evidoria

← Back to browse

Good practice Imported

Nadia — Australia's Cate Blanchett-Voiced NDIS Virtual Assistant That Never Launched

Australia · Geelong · See the Australia profile · See the Geelong profile

Evidence: Descriptive / self-reported Top 82% 52/100 · Ask Evidence Copilot about this practice

Australia's NDIS built an AI avatar voiced by Cate Blanchett to help citizens navigate an 8,000-call weekly hotline. After AU$3.5m+ and 18 months, co-design collapsed under bureaucratic caution and Nadia was shelved before ever launching.

8,000 calls/week
NDIS hotline call volume
3.5 AUD million+
Development spend before discontinuation (by mid-2017)
Nadia — Australia's Cate Blanchett-Voiced NDIS Virtual Assistant That Never Launched

Details

Maturity
Discontinued
Promoter
National Disability Insurance Agency (NDIA), Australian Government
Period
2016–2017 (development); discontinued 2017
Keywords
government, disability services, AI virtual assistants, digital government

Context

Australia's NDIS hotline handles 8,000 calls a week; the NDIA commissioned Nadia, an AI avatar built on Soul Machines' facial-animation technology and voiced by Cate Blanchett, to help people navigate the scheme.

Activities

The project was publicised in February 2017 with a planned 12-month supervised traineeship before public release.

Results

By September 2017 no launch date had been set; more than AU$3.5 million had been spent, the traineeship never started, and Nadia was never made publicly available.

Conclusions

An academic case study concluded that whole-of-government risk-aversion after the Census outage and 'robo-debt' controversy crowded out the disability co-design process that had originally shaped Nadia.

Implementation
Indicative cost
Medium (€50k–€500k)
Time to results
Short (< 1 year)
Staffing & skills
Soul Machines technical team, NDIA project staff, disability co-design participants

Conditions for success

  • protecting the disability co-design process structurally, not just rhetorically

Common failure modes

  • whole-of-government risk-aversion crowded out co-design
  • immature speech-recognition technology at the time

Commonly funded by

National / regional programmes

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful