No. 03 · The observatory
Pension, provident fund and superannuation administration
HESTA’s provider switch left members disrupted for months. Regulators stepped in.
- Status
- Open
- Rank
- 16 of 35 in the atlas
- Scored
- 2026-10-05
- Report
- Back to the study
- Rank
- 16 of 35, published score 57.5
- Scoring date
- 2026-10-05, reported 2026-10-09
- Buyable
- $2.2B buyable midpoint, $1.4B to $3.6B recorded legacy spend
- Factors
- size 3, pain 3.2, AI fit 4, lock-in 4, crowding 1.7
- Coverage
- none recorded; AI-native challenger recorded
Pension, provident fund and superannuation administration
Ranked 17 of 35 at 57.5 out of 100.
Pension, provident fund and superannuation administration systems of record.
What the evidence says
The checked evidence corroborates bounded buyer pain. A prudential regulator documents severe member-service disruption at HESTA. The disruption followed an inadequately overseen provider transition. A vendor release adds reported census-workload gains after adoption.
Providers
The card checked three incumbents: Aquila Heywood, Capita and Civica. A 2017 council review shows Altair at 57 England LGPS customers. Civica’s UPM held 14; Capita’s Hartlink held 4 in the same table. That ordering is historical and scoped to England’s LGPS only.
Stax.ai is the recorded challenger. Its release reports 8 of the top 10 independent TPAs standardized on it. Adoption and AI claims are vendor self-reported.
Sources:
- 8 of the top 10 retirement TPAs choose Stax.ai (Stax.ai vendor release)
- Pensions Administration Software Market Review (East Sussex County Council procurement review)
The pain
Retirement administration carries heavy manual census workloads. One TPA reported census work dropping from days to minutes.
Provider transitions carry member-service and regulatory risk. APRA found deficiencies in HESTA’s board governance over the transition.
Sources:
- 8 of the top 10 retirement TPAs choose Stax.ai (Stax.ai vendor release)
- APRA imposes additional licence conditions on HESTA
Events
APRA imposed additional licence conditions on HESTA in December 2025. Its provider transition had finalised in June 2025. The regulator recorded direct member harm from the disruption.
Sources:
- 8 of the top 10 retirement TPAs choose Stax.ai (Stax.ai vendor release)
- APRA imposes additional licence conditions on HESTA
Limits
The census-workload figure is a vendor-published testimonial. APRA’s release does not name HESTA’s successor provider. Provider figures are October 2017 historical counts, not a current census.
Sources:
- Pensions Administration Software Market Review (East Sussex County Council procurement review)
- 8 of the top 10 retirement TPAs choose Stax.ai (Stax.ai vendor release)
- APRA imposes additional licence conditions on HESTA
In their words
our census work dropped from days to minutes
Source: PlanPerfect chief executive, 2025-10-07
a severe, prolonged disruption to member services
Source: recorded speaker, 2025-12-11
Relevant workflows
6 recorded openings. A missing mapped round is a coverage gap, not an empty field.
| Workflow | Index | Labour | AI use | Core | Mapped |
|---|---|---|---|---|---|
| Employer arrears, inspections and recovery pension-provident-fund-admin:employer-arrears-compliance | 403 | $41.8B | 28% | non-core | 0 |
| Member service: calls, counters and complaints pension-provident-fund-admin:member-service-contact | 352 | $92.9B | 82% | non-core | 3 |
| Benefit payment runs, interest crediting, statements and regulatory reporting pension-provident-fund-admin:payments-crediting-reporting | -223 | $74.9B | 18% | touches core | 0 |
| Withdrawal, retirement, death and disability claims pension-provident-fund-admin:benefit-claims-processing | -343 | $26.5B | 8% | touches core | 0 |
| Member and employer registration, onboarding and data maintenance pension-provident-fund-admin:member-employer-data-maintenance | -386 | $17.7B | 29% | touches core | 1 |
| Employer contribution intake and reconciliation pension-provident-fund-admin:contribution-intake-reconciliation | -397 | $42.4B | 9% | touches core | 1 |
Recorded providers
Provider claims carry their recorded kind and citations. Undisclosed figures stay undisclosed.
| Provider | Products | Parent | Claims |
|---|---|---|---|
| Aquila Heywood | Altair | unknown | cl-pen-rel-heywood (inference) source cl-pen-inc-heywood (inference) source cl-pen-m-heywood-57 (observation) source cl-pen-m-heywood-members (observation) source |
| Capita | Hartlink | unknown | cl-pen-rel-capita (inference) source cl-pen-inc-capita (inference) source cl-pen-m-capita-35 (observation) source cl-pen-m-capita-4 (observation) source |
| Civica | UPM | unknown | cl-pen-rel-civica (inference) source cl-pen-inc-civica (inference) source cl-pen-m-civica-14 (observation) source cl-pen-m-civica-members (observation) source |
Recorded inputs
Raw scored inputs exactly as frozen. Values and units keep their recorded definitions.
| Factor | Input | Value | Count | Score |
|---|---|---|---|---|
| pain | keep_alive_share | unavailable | 2 | unavailable |
| pain | backend_failure_share | 0.17666666666666667 | 300 | 2 |
| pain | app_rating_inverted | 3.500496956521739 | 23 | 3 |
| pain | failed_rate | 0.6666666666666666 | 9 | 5 |
| pain | s3_grade | weak | 1 | 3 |
| pain | mean | 3.25 | unavailable | 3.2 |
| lockin | award_duration_months | unavailable | 4 | unavailable |
| lockin | noncompetitive_share | 0.1 | 10 | 4 |
| lockin | replacement_duration_months | 57.5 | 8 | 3 |
| lockin | regulator_approval | true | 1 | 5 |
| lockin | mean | 4 | unavailable | 4 |
| crowding | ai_native_traction | 2 | 2 | 2 |
| crowding | funding_per_billion | 17911824.66510497 | 5 | 2 |
| crowding | yc_2024_plus | 0 | 0 | 1 |
| crowding | mean | 1.6666666666666667 | unavailable | 1.7 |
| size | users_midpoint | 1555 | unavailable | unavailable |
| size | buyers_mid | 7796 | unavailable | unavailable |
| size | midpoint | 2196872777.3815217 | unavailable | unavailable |
| ai_fit | mean | 3.7023076923076923 | 12 | 4 |
| ai_fit | unanswered | unavailable | 0 | unavailable |
Recorded sources
24 recorded citations. Each link lands on the sources catalogue.
Back to the 35-category table or the explorer.