Answers up front.

MyDataWork connects analytical assets to the business outcomes they support. It gives practitioners a place to organize work across existing tools, and gives leaders context for discussing priorities, dependencies and recorded value.

Start with connected assets, link them to a use case, record stakeholders and objectives, and review the relationships and value behind the work. Portfolio exports carry selected use cases into a wider discussion.

MyDataWork supports analysts, analytics engineers, team leads and leaders across commercial organizations and government agencies. It connects analytical work to business or agency outcomes, preserving the purpose, relationships and recorded value needed for handoffs, reviews and planning.

Yes. In the workspace, open your account menu at the top right and choose Take the guided tour. It explains how to connect work, frame a use case, link assets, review the Dashboard and explore an AI action. You can follow along or move through the explanation, skip it and start it again later. The header Help button opens the Help page, where you can also start the guided tour.

No. A modern data catalog centers on the governed data estate: data assets, definitions, lineage, quality, policy, access and increasingly business and AI context. MyDataWork centers on the analytical work practitioner: the use case, cross-tool assets, stakeholders, practical dependencies, activity, outcomes and value.

An analyst may use a catalog as a trusted data resource while using MyDataWork as the workspace around the work itself. They overlap by design; the difference is what each system is built around.

Not quite — though the comparisons are understandable. MyDataWork shares some surface-level characteristics with personal knowledge management tools, project trackers, and collaborative documentation platforms. But it is designed around a specific problem none of those tools were built to solve.

vs. personal knowledge management tools: the fundamental difference is that MyDataWork knows what your files actually are. Rather than asking you to describe your work in notes and build your own structure, MyDataWork reads the metadata from your actual working files — SQL scripts, Excel models, Python notebooks, Alteryx workflows, Tableau workbooks, and cloud assets — and builds the asset catalog and lineage map automatically.

vs. project trackers like Jira: trackers are excellent at tasks, tickets, and sprints, but they don't understand the file-level context behind data work — which Excel models feed which dashboards, which SQL scripts power which reports. MyDataWork captures that automatically from the files themselves, and integrates with Jira directly so teams don't have to choose between the two.

vs. documentation platforms: wikis are passive — someone has to write and maintain everything. MyDataWork is active: it detects relationships between files automatically, flags external dependencies that aren't yet tracked, surfaces outliers in reported values, and uses AI to identify automation and modernization opportunities.

What it's actually competing with: most analysts have no dedicated system for this layer of their work, so it lives in folders, spreadsheets, tickets, documents, messages, naming conventions and memory — context lives in spreadsheets, email threads, and institutional memory that walks out the door. MyDataWork is built for that majority: something that works immediately, understands their toolchain, and connects their work to business outcomes without becoming a project in itself.

When analysts document their use cases and connect them to the files and data sources they depend on, they are implicitly documenting which data platforms, datasets, and tools are generating real business value. Aggregated across a team, that tells a compelling story: which systems are actively used, what decisions they inform, and where investment is paying off.

For IT leadership, data governance teams, and platform owners, MyDataWork provides something they rarely have: bottom-up evidence of data utilization and business impact, built organically from the daily work of the analysts who depend on their systems. This supports budget justification, governance reporting, and strategic planning adding bottom-up work evidence that top-down inventory and governance views may not provide on their own.

Architecture & Value is a leadership view — on Team and Enterprise plans — that rolls your analytics estate into one picture, built entirely from what your team already tracks in MyDataWork: your tools, your use cases, the value you’ve recorded, and how heavily each platform is used. No new data collection, and no AI credits — it’s deterministic.

Four lenses: a Capability map (value delivered by each business function), a Tool portfolio (every platform placed by the value it supports against how heavily it’s used, on a TIME-style Invest / Migrate / Tolerate / Eliminate quadrant — positioned relative to your own estate, a lens for discussion rather than a directive to cut or migrate), Dependencies (where value concentrates and which platforms are load-bearing), and Change impact (pick any tool and see the value and use cases that depend on it before you touch it).

Architecture & Value gives workspace administrators a leadership summary of recorded value and relationship evidence. Administrators can also inspect individual member records through the audited Workspace view. The summary supports decisions; it does not audit the figures, measure all runtime usage or establish a privacy boundary against administrators.

No. MyDataWork is designed to work alongside the tools your team already uses — Alteryx, Tableau, Power BI, Python, SQL, Excel, and others. It does not replace them or require any changes to how your team works today. Analysts continue working in the tools they know; MyDataWork simply helps them organize, document, and communicate the context around that work.

MyDataWork connects to twelve cloud sources and to local files.

Cloud sources (Setup → Connect your data → Cloud Sources; one-click setup, no installation, read-only credentials):

• Data warehouses — Snowflake, BigQuery, Redshift (read INFORMATION_SCHEMA / metadata APIs for catalog, schema, table, and column structure), and Databricks (parses notebook source for table references).
• Repositories & transformation — GitHub (file source parsed for tool type and table references) and dbt Cloud (latest run manifest.json for models, sources, and the dependency graph).
• BI & analytics — Power BI (read-only Scanner API), Tableau (Metadata API), Sigma, Looker, and ThoughtSpot — cataloging dashboards, workbooks, and datasets plus their source-table lineage.
• Data science — Dataiku (projects, datasets, and recipes with dataset table references).

Local files via the Windows Connector. Recognized types: Excel (.xlsx/.xlsm), CSV, SQL, Python (.py/.ipynb), Tableau (.twb/.twbx), Power BI (.pbix/.pbit), Alteryx (.yxmd/.yxwz/.yxmc), Dataiku exports (.zip), ThoughtSpot (.tml), and Looker LookML (.lkml).

Many other tools — Hex, DataRobot, Fivetran, Matillion, Coalesce, Prophecy, Datameer, Airbyte — write their outputs as tables into Snowflake, Databricks, BigQuery, or Redshift; once the warehouse is connected, those outputs are cataloged with no tool-specific integration. Tools that keep data only in their own backend (Smartsheet, Airtable, Monday, Notion, Salesforce, HubSpot, NetSuite) need to be replicated to the warehouse first. MyDataWork does not currently offer direct connectors for Google Drive, OneDrive, SharePoint, or Dropbox — files kept there are picked up when their sync folders sit on your local disk and are scanned by the Windows Connector.

MyDataWork builds cross-tool lineage from extracted references and file matches.

Shared source references provide evidence that assets are related. A shared table name alone does not prove which asset feeds the other. Inspect the relationship type, confidence and evidence shown in Lineage before treating a connection as a directional dependency.

File-level matches. When an Alteryx workflow or notebook references a local file also in your catalog, MyDataWork can link the matching assets as a Likely dependency.

Files can participate in the graph when their extracted references support a relationship. Inspect the evidence shown for each connection and add manual links where appropriate. A confidence label describes the relationship evidence, not the accuracy of an attached value estimate.

In Setup → Connect your data → Add cloud source, connect a supported BI tool to catalog its own dashboards, workbooks or datasets and available source references. A warehouse connection catalogs the underlying warehouse objects. Those are complementary views; inspect the displayed relationship evidence rather than assuming a shared table proves a directional dependency.

Any file-based data asset your analysts work with: Excel and CSV files, SQL scripts, Python notebooks, Alteryx workflows and macros, Tableau and Power BI workbooks, ThoughtSpot TML files, Dataiku DSS exports, Looker LookML files, and supporting documentation. Assets are connected to the use cases they support, the people who own them, and the lineage that shows how they relate to each other.

Yes. The MyDataWork Connector is a lightweight Windows application that indexes the files on your computer or network drive. It parses your files on your own machine to capture structure — file names, paths, sizes, creation and modification dates, plus tool type and schema-level references such as table, column, and sheet identifiers — and syncs only that structural metadata to your workspace. This local scanning path sends extracted metadata rather than source file contents or data values; manual uploads and cloud connections use different processing paths. The Connector is digitally signed (code-signed) by MyDataWork, LLC through a public certificate authority (Azure Trusted Signing), so Windows verifies the publisher's identity when you install it.

MyDataWork automatically filters files so you focus on current work while keeping everything accessible. Four options: Current (30 days, the default — your active work), Recent (30–90 days), All files (complete catalog, no filtering), and Archived (90+ days). You see the most relevant files first while retaining access to your full asset history.

Yes. In the Assets tab, any data source detected in a file that doesn't match a cataloged asset is flagged with an amber “Not in catalog” badge — for example a SQL script referencing an untracked database table. The Lineage tab surfaces these as ghost nodes (dimmed, dashed) automatically, and the Assets-tab preview has a “Show external dependencies” toggle. When exporting your portfolio you can optionally include external dependencies in the lineage diagrams — useful when sharing with IT teams who need the full dependency picture.

Yes — significantly. When an analyst leaves or a project changes hands, the context behind recurring work — which files matter, why they exist, what decisions they support — is often lost. MyDataWork captures that context as a natural byproduct of daily work, making onboarding faster and reducing dependence on institutional memory held by individuals.

In Setup, download and run the Windows Connector if it is not installed, or use the offered update if needed. Return to Setup in your workspace, choose Connect this computer, and approve the desktop prompt naming that workspace. Wait for the page to verify the connection before adding a folder. There is no setup code to type. Installation and connecting a workspace are separate steps.

Choose Add folder, open a drive and navigate to the folder you want, then check the full selected path and choose Add selected folder. MyDataWork previews the file count, main file types and skipped files before you confirm. If the count cannot be completed, it explains that and asks whether to continue. Canceling the picker or declining confirmation adds no folder and starts no scan. Choose a specific work folder rather than an entire drive.

The connector starts scanning automatically after the folder is added. Supported files are parsed on this computer and their structural metadata is sent to the workspace. Watch the scan status for progress, results or an error; adding a folder is not itself confirmation that its results have arrived. Use Scan now when you want to request another scan.

The connector skips common system and tooling folders beneath your selected folder, including Windows, Program Files, ProgramData, package caches, node_modules, virtual environments and .git. A folder you explicitly select inside one of those locations can still be scanned. Review the preview and selected path; these exclusions do not make selecting an entire drive advisable.

No. Removing an asset changes its catalog record, not the file on your computer or the object in a connected source. The record is hidden from ordinary catalog, use-case and lineage views and can be found using the Removed filter in Assets.

Yes. Select individual assets or all results matching your current filter, then review the count in the bulk action and confirmation. Use the Removed filter to find removed records and restore them. Restoration retains the record identity and its stored use-case links. If its folder is still scanned, metadata may have been refreshed while the record was removed.

No. A scan of a still-assigned folder may refresh a removed record’s metadata, but it does not restore the record or make it visible again. Use Restore if you want it back in the catalog.

Stop scanning beside a folder in Setup stops future scanning of that folder on this computer and leaves existing catalog records in place. Remove in Assets hides the selected records and does not unassign the folder. To retire both, stop scanning the folder and remove its unwanted records. View its assets helps you review the records associated with a folder path before removing them.

Folder assignments are specific to each computer and workspace. A second computer does not inherit or replace the first computer’s folders. To change a computer’s assignments, open Setup on that computer; changing this computer’s list does not remotely change another computer’s list. Both may contribute to the same workspace. Access to the resulting records follows the workspace’s normal permissions.

Repeated ingestion of the same source path updates its existing record. Identical content at different local or cloud paths can remain separate records with their source information; content alone does not establish that two records are the same asset.

The Windows Connector is digitally signed by MyDataWork, LLC. Signing identifies the publisher; it does not guarantee that every browser or computer will accept a download without a warning. If a warning appears, use the download help in Setup to check the publisher and published file fingerprint, or contact contact@mydatawork.com. Do not disable security protection. If the file or publisher does not match, do not run it.

Structure, never values. For a warehouse table we hold that the table exists, what it’s called, where it lives, that it has a column named order_date of type DATE and one named region of type VARCHAR, when it last changed, and what the analyst wrote in the column comment. We do not hold any of the dates in that column, or any of the regions — not one row, not a sample, not a summary, not a count of distinct values.

Column names and types are captured from Snowflake, BigQuery, Redshift, Databricks (Unity Catalog), Power BI, Tableau, Looker, Sigma, ThoughtSpot, dbt, Dataiku, Excel and CSV headers, and Parquet/Avro footers. Names and types only.

Where a source can’t publish something, we say so rather than leaving a blank. A CSV declares column names but no types, because the format has none — that’s a fact about CSVs, not a gap in your catalog. Tableau field metadata beyond published data sources needs Tableau’s Data Management add-on. Setup → Metadata shows this per source, alongside how many assets carry column schema and when each source was last refreshed.

Local files. If you install the desktop connector, it opens files on your own machine to read their structure — sheet names, column headers, and the tables a query references — and sends us only that structure. The service never receives your file. We say “we never receive or store your contents” rather than “we never open a file,” because the second wouldn’t be true.

The catalog stores extracted metadata and the context you add, which can itself be sensitive. The Windows Connector parses supported files locally and sends extracted metadata. Uploads and cloud connections are different: the service receives files, source definitions or metadata to extract structure. Source file contents and business-data rows are not retained as catalog content. Review Security & IT for processing, access and sharing boundaries.

Cloud connections fetch metadata or source definitions to extract structure: for example GitHub file content, Databricks notebook source, dbt manifests, warehouse INFORMATION_SCHEMA and read-only BI metadata. Source material is parsed and discarded; the catalog retains extracted structure rather than source file contents or business-data rows. See Security & IT for the connector-by-connector details.

AWS GovCloud hosting does not confer a government authorization on MyDataWork. MyDataWork does not claim FedRAMP or StateRAMP authorization, CMMC certification or ITAR compliance, and no agency ATO is implied. The service is not intended for classified information or Controlled Unclassified Information (CUI). Do not paste CUI, source data rows, credentials or file contents into workspace text fields.

MyDataWork organizes the structure of your work, not the data inside it — so the personal information it holds is limited and purposeful.

What it holds: your account details (name, email, and a securely hashed password), the names and emails of any team members an admin adds, and any stakeholder names or emails you choose to record against a use case.

Metadata can contain sensitive identifiers, and files must be read to extract their structure. The local Connector performs that extraction on your computer; uploads and connected cloud sources are processed by the service. Only add metadata and authored context that you are permitted to disclose. See Security & IT Questions for the processing details.

How it is protected: passwords are hashed, and workspace databases are encrypted at rest and browser connections use HTTPS. Stripe handles direct-plan payments; AWS handles Marketplace billing. Free-assessment eligibility uses an email hash. Explorer’s one-trial-per-email limit is also enforced using a one-way email hash retained after account deletion, as explained in our Privacy Policy.

AI features use authorized metadata and the text you enter around your work. Dedicated stakeholder fields are filtered, but names or confidential details may still appear in free text or metadata; review that context before using AI. Workspace administrators can export workspace JSON through Setup → Data portability. Available exports and deletion controls depend on your role and account state.

The GovCloud application, workspace database and built-in AI processing run in AWS GovCloud (US-West). Marketplace registration and agreement events pass through a relay in commercial AWS. Transactional email delivery and recipients’ mail systems are outside the GovCloud workspace boundary. Optional cloud connections, Jira and external agents communicate with the destinations you authorize. GovCloud hosting therefore does not mean every processing path stays within GovCloud.

In the Objectives & Progress tab of any use case you set three values: a baseline (where things stood before the work began), a current value, and a target. MyDataWork calculates progress automatically and shows a progress bar — no formula needed. The baseline is the most important value to set correctly, and it can only be truly accurate once, at the very beginning.

Progress uses professional value-measurement templates that go beyond monetary estimates: Time Savings, Quality Improvement, Efficiency Gains, Risk Reduction, and Cost Efficiency. Each provides recommended units, best-practice guidance, realistic improvement ranges, and industry-standard measurement approaches — a live, honest record ready to share at any moment.

Three actions: Archive (soft delete — removes a use case from active view while preserving all data and history), Restore (brings archived use cases back), and Delete (permanent removal). Switch between Active and Archived tabs in the Use Cases section. Useful for completed projects you want to keep for reference, seasonal work, or keeping a clean active workspace while preserving institutional knowledge.

My Work is your landing view for current work. The Dashboard summarizes recorded workspace information, including:

• Use case staleness — freshness buckets (under 30 / 30–60 / 60–90 / 90+ days).
• Value totals — estimated and recorded realized value for use cases with linked assets. Portfolio can include selected use cases without linked assets, so its totals can differ.
• Stakeholder coverage — how many active use cases have an assigned stakeholder.
• Value distribution — realized values against the workspace mean and median.
• Tool distribution — asset count by tool family.
• Workspace Agent activity — findings surfaced and acted on in the last 14 days.

Dashboard value totals cover use cases with linked assets. Portfolio totals cover the use cases selected there, whether or not they have linked assets, so the totals can differ. Check the selected scope and the distinction between estimated and recorded realized value. The Dashboard’s demo/real toggle applies to that view only.

Yes. Use Assets to browse and search your catalog, and My Work to return to the use cases you are working on. Linking an asset to a use case records the context for that work; it does not change the asset’s underlying source or grant new access to it.

An AI-powered review of your workspace, runnable on demand from Setup → Asset Estate Assessment. The first run is free on every plan — one free assessment per email, forever. Subsequent runs use AI credits like other AI features.

What it produces: a five-section report — Estate overview, Connections worth making explicit, External dependencies, Use case opportunities, and Health observations — plus a ranked “Concrete next actions” list. Exportable to PDF.

The Asset Estate Assessment uses authorized asset metadata and use-case context to suggest areas for investigation. Review its findings against your own knowledge of the work; it is not an independent audit of source data.

Both are AI-powered and can be scoped to your whole workspace or a specific client/project — but they do different jobs.

The Workspace Agent is a detector. It runs rule-based checks and surfaces discrete, specific findings worth acting on — a running worklist you run any time. The Asset Estate Assessment is an analyst. It produces a point-in-time synthesis — a connected, narrative review of your whole estate's health. In short: run the Agent for an ongoing to-do list; run the Assessment for a periodic, comprehensive read.

Portfolio provides value summaries to help you review recorded figures. Its mean and median realized-value comparison excludes zero and unrecorded amounts; read the displayed population. These summaries do not verify or audit an estimate or outcome.

Yes. Each use case has a Notes URL field on the Overview tab. Paste a Google Doc, Notion page, Confluence URL, or any HTTPS link. When set, an “Open notes” button appears and a small indicator shows next to the use case in the list view.

Solution Packs provide editable starter use cases you adapt to your work. Choose a scenario in Setup → Solution Packs, then add your objectives, stakeholders and supporting assets. They use no AI credits.

The commercial service includes scenarios for FP&A, Power BI portfolio review, agentic readiness, consulting discovery, Alteryx modernization and value realization. The GovCloud offering includes Agency Operations & Public Outcomes, with agency-focused starting points. Available packs are shown in your workspace. Example figures are starting assumptions, not measured results.

A project or program that groups related use cases — the use-case counterpart of an asset group. Initiatives are personal to you, and each use case belongs to one initiative, so dashboard value roll-ups never double-count. Manage them on the Use Cases tab's By initiative view, and filter the Dashboard, Portfolio, and Insights to a single initiative.

Use asset groups to organize assets and initiatives to organize use cases. Where a screen offers a scope selector, choose the relevant project and verify the population shown before running an analysis or exporting. These groupings are not confidentiality boundaries. Use separate workspaces where client information must be kept apart, and review the displayed credit cost for each action.

Workspace administrators can export supported records as JSON from Setup → Data portability. The export covers assets and relationships, stakeholders, use cases and recorded context, saved recommendations, action plans and other supported planning records.

It is not a complete system backup: source files, connection credentials and configuration, audit history, Assessment reports and Workspace Agent finding history are not included. Import requires administrator access to an eligible active workspace and is subject to plan limits. It merges supported records; review the destination because matching records can be updated.

Portfolio PDF and PowerPoint exports communicate selected work within your permitted scope. Explorer exports remain available during the read-only period before deletion.

No. AI is enabled by default for new workspaces, with controls in Setup → AI Assist. AI usage draws from your plan’s allowance. Direct plans use per-person daily allowances and, on paid plans, weekly limits; AWS Marketplace uses a shared monthly pool.

Commercial workspaces use OpenAI for built-in AI. GovCloud workspaces use Amazon Bedrock in AWS GovCloud (US-West). AI can use relevant metadata and the context you write; see AI processing and access details.

Explorer includes 3/day, capped at 60 across its trial. Solo includes 5/day and 25/week. Direct Team includes 20/day and 100/week per member. Credit costs depend on the action; check the displayed cost before running it. Solo and direct Team administrators can buy shared packs of 10 for $5, 25 for $10 or 50 for $15. Purchased credits do not expire. Explorer and AWS cannot buy these packs.

The commercial AWS package includes 9,000 pooled AI credits per month; GovCloud includes 18,000. Each pool belongs to the workspace and resets on the first calendar day of the month at 00:00 UTC. Unused included credits do not roll over. Neither AWS package offers self-service credit packs. Operational safeguards may pause AI before the credit pool is exhausted; non-AI work continues within plan limits.

Propose Use Cases (1 credit each) generates specific, actionable proposals from your actual files and workflows.

Leverage has five modes: Find reuse opportunities; Identify automation candidates; Discover marketplace data; Migration Assist; and Public datasets. Use these to explore overlap, modernization options and external data relevant to your work, then review the suggestions before acting.

AI Recommendations use use-case descriptions and linked assets to suggest improvements. Workspace Agent reviews the workspace when you run it and surfaces items worth attention. Open MyDataWork Assistant from the Assistant button in the page header.

Run a review from the Workspace Agent tab using “Analyze.” It checks recorded work for findings such as:

• Cleanup — use cases with assets and progress but no stakeholder; active use cases not updated in 90+ days.
• Activity — recently-added assets (last 14 days) not yet linked to a use case (with a suggested best fit); assets you removed in the last 30 days still referenced by an active use case or lineage edge.
• Insight — high-value use cases with low realized value not updated in 60+ days; assets linked to three or more active use cases (your “hidden infrastructure”).

Each finding includes a one-click deep-link. You can dismiss with a reason or give a thumbs up/down, and the agent auto-resolves a finding when its condition no longer applies. The checks are rule-based; AI is used only to phrase each finding in plain language. You get up to two free attempts; after that, 1 credit per finding, capped at 8 per run.

Yes. The Leverage tab's “Identify automation candidates” mode uses AI to surface specific processes that could be candidates for intelligent automation. Each opportunity includes what could be automated, which assets are involved, an estimated value and complexity, and a concrete next step. You can share any opportunity by email directly from the app, and all opportunities are saved so you can revisit them without regenerating.

Yes, and that is the point of it. There is a great deal of genuinely useful public data — inflation, employment, weather, energy prices, demographics, company filings — published free on open APIs. What stops teams using it is the step before the work: knowing which of thousands of published series is about your problem, and whether your data has a column it could be joined on. MyDataWork answers that from your metadata in seconds, so the sources worth using actually get found. It does not ingest public data or blend it into yours.

The Leverage tab’s External data mode scans a use case and suggests free public datasets that may improve it — FRED, U.S. Census, BLS, NOAA, EIA, SEC EDGAR and World Bank. It works on every plan including the free Explorer trial, and needs no cloud provider configured.

Each suggestion carries a match strength — Strong match, Worth testing or Speculative — and a plain-language reason that cites the specific assets it read. The tier describes how well the source fits the shape of your work. It does not claim the enrichment will make your results better — only running it can show that.

You can preview a source, take a connection recipe for your own platform, and accept the ones worth keeping. An accepted source is cataloged as an external-dataset asset with an enriches relationship, so it appears in lineage and change-impact like any other dependency — and it does not count against your plan’s asset limit.

Yes. All seven are included by default, and you can narrow that to any subset — either from the panel in the Leverage tab or in Setup → Public data sources. Both places list what each source covers, how often it updates, and what it takes to get access: FRED, BLS, Census and EIA need your own free API key; NOAA and SEC EDGAR need no key at all, only an identifying email in the request header; World Bank needs nothing.

Your choice is remembered, and narrowing the scan does not discard results you already have — a scan limited to one source replaces only that source’s result.

The matching uses asset metadata and the work context you record, such as a linked use case and its objective. It does not require your underlying business-data rows. AI-assisted suggestions can send relevant metadata and authored context to the AI provider; this is separate from running a connection recipe in your own environment.

The blending happens in your environment, not ours. MyDataWork generates a connection recipe — the API call, a landing table, the suggested join and the attribution the source requires — and you run it in your own warehouse or BI tool, with your own API key. We never execute it and never store the result.

Yes. Before any AI runs, match readiness scores the use case on two things, free and with no credit used: topic match — is there a relevant source at all? — and connection key — is there a shared column, usually a date and often a place, to match the two datasets on?

These fail independently, and the message differs. Weather may clearly suit a demand forecast, but if no asset declares a date or region column the enrichment cannot land, so we tell you that and what to add. Where a use case simply has no external-data shape, such as a reporting-delivery process, we say so plainly and decline the scan instead of returning suggestions that look confident and are not.

You also choose which public sources are scanned — all seven by default, or a subset you pick in Setup or in the tab itself.

Readiness, previews, recipes, accepting and dismissing are all free. Only the scan itself uses an AI credit.

Yes. The Leverage tab's “Discover marketplace data” mode recommends external datasets available through your cloud providers. First set your platforms in Setup → Cloud Providers (AWS, Google Cloud, Snowflake, Microsoft Azure, Databricks); recommendations then include the dataset name, the platform, why it's relevant to your specific use cases, and where to find it. Recommendations are saved across sessions.

Yes. The Leverage tab's Migration Assist mode analyzes your assets or use cases and identifies opportunities to migrate or modernize — for example moving a complex Excel model to Python, replacing an Alteryx workflow with dbt, or converting ad-hoc SQL into structured dbt models. You can analyze individual assets or a full use case end-to-end. Each recommendation includes the current tool, the recommended alternative, migration effort (Low/Medium/High), estimated benefit, key risks, a confidence rating, and a concrete next step. Results are saved and shareable by email.

Yes. Agent access (Setup → Agent access) lets you give an AI agent you choose read-only access to the structure of your data work — your use cases, their linked assets, and the lineage between them — over the Model Context Protocol (MCP), an open industry standard (for example Claude Code, your team's own agent, or another MCP-capable tool).

You create an agent key, choose which use cases it may read, and preview the exposed context. Agent Access shares selected structure and relationships, plus descriptions and objectives you have written. It does not share source file contents or values from connected sources. Authored text can contain names or figures, so review it before granting access.

You choose an access mode per key. Live mode lets the agent read your current context on demand (paid plans, daily read cap; works with MCP clients that accept a bearer key). Export mode lets you download a point-in-time context file and hand it to your agent (the only mode on the free Explorer plan; works with any agent, including Claude Desktop). Every grant change and every agent read is recorded, and you can revoke a key at any time.

Yes. The Agent activity panel at the bottom of Setup → Agent access shows the record: how many times your agents have read your context over the last 7, 30, or 90 days, a per-day chart, which key each read used, and which tool the agent called — plus a recent-activity feed that also covers your own grant changes, previews, and exports.

A key you delete keeps its history, so removing a key does not erase the record that it read your work. Day buckets are UTC, matching the daily read cap, so the “today” figure always agrees with the “reads left today” number shown at the top of the panel.

You see your own agent activity only. In a Team workspace, other members see theirs — a key exposes only its owner’s context, and the activity record follows the same boundary.

No. Agent reads (in Live mode) and exports draw on their own dedicated daily allowance — the Live reads and Exports caps shown in the Agent access panel — not your AI credits. The AI credit pool that powers the chatbot, Leverage analyses, and recommendations is never touched by agent activity.

Agent Studio is where you scope an agentic use case before anyone builds it. You catalog the MCP sources your organization exposes, pick the tools an agent could use, arrange them on a simple canvas (trigger, inputs, agent, actions, outcome), and let AI help draft the surrounding definition — the value case, agent persona, reasoning approach, guardrails, and human-in-the-loop checkpoints. The result is a shareable definition you can export as a PDF or PowerPoint.

It deliberately stops short of building or running anything. Agent Studio does not deploy agents, does not call tools, and does not read your business data — when it catalogs an MCP source it reads only what that server advertises: tool names, descriptions, and input-parameter shapes. It also includes an anti-“agent-washing” check: if the work is stable, rules-based, and deterministic, it says so rather than forcing an agent where a script would be better.

They serve different purposes. Agent Access lets an approved agent read selected work context over MCP, including asset structure and authored descriptions and objectives. Agent Studio catalogs advertised tool capabilities so you can design an agentic use case. It designs the workflow; it does not run it. Your team implements actual tool access through its own governed environment.

Team members use individual accounts within a shared workspace. Their own asset records are private from other ordinary members unless shared; workspace administrators can inspect them through the audited Workspace view. Connecting or copying an asset records ownership, but does not grant access to its underlying source file.

Members can share selected assets with colleagues, who can use Copy to mine to create an independent copy. Removing the original share does not remove copies already made. Ordinary members do not directly share use-case records; communicate selected use cases through Portfolio exports. Stakeholders are workspace business contacts, not a private address book. Administrators can inspect individual member records through the audited Workspace view.

A Team workspace has an administrator who manages membership in Setup → Workspace management. The admin can create a member account or link an existing account, subject to seat limits. Linking an existing account preserves its credentials. Administrators can transfer ownership and remove access; removing a member does not delete their contributed workspace records.

Team Metrics summarizes adoption and reuse. That aggregate panel is not the limit of administrator access: admins can also inspect individual member records through the audited Workspace view. Explain both capabilities to members, especially when participants come from different organizations.

If the admin leaves without first transferring admin to another member, the workspace can become orphaned. If that happens, email contact@mydatawork.com and we can reassign admin access. To avoid it, admins should transfer ownership before they leave or are offboarded.

Yes. MyDataWork includes a Jira integration (available on all paid plans) to create structured tickets and push progress updates from your use cases, with professional templates for automation projects, reporting enhancements, and data-quality initiatives. Each ticket includes objectives, stakeholders, linked assets, acceptance criteria, and a link back to your workspace. Set it up in Setup → Integrations with your Jira URL and API token. The integration is one-way — MyDataWork sends to Jira; your workspace remains the source of truth.

Use the web app on mobile to review assets, use cases, lineage and portfolio information. Local-file scanning and file-system browsing require the Windows Connector on a computer. Cloud connections are configured in the web app, subject to your permissions and plan.

Local-file scanning requires the Windows Connector. Supported cloud sources connect through the web app without that local installation. Once cataloged, assets can be viewed from other devices.

The Windows Connector discovers supported local files and network-drive files. Supported cloud sources are a separate discovery path configured in the web app; they do not require the Windows Connector.

Use landscape mode on tablets for detailed asset information and lineage diagrams. Portrait mode works well for portfolio overviews and use case tracking.

Direct plans run on the commercial service: Explorer is free for 90 days (one user, 30 assets, one cloud source); Solo is $50/month or $480/year (one user, 300 assets); Team Starter is $2,100/year (2–5 users, 2,000 shared assets); Team Growth is $3,900/year (6–10 users, 5,000 shared assets). Only direct paid plans include the 14-day card trial. Eligible Solo/direct Team plans can buy credit packs; Explorer and AWS cannot.

Commercial AWS is $7,500 per 12 months for up to 20 users and 25,000 assets. GovCloud is $15,000 per 12 months for up to 40 users and 50,000 assets. Neither AWS offer has a trial. See the full comparison.

Yes. For larger teams we offer negotiated Enterprise plans with custom seat counts and higher asset, AI-credit, and agent-access limits, billed by invoice (no credit card required). Contact contact@mydatawork.com to discuss your needs.

They're two distinct offerings:

• Explorer plan is our 90-day free workspace for individuals. No credit card. 1 seat, up to 30 user-added assets, 1 cloud source, 3 AI credits/day (60-credit total cap). New self-service signups start here by default.
• 14-day free trial applies to our paid Solo and Team plans — the standard “try before you're billed” trial. After 14 days, billing begins on the chosen plan.

Explorer runs for 90 days and is limited to one trial per email address. When your trial ends, your workspace becomes read-only for a further 30 days. That period starts at trial end, not at signup. You can still browse and export your work throughout the read-only period.

Upgrade to a paid plan during the trial or read-only period to keep your work and restore full access. If you do not upgrade by the end of the read-only period, your trial workspace content is scheduled for permanent deletion. Your account and sign-in remain, and you can start a new workspace on a paid plan.

Before deletion, save a JSON export of supported workspace records from Setup → Data portability, or export a PDF or PowerPoint presentation from Portfolio. These exports remain available during the read-only period.

Explorer and direct Solo/Team trials run on our commercial service, not in AWS GovCloud (US). Evaluate with sample or non-sensitive data. A direct trial is separate from any AWS purchase and does not transfer automatically to GovCloud.

For commercial direct plans, register at app.mydatawork.com: Explorer has a 90-day no-card trial, and direct Solo/Team have 14-day card trials. For AWS purchases, use the listing for the intended commercial or GovCloud environment, then complete its registration and email-confirmation flow. AWS packages have no free trial.

Edit your own name in Account settings. The change applies to future uses of your profile; it does not rewrite emails already sent.

Cancellation, stopping renewal, deleting an account and deleting a workspace are different actions. Direct subscription retention follows our Privacy Policy. Shared workspace contributions and backups are handled separately from personal account deletion.

Manage renewal and cancellation through AWS Marketplace under your accepted agreement. Turning off renewal prevents a future renewal; it does not itself end the current contract early. Early termination follows the applicable AWS process and agreed terms. When the agreement actually ends, product access stops. A refund does not itself cancel an agreement, and deleting a MyDataWork account or workspace does not cancel AWS billing.

For an AWS purchase, the accepted agreement governs retrieval and deletion. Where the agreement incorporates the Standard Contract for AWS Marketplace retrieval provision, the retrieval period is 45 days after termination, or another period agreed in writing; this includes the GovCloud agreement. An administrator should contact contact@mydatawork.com to arrange retrieval or deletion after access ends. Data is then deleted within a reasonable time under the applicable agreement through a reviewed process. Other agreements retain their own terms; a website update does not change them.

Use Account settings and read the confirmation. Administrators may need to transfer ownership or remove other members first. If your account belongs to multiple workspaces or self-service deletion is unavailable, contact contact@mydatawork.com. Shared workspace contributions can remain with that workspace. Deleting your account does not cancel an AWS agreement or create a refund entitlement. See our Privacy Policy.

Direct annual plans are non-refundable; direct monthly cancellation ends access at the end of the paid period without a partial-period refund.

AWS Marketplace offers have no free trial and no money-back guarantee. Refund requests are considered case by case: contact contact@mydatawork.com. Approved AWS-billed refunds are issued through AWS Marketplace, subject to the applicable offer and AWS terms. Cancellation and a refund are separate actions.

Check the allowance for your plan: Explorer has a daily and trial-total cap; Solo and direct Team plans have daily and weekly limits. Their eligible paid plans can use purchased credits after included allowances. AWS Marketplace uses a monthly workspace pool that resets on the first at 00:00 UTC, with no top-up packs. Non-AI features remain available within plan limits.

Some features have rate limits to protect platform resources: lineage rebuilds 5/day, Jira integration calls 20/hour (shared across push and progress updates), and exports 10/day. If you hit one, you'll see a message and can retry after it resets. Normal workflows rarely hit these — they exist mainly to prevent accidental runaway usage.

Contact us at contact@mydatawork.com. We aim to respond within 2 business days.

AWS manages the agreement and billing. MyDataWork manages workspace access and members within the package limits. Commercial AWS uses the commercial service; the GovCloud offer uses a separate service in AWS GovCloud (US-West), with Amazon Bedrock for AI. Private offers require separately agreed terms and capacity review.

The commercial AWS package includes 9,000 pooled AI credits per month; GovCloud includes 18,000. Each pool belongs to the workspace and resets on the first calendar day of the month at 00:00 UTC. Unused included credits do not roll over. Neither AWS package offers self-service credit packs. Operational safeguards may pause AI before the credit pool is exhausted; non-AI work continues within plan limits.

Manage renewal and cancellation through AWS Marketplace under your accepted agreement. Turning off renewal prevents a future renewal; it does not itself end the current contract early. Early termination follows the applicable AWS process and agreed terms. When the agreement actually ends, product access stops. A refund does not itself cancel an agreement, and deleting a MyDataWork account or workspace does not cancel AWS billing.

For an AWS purchase, the accepted agreement governs retrieval and deletion. Where the agreement incorporates the Standard Contract for AWS Marketplace retrieval provision, the retrieval period is 45 days after termination, or another period agreed in writing; this includes the GovCloud agreement. An administrator should contact contact@mydatawork.com to arrange retrieval or deletion after access ends. Data is then deleted within a reasonable time under the applicable agreement through a reviewed process. Other agreements retain their own terms; a website update does not change them.