uExcelerate AI Service (1.0.0)

Generates AI-suggested next actions based on user goals + recent activity. Backed by GPT-4 via the OpenAI service module. If the goal value is UNKNOWN or empty, fallback action generation is used (no goal context).

Generates AI-suggested next actions based on user goals + recent activity. Backed by GPT-4 via the OpenAI service module. Returns a list of behavioural skills/competencies inferred from the supplied action description.

Infers user growth areas from assessment + interaction history. Used by the coach to populate the discovery flow. Accepts a learner/tool data payload (assessment scores, low-rated competencies, prior feedback) and returns ranked growth areas with rationale.

AI-powered coach recommendation. Ranks coaches by match score against the learner's profile. Step 1 takes the learner's high-level goal and returns an initial candidate pool plus suggested follow-up questions for step 2.

AI-powered coach recommendation. Ranks coaches by match score against the learner's profile. Step 2 narrows the candidate set using the learner's current work context (role, team, challenges).

AI-powered coach recommendation. Ranks coaches by match score against the learner's profile. Step 3 finalises the recommendation by incorporating the learner's achievements / strengths.

AI-powered coach recommendation. Ranks coaches by match score against the learner's profile. This endpoint maps the learner's free-form statements (typically three discovery-flow answers) to canonical behavioural skill competencies used downstream by the matcher.

Cross-cohort analytics comparing tool/program performance against benchmarks. Supports two auth modes: (1) JWT bearer token (standard in-app calls); (2) encrypted token field in body (for emailed report links — JWT is skipped). When token is supplied, the handler auto-detects PDF tokens (multiple tool IDs) vs single-tool tokens.

AI-generated personalized email content for outreach or follow-up. This endpoint dispatches a support email (to support@uexcelerate.com) with the supplied subject + body + optional attachments.

Growth partner (peer) matching algorithm. Returns ranked candidate peers (growth partners) for the requesting learner, filtered by seniority level, industry, and function/department.

Google Gemini chat backend for in-app coaching conversations. Returns the complete persisted chat history (prior conversation messages, array of {role, content}) for the given user/role pair.

Google Gemini chat backend for in-app coaching conversations. Submits the user's prompt (with role + persisted conversation context) and returns the assistant's reply.

Google Gemini chat backend for in-app coaching conversations. Returns a condensed summary of the user's recent conversation - used to seed context on the chat resume screen.

Goal suggestion and tracking AI. Health-check / liveness endpoint that returns a static greeting - useful for verifying the AI service is reachable through the gateway.

Goal suggestion and tracking AI. Expands a partial goal stem into a full, well-formed SMART-style goal sentence - powers the goal-completion suggestion in the goal editor UI.

Goal suggestion and tracking AI. Generates a ranked list of SMART goal candidates for a learner, given a seed goal/skill set and the learner's leadership level.

Goal suggestion and tracking AI. Returns a SMART-criteria evaluation (specific/measurable/achievable/relevant/time-bound) for the supplied goal text. Endpoint is currently not wired into the UI - retained for future use.

Goal suggestion and tracking AI. Single-word autocomplete helper for the goal entry field. Endpoint is currently not wired into the UI - superseded by /complete-goal.

Goal suggestion and tracking AI. Returns a list of behavioural skills / competencies relevant to the supplied goal. Endpoint is currently not wired into the UI - retained for future use.

Goal suggestion and tracking AI. Token-level next-word predictor used inline as the learner types - lighter weight than /complete-goal, called per keystroke at typing pauses.

OpenAI GPT chat backend (parallel to Gemini for A/B/redundancy). Submits a single user prompt to the assistant (continuing the user's persisted thread) and returns the assistant reply.

OpenAI GPT chat backend (parallel to Gemini for A/B/redundancy). Fetches the active conversation record (thread + metadata) for the given user/role pair - distinct from /chat_history which returns only the message list.

OpenAI GPT chat backend (parallel to Gemini for A/B/redundancy). Provisions a new conversation thread for the given user/role and returns its thread ID, which is then used by /chat.

Individual Development Plan generation. Generates narrative blind-spot insights and a structured development plan, given the learner's identified blind-spot competencies plus the subset they scored lowest on.

AI summarization of free-form coaching notes. Accepts raw notes text and returns a structured summary (key themes, action items, follow-ups).

AI summarization of free-form coaching notes. Auto-detects the file type from the URL's extension and routes to the matching extract+summarise pipeline. Supported extensions: .pdf, .doc, .docx, .txt, .ppt, .pptx.

AI summarization of free-form coaching notes. Performs OCR + summary on a handwritten / scanned note image (e.g. whiteboard photo) and returns the structured summary.

Tool report insight generation - interprets numeric scores into qualitative narrative. Generates coaching reflections (themes, hypotheses, conversation starters) for a learner's tool report, given the tool, learner, and requester IDs.

Tool report insight generation - interprets numeric scores into qualitative narrative. Token-authenticated variant for emailed report links; no JWT required. The encrypted token carries toolId, learnerId, and requestedBy.

AI tool generation (custom assessment / reflection tool builder). Composes a complete assessment instrument (questions + scale) given a target leadership level, competency set, tool category, and response style (defaults to Likert).