M2 Internal Adoption + Design Partners

Due Apr 30, 2026

Planned9 / 77 (12%)

Exit criteria (0/4)

• ☐ Multiple internal teams using product weekly.
• ☐ 3-5 influencer companies active as design partners.
• ☐ Weekly feedback loop closes top blockers quickly.
• ☐ Pricing/packaging tested with real buyer conversations.

KPI targets (0/3)

• ☐ Activation trend improving week-over-week.
• ☐ Internal + design partner WAU stable or growing.
• ☐ Reliability metrics within pre-launch thresholds.

Tasks by Workstream

dft core (Sr Engineer Architect)

the YAML contract, compiler/normalizer, execution adapters, and release/versioning is hardened enough for regular use by multiple internal teams and initial design partners, with a predictable response loop for issues and requests. Quality expectations are documented, and prioritized improvements from real usage are actively incorporated into delivery.

• ☐ Implement YAML versioning and migrations — Add built-in schema version handling and migration workflow for dashboard YAML changes.
• ☐ Adoption hardening for internal teams — Harden YAML contract and normalizer for repeated use across multiple internal teams and first design partners.
• ☐ Design-partner feedback loop operations — Operationalize rapid feedback-to-fix loop for execution/runtime adapters with explicit decision logs.
• ☐ feat: chart decisions Phase 5 — YAML annotation overrides — Support YAML annotation overrides for chart decisions so analysts can steer defaults without patching generated specs.
• ☐ Quality standards and guardrails — Define and enforce quality standards for versioning and migrations to keep output consistent as contributors expand.
• ☐ refactor: Consolidate config system - single source of truth in YAML — Consolidate config loading into one YAML-driven source of truth to eliminate conflicting runtime settings paths.
• ☐ Return normalized Dataface format for JSON export (not Vega-Lite specs) — Make JSON export return normalized Dataface spec objects instead of raw Vega-Lite so tooling can round-trip safely.
• ☐ Add settings: attribute pattern for all chart types — Standardize a settings attribute pattern across chart types so YAML is consistent and easier to generate.
• ☐ On-demand result-set profiling in chart decisions pipeline — Expand the ColumnProfile in decisions.py _profile_columns to compute richer statistical characteristics on-demand from…

cloud suite (UI Design and Frontend Dev)

hosted user experience for onboarding, sharing, collaboration, and account/project flows is hardened enough for regular use by multiple internal teams and initial design partners, with a predictable response loop for issues and requests. Quality expectations are documented, and prioritized improvements from real usage are actively incorporated into delivery.

• ☐ Suite git branch and sync workflow for pilot teams — Enable pilot teams to branch, sync, and reconcile dashboard changes with git from Suite without manual backend interven…
• ☐ Adoption hardening for internal teams — Harden workspace and onboarding UX for repeated use across multiple internal teams and first design partners.
• ☐ Design-partner feedback loop operations — Operationalize rapid feedback-to-fix loop for sharing and collaboration surface with explicit decision logs.
• ☐ GitHub integration auth for repo sync — Add GitHub OAuth or GitHub App based integration authorization for Cloud Suite repo sync flows on top of the production…
• ☐ M1 Visual ASQL mode in cloud chart editor — Implement SQL/Visual query mode in cloud chart editor, integrating Visual ASQL alongside existing CodeMirror SQL mode.…
• ☐ Quality standards and guardrails — Define and enforce quality standards for account/project lifecycle flows to keep output consistent as contributors expa…
• ☐ Run internal analyst adoption loop — Set up weekly analyst usage review, friction capture, and release response cycle.
• ☐ Suite Git Integration (dbt Cloud Model) — Define and deliver Suite Git integration using a dbt Cloud-style collaboration model for analytics workflows.
• ☐ Suite Okta login via OIDC — Add Okta OIDC login to Cloud Suite as a secondary provider for internal Fivetran access, layered on top of the producti…
• ☐ Suite role sync and permission enforcement follow-ups — Track post-login authorization, role mapping, and audit/drift follow-up work after M1 Google login.

inspect profiler (Sr Engineer Architect)

warehouse profiling, semantic inference, and analyst-facing data context surfaces is hardened enough for regular use by multiple internal teams and initial design partners, with a predictable response loop for issues and requests. Quality expectations are documented, and prioritized improvements from real usage are actively incorporated into delivery.

• ☐ Adoption hardening for internal teams — Harden profiling pipeline for repeated use across multiple internal teams and first design partners.
• ☐ Design-partner feedback loop operations — Operationalize rapid feedback-to-fix loop for semantic inference and context quality with explicit decision logs.
• ☐ Increase profiler semantic coverage — Improve semantic typing and profile output quality used by analysts and agent workflows.
• ☑ Inspector template customization with eject command — Provide an inspector template eject workflow so teams can customize profile UI safely while retaining upgrade paths.
• ☐ Quality standards and guardrails — Define and enforce quality standards for analyst-facing inspector experience to keep output consistent as contributors…
• ☐ Float confidence scores for statistical column characteristics — Replace binary boolean flags in ColumnInspection with float confidence scores 0.0-1.0 for statistical properties: is_se…

mcp analyst agent (Data AI Engineer Architect)

AI agent tool interfaces, execution workflows, and eval-driven behavior tuning is hardened enough for regular use by multiple internal teams and initial design partners, with a predictable response loop for issues and requests. Quality expectations are documented, and prioritized improvements from real usage are actively incorporated into delivery.

• ◔ Add 'describe' or 'text' render output format for AI agents — Add a describe/text render mode so AI agents can request compact textual dashboard outputs instead of visual payloa…
• ☐ Adoption hardening for internal teams — Harden MCP tool execution model for repeated use across multiple internal teams and first design partners.
• ☑ Build text-to-SQL eval runner and deterministic scorer — Build a Dataface text-to-SQL eval harness that runs agent/model prompts against the cleaned benchmark and scores output…
• ☑ Chat-First Home Page - Conversational AI Interface for Dataface Cloud AI Agent Surfaces — Replace the current org home page (dashboard grid) with a chat-first interface. The home screen shows existing dashboar…
• ☑ Create cleaned dbt SQL benchmark artifact — Create a reproducible benchmark-prep step that imports the raw dbt dataset from cto-research, filters out AISQL rows, r…
• ☐ Design-partner feedback loop operations — Operationalize rapid feedback-to-fix loop for agent prompt/workflow behavior with explicit decision logs.
• ☑ Embeddable Dashboards in Chat - Inline Preview, Modal Expand, and Save to Repo AI Agent Surfaces — Dashboards generated during chat conversations can be embedded inline as interactive previews. Users click to expand in…
• ☑ Extract shared text-to-SQL generation function Benchmark-Driven Text-to-SQL and Discovery Evals — Extract a shared generate_sql(question, context_provider, model) function, wire render_dashboard and cloud AIService to…
• ☑ MCP and skills auto-install across all AI clients — Expand dft mcp init to cover VS Code, Claude Code, and GitHub Copilot Coding Agent. Register MCP server programmaticall…
• ☐ Quality standards and guardrails — Define and enforce quality standards for eval and guardrail framework to keep output consistent as contributors expand.
• ☐ Run agent eval loop with internal analysts — Establish repeatable agent-level eval workflow that tests the full loop (prompt → tool use → SQL generation → dashboard…
• ☐ Set up eval leaderboard dft project and dashboards Benchmark-Driven Text-to-SQL and Discovery Evals — Create a dft project inside the eval output directory with dashboard faces that visualize eval results as a leaderboard…
• ☑ Terminal Agent TUI - dft agent — Build a Claude Code-like terminal AI agent as a dft subcommand. The agent comes pre-loaded with Dataface MCP tools and…
• ☐ Add catalog discovery evals derived from SQL benchmark — Adapt the dbt SQL benchmark into search/catalog discovery eval cases by extracting expected tables from gold SQL and ge…
• ☐ Add persistent analyst memories and learned context AI Quality Experimentation and Context Optimization — Design and implement a memories file that accumulates knowledge from analyst queries — table quirks, column semantics,…
• ☑ Chat Conversation Persistence and History AI Agent Surfaces — Add ChatSession and ChatMessage Django models so chat conversations survive page refreshes. Show recent conversations i…
• ☐ Curate schema and table scope for eval benchmark AI Quality Experimentation and Context Optimization — Decide which schemas, tables, and data layers (raw, silver/staging, gold/marts) to include in the eval scope and catalo…
• ☐ Persist eval outputs for Dataface analysis and boards — Define the canonical eval artifact schema for run metadata, per-case results, retrieval results, and summaries. Add loa…
• ☐ Run context and model ablation experiments AI Quality Experimentation and Context Optimization — Define and execute the initial experiment matrix using the eval system. Compare models (GPT-4o, GPT-5, Claude Sonnet, e…

ft dash packs (Data Analysis Evangelist and AI Training)

connector-specific dashboard packs and KPI narratives for Fivetran sources is hardened enough for regular use by multiple internal teams and initial design partners, with a predictable response loop for issues and requests. Quality expectations are documented, and prioritized improvements from real usage are actively incorporated into delivery.

• ☐ Adoption hardening for internal teams — Harden connector pack coverage for repeated use across multiple internal teams and first design partners.
• ☐ Design-partner feedback loop operations — Operationalize rapid feedback-to-fix loop for dashboard narrative quality with explicit decision logs.
• ☐ Quality standards and guardrails — Define and enforce quality standards for pack publishing workflow to keep output consistent as contributors expand.

ide extension (Head of Engineering)

analyst authoring workflow in VS Code/Cursor with preview, diagnostics, and assist is hardened enough for regular use by multiple internal teams and initial design partners, with a predictable response loop for issues and requests. Quality expectations are documented, and prioritized improvements from real usage are actively incorporated into delivery.

• ☐ Adoption hardening for internal teams — Harden editor + preview workflow for repeated use across multiple internal teams and first design partners.
• ☐ Copilot + MCP dashboard and query workflow in extension — Enable a reliable Copilot-driven flow in the extension where MCP tools can generate and iterate on dashboards and SQL q…
• ☐ Design-partner feedback loop operations — Operationalize rapid feedback-to-fix loop for IDE diagnostics and guidance with explicit decision logs.
• ☐ Quality standards and guardrails — Define and enforce quality standards for inspector/agent integration in IDE to keep output consistent as contributors e…
• ☐ Unify IDE preview and inspector on one extension-managed dft serve — Design and implement a single extension-managed dft serve runtime for both preview and inspector so the IDE stops mixin…

graph library (Data Visualization Designer and Engineer)

visual language, chart defaults, interaction behavior, and differentiated styling is hardened enough for regular use by multiple internal teams and initial design partners, with a predictable response loop for issues and requests. Quality expectations are documented, and prioritized improvements from real usage are actively incorporated into delivery.

• ☐ Lock visual language v1 with RJ — Finalize differentiated chart styling system, typography, and interaction defaults with RJ Andrews.
• ☐ Adoption hardening for internal teams — Harden visual language system for repeated use across multiple internal teams and first design partners.
• ☐ Design-partner feedback loop operations — Operationalize rapid feedback-to-fix loop for chart default behavior with explicit decision logs.
• ☐ Quality standards and guardrails — Define and enforce quality standards for interaction/accessibility polish to keep output consistent as contributors exp…
• ☐ Terminology standards and synonym handling — Define strict product terminology for chart and visualization concepts while preserving backward-compatible interpretat…
• ☐ M2 research perceptual graphic-emphasis analysis beyond contrast — Research whether Dataface should develop a chart-analysis tool that goes beyond standard accessibility contrast checks…

context catalog nimble (Data AI Engineer Architect)

context schema/catalog contracts and Nimble enrichment flows across product surfaces is hardened enough for regular use by multiple internal teams and initial design partners, with a predictable response loop for issues and requests. Quality expectations are documented, and prioritized improvements from real usage are actively incorporated into delivery.

• ☐ Adoption hardening for internal teams — Harden context schema model for repeated use across multiple internal teams and first design partners.
• ☐ Design-partner feedback loop operations — Operationalize rapid feedback-to-fix loop for context enrichment rules with explicit decision logs.
• ☐ feat: chart decisions Phase 4 — SQLGlot column lineage Layer 6 Relationship Mapping — Implement SQLGlot column lineage integration to enrich chart decisions with column-level dependency context.
• ☐ Quality standards and guardrails — Define and enforce quality standards for cross-surface context contract to keep output consistent as contributors expan…
• ☐ Stabilize context catalog schema v1 — Finalize context schema contracts and integration points across inspect, generation, and agent flows.
• ☐ Static semantic type propagation through SQL queries via SQLGlot — Use SQLGlot Expression.meta to propagate profiler-detected semantic types like CURRENCY, EMAIL, CREATED_AT through SQL…

dashboard factory (Data Analysis Evangelist and AI Training)

repeatable process for producing, reviewing, and publishing quickstarts/examples is hardened enough for regular use by multiple internal teams and initial design partners, with a predictable response loop for issues and requests. Quality expectations are documented, and prioritized improvements from real usage are actively incorporated into delivery.

• ☐ Adoption hardening for internal teams — Harden template production pipeline for repeated use across multiple internal teams and first design partners.
• ☐ Artifact digest and Slack sharing workflow — Design a durable artifact review workflow that publishes shareable PNG/SVG dashboard snapshots with brief commentary to…
• ☐ Define dashboard quality rubric v1 — Create rubric/checklist for quickstarts/examples used for internal and design-partner review.
• ☐ Design-partner feedback loop operations — Operationalize rapid feedback-to-fix loop for quality rubric + review process with explicit decision logs.
• ☐ Quality standards and guardrails — Define and enforce quality standards for publication throughput operations to keep output consistent as contributors ex…
• ☐ Define dashboard reference boundary and canon strategy — Decide which third-party dashboard artifacts stay external, which lessons should be distilled into Dataface guidance, a…
• ☐ Study dashboard composition references and extract reusable lessons — Define a repeatable way to study admired dashboard compositions larger than single charts, collect reference artifacts,…

infra tooling (Sr Engineer Architect)

developer tooling, cbox runtime reliability, and deployment execution safety is hardened enough for regular use by multiple internal teams and initial design partners, with a predictable response loop for issues and requests. Quality expectations are documented, and prioritized improvements from real usage are actively incorporated into delivery.

• ☐ Master Plans CLI next-stage guidance command — Add an advisory plans task check command that inspects a task's narrative sections, reports which are incomplete, ide…

integrations platform (Head of Engineering)

deployment, billing, 5T connectivity, and operational reliability/launch integration is hardened enough for regular use by multiple internal teams and initial design partners, with a predictable response loop for issues and requests. Quality expectations are documented, and prioritized improvements from real usage are actively incorporated into delivery.

• ☐ Define pricing and packaging model — Establish pricing tiers, packaging constraints, and internal approval for launch packaging.
• ☐ Finalize product name and procure domain — Decide final product name, verify legal/domain availability, and secure primary domain + redirects.
• ☐ Adoption hardening for internal teams — Harden platform deployment/integration path for repeated use across multiple internal teams and first design partners.
• ☐ Design-partner feedback loop operations — Operationalize rapid feedback-to-fix loop for billing/connectivity operations with explicit decision logs.
• ☐ Quality standards and guardrails — Define and enforce quality standards for reliability + launch operations to keep output consistent as contributors expa…