Know who needs which support now. See whether it arrived.
EVA is an AI agent that follows up with participants and field teams, joins their answers to your programme register, and shows where every finding comes from.
EVA provides evidence and recommendations. Your people make the decisions.
Live within the United Nations and LVMH, across 120+ countries.
Who needs which support, and what has changed?
A baseline describes one moment. EVA maps what people can do and shows the evidence. Here the people are your programme's participants and field teams. Start with one programme and one support decision. The same follow-up can be scoped for enterprise-support, training, cash-assistance and livelihoods programmes, each with its own record and safeguards.
Update the record, review the support, track delivery.
Start from the existing programme register.
Agree the population, purpose, record identifiers and service options with the programme owner. Reuse agreed baseline information and avoid asking people to repeat it. Keep the current survey or case-management system as the agreed record of reference.
Ask what has changed.
Scope short follow-up conversations with participants or business owners, and a separate delivery-team view where useful. Date each account, retain its source and show unanswered questions. Offer an assisted route when language, connectivity or access makes direct participation difficult.
Review support options and track delivery.
Compare expressed needs and programme criteria with available interventions and capacity. A caseworker reviews eligibility and exceptions. Distinguish a suggested referral from support accepted, delivered and followed up. Report coverage and missing evidence alongside outcomes.
See an updated record and its next support action.
This example follows a small-business support programme. Review what changed, which support is available and what has actually been delivered.
| Programme record | Latest evidence | Option to review | Owner and open check |
|---|---|---|---|
| SME-042 · Food producer (owner, 10 Sep) | Reports a new buyer; adviser notes a packaging gap (field visit 12 Sep) | Packaging support, subject to programme criteria | Adviser confirms provider capacity and owner interest |
| SME-087 · Repair business | Baseline six months old; no response to two follow-ups | Assisted follow-up by the field team | Field team checks access; silence is not ineligibility |
| SME-113 · Tailoring workshop (owner, 8 Sep) | Bookkeeping training completed 30 Aug (provider record 2 Sep); owner reports first records kept | Follow-up at three months | Delivered; recorded; next check December |
Keep support accessible and decisions accountable.
Support decisions stay with people
EVA does not make final decisions about assistance, benefits or eligibility. Keep the reviewer, criteria and route to challenge a decision clear.
Participation must remain accessible
Do not treat a missing digital response as lack of need. Scope assisted participation and avoid making an AI conversation the only route to support.
Show who has been reached and whose needs are still unknown.
Show who was invited, who responded, when records changed and what is unconfirmed. A programme register is not automatically a population census.
Start with one programme and one follow-up cycle.
Bring one programme, its register structure and the support decision you make most often. Agree a follow-up cycle that shows who responded and by which channel, what changed, which referrals were accepted and which support was delivered. A referral count alone is not an impact result.
Proposed application. Delivery, language coverage and connections are agreed in scope.
Does follow-up change what the programme delivers?
Compare the effort to obtain useful updates with the support decisions they inform. Track referrals through acceptance and delivery, while keeping unconfirmed needs and missing responses visible.
Usable updates and response by channel
Count usable updates among invited participants and compare response by channel.
Reviewer effort
Record the time spent reviewing updates, resolving gaps and arranging follow-up.
Accepted referrals
Track which referrals participants accepted, separately from support delivered.
Completed support
Confirm delivery and follow-up with the responsible provider and participant.
Measures are agreed with you against your baseline; they are not promised results.
- Decision owner
- Programme director allocates support and answers for participant outcomes.
- Working lead
- Monitoring and evaluation lead separates current needs, completed support and missing follow-up in donor reporting.
- Review and control partners
- Delivery-team lead coordinates field teams and service providers under the programme framework.
Questions before you start.
3 questions to explore
How does this work with our survey or case-management system?
No replacement is assumed. Begin with the data and workflows you already use. The trial should establish whether conversational follow-up adds useful evidence and reduces chasing. Any connection or automatic write-back is separately scoped.
You might also want to know
Where this connects.
National Skills & Employment
For employability programmes: the same conversation, aimed at vacancies and training places.
Explore this decision →Surge & Emergency Rosters
If your agency also holds a response roster: a different decision on the same evidence record.
Explore this decision →Competency Mapping
The evidence record underneath every EVA conversation.
Explore this decision →Discuss one programme's follow-up cycle.
Use the call to describe one programme and follow-up cycle, the people involved and what a useful first output would look like. No personal records are needed.