VECTIQ
← Vectiq Product Platforms

AI Mentor - SOP Training That Remembers

This is not a chatbot over your documents - we build those too, under Gen AI & Agents, and they answer a question and forget you. AI Mentor is a training system: it turns SOPs into assessment, tracks what each person got wrong, and schedules the next review for the day they are about to forget it. A chatbot makes a person competent for one minute. This makes them competent in six weeks, and shows a manager the proof.

Launch app ↗
EnterpriseEducationBanking & Finance

Interactive demo

Adaptive review scheduling

One card, four learners, four different schedules

Rate how well you recalled each card and watch the schedule recompute. This is where AI Mentor differs from a chatbot: it does not only answer, it decides when you have to meet the fact again.

2
Due today
87%
Recall probability
0
Reviews logged
SOP-CS-014 §3.2

A customer asks for a refund on day 9 after delivery. What do you do?

Schedule across the deck

The scheduling formula here is a simplified illustration. The real system uses FSRS with parameters fitted to each person's actual review history. The SOP text is an example, not any customer's procedure.

Technical capabilities

01

SOP Ingestion & RAG

Upload procedure documents; they are split into knowledge items, embedded into pgvector, and retrieved as context so answers cite the source paragraph instead of inventing one.

02

Spaced Repetition (FSRS)

Cards are scheduled by FSRS, not by a fixed interval. A person reviews what they are about to forget, on the day they are about to forget it.

03

Auto-generated Assessment

Quiz questions and flashcards are generated from the SOP itself, so a new procedure becomes trainable material the day it is written.

04

Scenario Practice

A practice chat mode plays the difficult customer. The trainee handles the case, and the response is scored against the procedure rather than against a vibe.

05

Gap Analysis

Analytics show the forgetting curve per person and per topic, and flag which SOPs the team consistently gets wrong - usually a sign the document is unclear, not that the staff is careless.

06

Choose Your Own Model

The LLM provider is configuration: OpenAI, Gemini, or a self-hosted Ollama/vLLM endpoint when the procedures must not leave your network.

Processing pipeline

01

SOP Documents

PDF / DOCX

02

Chunk & Embed

pgvector

03

Retrieval

RAG

04

Agent

LangGraph

05

Scheduler

FSRS

06

Analytics

Gap / Retention

Deployment scenarios

Customer Support

Onboarding a Support Agent in Days, Not Months

A support team's quality is bounded by how well the newest person knows the procedures. New hires ask AI Mentor instead of interrupting a senior, practise against generated scenarios, and are reviewed on exactly the procedures they got wrong.

  • Answers cite the exact SOP paragraph
  • Scenario practice with scoring against the procedure
  • Review scheduled by FSRS per person
  • Manager dashboard of team weak spots
  • New SOP becomes training material the same day
RAGFSRSLangGraphOnboarding
Compliance & Regulated Work

Prove the Team Was Trained

When a procedure changes, compliance needs evidence that every affected person saw it and understood it. Assign the document, generate the check, and keep the review log - the record is a query, not a spreadsheet someone maintains by hand.

  • Assign documents to specific learners or teams
  • Review log with timestamps per card
  • Quiz history and scoring per person
  • Self-hosted model option for sensitive procedures
  • Retention statistics per procedure
ComplianceAudit TrailSelf-hosted LLMAnalytics
Training Providers

A Tutor Layer Over Existing Course Material

Training centres already own the material; what they lack is one-to-one attention between sessions. AI Mentor sits on top of the existing library and gives every learner a tutor that knows what they answered wrong last week.

  • Library of documents shared across cohorts
  • Per-learner progress and retention view
  • Team comparison for instructors
  • Streaming chat over WebSocket
  • Runs fully on your own infrastructure
EducationpgvectorWebSocketSelf-hosted

Tech stack

FastAPINext.jsPostgreSQLpgvectorRedisCeleryLangGraphFSRSOpenAIOllamavLLMDocker

Related projects