
submitted, but then held in a “virtual lobby”, until the back-end processing development was
ready to begin visa assessments.
43. Overall, the ADEPT system coped well with the incoming demand, allowing for more efficient
processing. ADEPT was able to automatically identify and trigger requests for Medical Certificates,
and NZ Police checks. It also allowed for a higher capacity of additional documentation to be
uploaded, and visas could automatically be issued. The success of the implementation of ADEPT
for the RV21 programme meant that productivity rates more than doubled. It can be concluded
that the RV21 programme’s key objective may not have been achieved if INZ relied on AMS, IGMS,
and Plone, and as such ADEPT was critical in the success of the RV21 programme.
44. ADEPT was not without its own issues. Although AOT reported that once issues were raised, they
were fixed as quickly as possible either with a permanent fix or a workaround. This was made
possible by the level of built-in flexibility and the team of technical and subject matter experts
available for support. This however was in contrast to the user experience at the processing
offices. Processing staff reported that there were major difficulties in resolving issues, especially
during the beginning of Phase 2. This view did ease over time as workarounds were deployed.
There was some consensus at the processing office level that the ADEPT related issues were due
to a lack of design forethought. It was suggested that the ADEPT product owner be made more
accessible, and more efficient project management to enable smoother running of the
programme.
45. A number of specific issues are described in Table 1. This was expected as ADEPT was not fully
developed, and so the system was not systematically tested. A lot of the critical errors were
identified and fixed earlier in the year, when the Visitor Visa product transitioned onto ADEPT.
Table 1.
Summary of some ADEPT related issues
Initial ADEPT issues during Submission
Initial ADEPT issues during Processing
Payment failure error message when trying to Training for ADEPT was insufficient, and
submit seven or more individuals on one commonly occurring issues had inadequate
application. This issue is still being monitored.
responses.
Some customers were incorrectly recorded as The ADEPT User Guide didn’t cater for all
unlawfully in New Zealand and were not able to scenarios, including the differences between
submit an application. Customers were unaware applications being ‘pushed’ to immigration
of this error until they contacted INZ. Clearer officers vs applications being ‘pulled’ by
communication has since been implemented.
immigration officers.
The application form had the National ID As ADEPT was a new technology solution, there
Number as a mandatory field, despite some was a steep learning curve alongside a steep
nationalities not using national ID numbers. cultural change due to the streamlined
Customers had to enter ‘Nil’ to progress the approach. This amount of change was difficult
application.
for some to deal with.
Because of the use of the ‘virtual lobby’, interim The ‘Help’ infrastructure was not designed well
visas were not being automatically issued. A enough, especially for those who were not tech
manual workaround was put in place to savvy. Those staff who were tech savvy,
Page
11 of
25
IN CONFIDENCE
generate interim visas on a daily basis for any unofficially became ‘experts’, as it wasn’t easy
affected customers.
to get support. A lot of reliance was placed on
them to raising tickets with the MBIE Service
Desk.
Some customers who were eligible to be There was a lack of transparent communication
included in an RV21 application, were unable to about the limitations of ADEPT and a perceived
submit an application due to eligibility screening lack of ownership from ADEPT product owners
questions. Given the low number of cases, a of the staggering amount of issues. A better
workaround was deployed directing these escalation process was suggested for future
customers to a paper form.
programmes.
The form requested for details of the parents of It was suggested that ‘Super User’ type access
the customer, despite there being no field to could have helped with trouble shooting.
enter such information. This was later fixed.
Immigration Risk
46. To enable timely delivery of the RV21 programme it was agreed that any proposed risk
management process would need to support the overall streamlined approach and enable
efficient decision making. The Risk Monitoring Review Governance Group (RMRGG), in
consultation with the Risk and Verification (R&V) network and the Risk Control Group (RCG),
agreed to risk appetite thresholds that would support a streamlined approach. As a result, several
adjustments were made to tolerance settings and risk controls.
Phases 1 & 1.5
47. The key risk areas identified in preparation for Phase 1 include: national security risk, non-genuine
partnership claims, identity fraud, and character concerns. Of the 41 triage rules that would
normally trigger on regular residence visa applications, 33 rules were excluded from triggering on
RV21 applications, aligning with the key risk areas identified by the RMRGG. The same risk
management approach was deployed for Phase 1.5, with the inclusion of an additional risk rule
(RR1013)
s 6(c)
For all phases, the approach to adverse information
triggering within the system was maintained. It was also agreed that applicants would still require
New Zealand Police Checks (NZPC) and specified applicants would also require National Security
Checks (NSC). To support streamlined decision making it was agreed that certain documents could
be accepted at face value and that offshore police and full medical certificates would not normally
be required for assessment.
48. The RMRGG also recommended post-decision monitoring of outcomes to refine risk management
at the application stage and to drive compliance with immigration standards. R&V completed risk
monitoring activity on 5% of the approved RV21 Phase 1 and 1.5 cohort, where a partner was
included, to better understand the risks and determine whether any changes to risk treatments
for this category were required. The results of this monitoring activity concluded that the current
risk treatment was appropriate and would continue to be used for Phase 2.
Page
12 of
25
IN CONFIDENCE
54. Of the risk rules that triggered for Phase 1, RR0604 triggered with the highest frequency. s 6(c)
. In Phase 2, RR1013 triggered
with the highest frequency. s 6(c)
This was likely due to a large amount of RV21 applicants’ employment not having ever been
assessed by INZ, relevant for the Skilled and Scarce criteria. The risk rule with the second highest
frequency of triggering during Phase 2 was RR1009, which was designed to surface AMS alerts and
warnings in the ADEPT platform.
55. Of the applications that were triaged as ‘High Risk’, 0.46% were declined and 1.2% were
withdrawn. Of the applications that were triaged as ‘Medium Risk’, 0.34% were declined and 0.3%
were withdrawn. Of the applications where a triage risk was not applicable, 0.35% were declined
and 0.7% were withdrawn (see Table 5).
Table 2.
Risk Level of RV21 Applicants as of 26 July 2023
Decision
Risk Level
Approved
Declined
Withdrawn
Total
14,938
68
183
15,189
High
(98.35%)
(0.45%)
(1.2%)
(100%)
39,595
136
118
39,849
Medium
(98.36%)
(0.34%)
(0.3%)
(100%)
148,330
523
1,124
149,977
Not Applicable
(98.90%)
(0.35%)
(0.75%)
(100%)
633
7
603
1,243
Other
(50.93%)
(0.56%)
(48.51%)
(100%)
203,496
734
2,028
206,258
Total
(98.66%)
(0.36%)
(0.98%)
(100%)
56. While the applications that triaged as ‘Not Applicable’ had a higher percentage of decline and
withdrawal decisions than those that triaged medium risk, it is likely that these applications did
not meet immigration instructions, such as for Character or Health, and thus have not been
analysed in this report.
57. The intended streamlined approach to the RV21 category is clearly reflected in the interventions
and outcomes. In summary, 73% of all applications did not automatically trigger any risk flags and
only 0.36% of all applications processed (by 26 July 2023) have resulted in a decline decision.
58. Risk monitoring activity completed between phases did not reveal any significant gaps and
supported the continuation of the proposed risk management approach for Phase 2.
59. Quality assurance was completed on 152 or 0.07% of the completed RV21 applications, post-
decision. These checks resulted in a Risk specific score of 92.1% and did not identify any concerns
with the risk mitigation of the completed RV21 visas. While the overall score for risk mitigation
was 4% lower compared to that of regular residence visas, it is not a significant difference to be of
concern (see
Visa Decision Quality section below).
Page
14 of
25
IN CONFIDENCE
60. It is also worthy to note processing office staffs’ views on management of risk. The view was mixed.
61. Some processing office staff agreed that the design of the RV21 category mitigated immigration
risk appropriately. It was mentioned that the category was designed to only include applicants
who met certain criteria and immigration instructions were simple and easy to assess. It was
agreed that both instructions and processes were clear, and R&V were available to provide advice
for applications that were identified as High risk.
62. There was however, a small amount of scepticism regarding the propensity for risk, but this eased
over time as processing staff became accustomed to the streamlined approach. This resulted in a
change of culture of application processing, as decision makers were “less defensive” and more
open to approving applications with a lesser degree of checks undertaken.
63. Some processing staff noted that there were differences in the practical handling of risk between
applications assessed on AMS vs ADEPT. ADEPT enabled automation of some tasks and this meant
that occasionally, applications with risk would ultimately be approved. A minority continued to
feel uncomfortable with the fact that more risk-laden cases were being approved.
64. Despite the residual concerns, there is no evidence to suggest that risk was handled
inappropriately throughout the implementation and delivery of the RV21 programme.
65. It is not yet possible to fully assess the actual impacts of the applied risk management approach
for RV21 until sufficient time has lapsed to gather data on adverse outcomes. Such outcomes
would usually first become visible during Permanent Residence Visa processing when assessing
whether the Residence Visa holder is eligible for Permanent Residence based on the commitment
demonstrated while holding a Residence Visa.
Visa Decision Quality
66. Alongside the organisational RV21 objective of completing all applications within 12 months, INZ
was also committed to maintaining high visa decision quality for RV21.
67. Visa decision quality is measured by INZ at two points: pre-decision and post-decision. The quality
model aims to ensure that quality performance is sustained or improved over time.
Quality Control
68. The quality control (QC) check is completed pre-decision i.e. before a visa assessment is
completed. The purpose of the QC check is to confirm that the necessary decision-making
processes were followed, and if not, are completed before a final decision on the application is
made.
69. Because the RV21 programme was a one-off visa category ultimately aiming to automate some
decisions, it was decided that the QC component of the visa assessment procedure would be
simplified as much as possible (streamlined) while still meeting INZ’s quality standards. The
streamlined approach included reducing the number of questions in the question set, and
decreasing the QC rate below 100%.
Page
15 of
25
IN CONFIDENCE
using the QC Tool, which sits on an online platform independent to AMS, ADEPT, IGMS, and Plone
altogether. QCs for RV21 Phase 1 were completed on this QC Tool.
76. QCs for RV21 Phase 2 were completed within ADEPT. This was the first time QC checks were being
conducted on the ADEPT platform. As such, there were critical technical issues that prevented
conducting QC checks in an efficient manner. One such issue led to a backlog of 2,000 QCs and
often taking longer to complete a QC check than to complete the manual assessment activity.
Therefore, in June 2022, INZ decided that those RV21 applications that fell under the ‘Settled’
criteria would not undergo a QC check. This decision allowed INZ to focus 100% of its QC efforts
on the RV21 subtypes that presented more risk (Skilled, Scarce).
77. This does not mean that Settled application had no oversight. The Risk and Assurance team
reviewed a sample of ‘Settled’ RV21 applications post-decision each quarter as a discrete exercise.
To date, no issues have been found. This further indicates that the decision to drop QC checks for
‘Settled’ RV21 applications did not have an adverse impact on visa decision quality. This decision
was also well-received by TAs and the instructions were clear for them to quickly change the QC
process.
78. For Phase 2 specifically, some TAs felt that quality control was difficult to manage in ADEPT as they
were less able to truly control quality effectively, which led to some cases having to be reassessed,
including after a decision had been made. This may have been an unintended consequence of
either the truncated question set not being able to capture a critical error, or not identifying
potential issues for ‘Settled’ applications.
79. It was suggested that any future designs of the QC process should consult TAs, as they do the large
majority of the quality control tasks, and understand the subtleties of QC related issues the best.
80. Although the attempts to streamline the QC process were successful, contributing to achieving the
Key objective, it is important to evaluate whether visa decision quality was sustained or adversely
impacted. This can be observed from Quality Assurance data.
Quality Assurance
81. Quality Assurance (QA) is performed post-decision on a sample of decided visa decisions. The
scores are aggregated into two scores: one for Residence applications and one for Temporary
decisions. Analysing quality of RV21 quality over time, allows to observe whether visa decision
quality was sustained from the implementation of the programme onwards.
82. Between the Dec 21 – Jun 23 period, 152 RV21 quality assurance assessments were conducted,
analysed over five quarters. Figure 7 illustrates the overall RV21 quality scores compared to the
aggregate Residence score excluding RV214.
4 Data for the Apr-Jun 2022 quarter is missing as a business decision was made not to quality assess this
particular quarter. Instead qualitative analysis was conducted on a sample of ‘Settled’ RV21 applications. No
issues were found for all 30 applications.
Page
17 of
25
IN CONFIDENCE
110.
Process delays are overwhelmingly the issue that most RV21 customers complained about,
with half of all complaints containing some dissatisfaction with the time taken to process an
application. This is not an extraordinary finding, as delays are the most common case ground
overall.
111.
Additional reports from processing offices indicated that the extension of the timeliness
objective from December 2022 to June 2023 as well as the decision to prioritise some applicant
skills over others led to decreases in customer satisfaction.
112.
Taking into both sources of data from the CXS and CFT, it appears that customer satisfaction
with the RV21 product was more positive than negative.
Page
24 of
25
IN CONFIDENCE
Document Control
Version
Date
Responsible
Description of change
1.0
11 Aug 23
Divesh Prakash
Initial Draft
1.1
15 Aug 23
Divesh Prakash
Second Draft
2.0
25 Aug 23
Divesh Prakash
Final Draft
3.0
29 Aug 23
Divesh Prakash
Final
Document Handling Instructions
Security Procedures
This document must be handled, stored and transferred in accordance with the security procedures
applicable to its security restriction as detailed in the MBIE Security Hub.
Dissemination
Primary Customers (as noted on the front page) are welcome to share this document internally
within INZ. This document (or attachments) must not be distributed to other work areas outside of
INZ without prior authorisation from a Manager from the Insights Team. Any requests for further
distribution of this document should be directed to the Manager Insights and include a reason for
the request.
Copying
This document may not be copied without authorisation from a manager within the Insights Team.
Information in this document may only be incorporated in other documents or otherwise used, subject
to the conditions in the Handling Instructions and provided that such use does not lessen the degree
of protection afforded this information.
Official Information Act 1982
This document remains the property of MBIE. The release of information contained in the document
may prejudice the maintenance of the law, including the prevention, investigation and detection of
offences.
Privacy Act 2020
This document may contain information relating to individuals that is covered by the principles of the
Privacy Act 2020. Accordingly, this document should be protected by use of the above security
measures to safeguard against its loss, or unauthorised access, use, modification or disclosure.
Page
25 of
25