A pilot designed to answer a question, not to test a platform.
Most software trials measure adoption — did the user log in, did they click, did they finish onboarding. The right pilot measures something else: what specific decision can you make differently in two weeks, that you could not make before. Starter is configured to deliver that. Pick the operating question — is the manager development gap real and where, or is the cost of routine HR/IT enquiry larger than we assume — and run the instrument that answers it. The free tier is persistent; the answer is what defines whether the platform is worth paying for.
The perimeter of the free tier
Both flagship instruments — knowledge-grounded chatbots and multi-rater feedback — operating on real users and real organizational data. No degraded mode, no time limit.
Resets monthly. Sized to absorb a representative volume of customer enquiries or one complete evaluation cycle.
Sufficient population for a statistically meaningful multi-rater cycle. Administrators, raters, and ratees included.
One production assistant. Embed it on the public site as a support channel, or expose it internally to your workforce.
Documents, transcripts, URLs and reference entries. The assistant retrieves only from sources you have authorised.
Pick the decision the pilot has to inform.
Two questions Starter is sized to answer with real data in two to four weeks. Decision A: 'Where are our manager development gaps actually concentrated — and is the cost of arbitrary performance reviews showing up in our attrition'. Run the 360°. Decision B: 'How much of our HR/IT load is genuinely routine enquiry the workforce can self-serve, and what is that worth at our cost-per-ticket'. Deploy the assistant. The second pilot remains available the moment the first delivers its read-out.
A structured read-out on how the team actually performs.
Configure the competency frame in minutes. Each ratee is evaluated by peers, line manager and self, anonymously at write. The platform reconciles the raw scores into quantitative dimensions and an AI-drafted narrative that surfaces convergent strengths and contradicted areas across rater groups — moving directly to the development conversation, without the spreadsheet detour.
- Configurable competency frameworks, question banks and rating scales.
- Anonymous-at-write architecture: individual responses are not visible to the ratee or their management chain.
- Automated invitations, reminders and response collection — administrator time near zero.
- AI synthesis: quantitative scores, evidenced strengths, growth areas, contradiction map — exportable to PDF or shared in-platform.
An assistant that retrieves only what you have documented.
Index your own documents, policies, product manuals or internal knowledge. Embed the assistant on the public site as a support channel, or expose it internally as a helpdesk for your workforce. Every answer is grounded in source material, cited at the document level, with out-of-scope queries refused by design rather than hallucinated.
- External-facing widget or internal portal — the same engine, different access perimeter.
- Configurable interaction modes: direct answer, guided procedure, onboarding, structured training.
- Ingests PDFs, Office documents, video transcripts, web pages and reference entries — all retrievable with citation.
From registration to first signal in one working session
Self-serve registration. No sales gate, no onboarding fee, no proof-of-concept timeline. The platform delivers production output within the first session.
Register
Provision the tenant with a corporate email. The Starter perimeter is applied automatically — no billing details, no procurement step.
Index the knowledge base
Upload documentation, attach URLs, ingest reference entries. The retrieval index is operational within minutes.
Deploy or convene
Embed the assistant on the site or distribute the internal link. Or convene raters and launch the first multi-rater cycle.
Read out and iterate
Read the synthesis. Audit the conversation logs. Refine the source material. The platform stays free until monthly volume genuinely outgrows it.
Frequently Asked Questions
Specifics on the Starter perimeter and the governance model. Contact our team if something's missing.
Run the pilot.
Provision the Starter tenant, deploy the assistant, run the first multi-rater cycle. The platform delivers production output before the procurement conversation begins.
Self-serve. No commitment. No procurement step.
