United StatesPublic sector2026 – ongoingInternational Rescue Committee
Signpost AI assistance for humanitarian responders
IRC reported nearly 70% higher staff productivity in early humanitarian AI pilots requiring human review
Problem
Signpost connected people affected by crises with localized information and personalized support. Providing that assistance through human responders was difficult to scale.
Approach
IRC tested AI assistance for responses, with human review required before sharing them safely. Early results informed Signpost AI, an agent-building platform developed with support from partners including Google.org and Zendesk. IRC reported that Signpost AI-enabled agents had piloted in 8 countries.
Results
Takeaway
The productivity result came from early pilots that retained human review. It did not establish that autonomous responses met the same quality standard.
Categories
- Size
- Large enterprise
- Industry
- Public sector
- Business function
- Customer serviceInternal work and knowledge management
- Innovation type
- Process innovation
- AI technology
- Generative AIAI agents
- Adoption
- Partner co-development
- Evidence
- Company's own claim
Sources
Source summaries
IRC described early AI response pilots and subsequent development of Signpost AI. It reported productivity and quality results while explaining that early responses required human review.
Full details are in the original sources.