Implementation and training approach

Access is granted, never self-service

VoiceInsights is an invitation-controlled platform. An organisation is approved and provisioned; people inside it are invited by someone who holds the authority to invite them; and roles are granted in the platform by a person who may grant them. There is no public sign-up, and nothing in this approach assumes one. If your procurement process expects a self-service trial account created from a web form, this is not that kind of product and the difference is worth surfacing early.

The five stages

1. Scope

   We establish what you are trying to find out, in which countries, with which teams, and under which governance. This determines the research modes, the collection routes, the languages and the country structure — and it is the stage where a requirement we cannot meet should surface, not the stage after.

   You provide: the programme, its countries, the teams and their roles, and any data-residency, ethics or safeguarding requirement that constrains the design.

   Produces: a scoped configuration and an explicit list of anything requested that the platform does not do.

2. Configure

   Your organisation is created, the country structure is set up, roles are defined, and access is granted — including any geographic grant that narrows a person to one country. Instruments are built or imported from XLSForm. Collection routes are enabled for the channels your respondents actually use.

   You provide: the people who will hold each role, and approval of who may see what.

   Produces: a working organisation with real users, real roles and real instruments.

3. Pilot

   A small, real collection runs end to end: field work, sync, quality review, analysis, evidence and a first report. A pilot exists to find the things a configuration review cannot — a question that reads differently in translation, a channel that does not reach the people you need, an enumerator workflow that does not survive contact with a field day.

   You provide: a small number of real enumerators and a real, low-stakes collection.

   Produces: a corrected configuration and a team that has completed the workflow once.

4. Train

   Role-based training against the actual product, using the same guides published in the Help Center rather than a separate slide deck that will drift from the software. Enumerator training is run on a phone, because that is where the work happens.

   You provide: the people, and a decision on who trains future joiners.

   Produces: trained teams, and your own trainer if you want one.

5. Operate and review

   The deployment runs. We review with you on an agreed rhythm: what is being collected, where quality is drifting, what evidence has been produced, and what decisions followed. Support runs through the enterprise support route rather than through a sales contact.

   You provide: an owner on your side who can make configuration decisions.

   Produces: a running programme and a record of what changed and why.

Training, by role

Training follows the roles the authorization model actually has, not a generic curriculum. Each of these is published as a guide, so training material and documentation cannot diverge:

• Enumerator — assignments, offline working, completing responses, sync, and what to do when something fails in the field.
• Organization Admin — people, access, projects, geographic scope, and what the organisation can and cannot see.
• M&E Officer — field progress, data quality, evidence review and the handoff to analysis.
• Analyst — datasets, analysis, what the statistics engine will and will not do, and the boundary on interpretation.
• Executive and decision-maker — recommendations, management response, decisions and what happened next.

What we are responsible for, and what you are

What this document does not promise

Back to the Procurement Center →