Data to AI™ / Gemini/HubSpot
Give RevOps engineers and finance analysts HubSpot deal, stage and owner data in the terminal, so configs, checks and CSVs come from extracted CRM records.
Better answers, fewer tokens.
Gemini CLI
How long did the deals we won in September spend in each stage?
The 23 deals won in September took 31 days on average from qualification to contract. Decision maker bought-in was the slowest stage, at 12 days or 39% of the cycle.
| Stage Label | Deals | Avg days in stage | Share of cycle |
|---|---|---|---|
| Qualified to buy Properties History | 23 | 6 | 19% |
| Presentation scheduled Properties History | 23 | 9 | 29% |
| Decision maker bought-in Properties History | 23 | 12 | 39% |
| Contract sent Properties History | 23 | 4 | 13% |
Days are measured between dealstage change Timestamps in Properties History, with Stage IDs mapped to Stage Label through Deal Pipelines.
HubSpot is a marketing, sales and service platform. Dataddo extracts its CRM records, such as deals, line items, quotes and owners, with custom fields on every object, plus pipeline stages and property change history.
Read Deal Pipelines through the dataddo server and write config/hubspot-stages.json with Pipeline ID, Stage ID, Stage Label and Stage Probability for every stage.
A revenue engineer keeps the stage map the billing service relies on in version control. Gemini CLI regenerates it after sales ops renames a stage, and the diff shows exactly which Stage Label or probability moved.
Compare hubspot_owner_id on open Deals with Owner ID in Owners and list any deal whose owner is missing or not Is Active.
Before shipping a change to lead routing, a developer runs the check in Gemini Code Assist agent mode in VS Code and fixes the orphaned assignments in HubSpot first, so no deal lands with an owner who has left.
List Line Items with an hs_lastmodifieddate after 2026-10-01 and save hs_object_id and hs_lastmodifieddate to reports/line-items-changed.csv.
A finance data engineer diffs the CSV against the invoicing table each week, so line items edited after a deal closed get re-billed or credited instead of surfacing at the quarterly audit.
Data to AI™ is priced per seat, like the AI assistants your team already uses, with ~25% off on annual billing. Need tailored contracts or volume pricing? Talk to us about Enterprise.
Serve governed business data to Claude, ChatGPT and other AI tools over MCP.
$10.49/month
per seat
3 seats minimum
Flexibly deploy Dataddo in any cloud or hybrid environment.
Custom
We have a payment model that works for you
| Without Dataddo |
With Dataddo
|
Outcome for you | |
|---|---|---|---|
| Stage IDs in code | A hand-written map of Stage ID to stage name sits in the repository and drifts silently when someone edits the pipeline in HubSpot. | Gemini CLI reads Deal Pipelines from the latest extraction and rewrites the mapping file, so the change shows up in a normal code review. | Config matches HubSpot |
| Secrets on developer laptops | Each script carries its own HubSpot API token in a .env file that gets copied between machines and projects. | settings.json holds only the httpUrl. HubSpot is authorized once in Dataddo, and the Dataddo sign-in is kept in ~/.gemini/mcp-oauth-tokens.json and renewed automatically. | No HubSpot token in the repo |
| Approving what runs | A one-off script calls the HubSpot API with whatever scope its token happens to have, and nobody reviews each request. | Gemini CLI asks you to confirm each Dataddo tool call, and you can mark a tool such as describe_entity as Always allow while keeping query on confirm. | Every call is visible |
| Token consumption | Gemini spends tokens exploring raw HubSpot columns, sampling rows and retrying until it understands the data. | Field definitions and relations are supplied up front, so the model skips the exploration and goes straight to the answer. | Fewer tokens per answer |
| Answer quality | Prone to wrong joins, invented fields and numbers that are hard to check. | Answers grounded in defined fields. In the Dataddo benchmark, 78.5% of questions were answered correctly, against 65.5% with plain CSV files. | Better answers |
The HubSpot datasets finance and RevOps teams use most, with their real field names. Gemini queries them by name through the Dataddo semantic layer.
| Dataset | Metrics and attributes |
Ask Gemini
|
|---|---|---|
| Deals |
|
List hs_object_id and dealname for every deal with a createdate since Monday, so I can compare them with our staging import. |
| Deal Pipelines |
|
Output each Pipeline ID with its Stage ID values and Stage Probability as JSON for our revenue service config. |
| Properties History |
|
Which Property Name on deals changed most often this month, counted by distinct Object ID? |
| Owners |
|
Return every Owner ID with its Is Active flag, so a script can validate the owner on each deal before a deploy. |
| Line Items |
|
Which hs_object_id values on line items have an hs_lastmodifieddate after our last billing sync ran? |
We asked 43 pre-registered questions of three real datasets - Google Search Console, HubSpot deals and Google Ads - and delivered the same rows to the same model in four ways. Only the data layer changed.
of questions answered correctly - the best of the four delivery methods tested
fewer confidently wrong answers - 14.0% of runs against 24.0% and 26.3%
more correct answers on messy CRM data - better on 8 questions, worse on none
Same model, prompts and rows in every condition; 759 graded runs, ground truth frozen before any run. On Google Search Console alone, documented CSV files did as well as Data to AI™ - the advantage comes from messier data such as HubSpot deals.
Read the benchmark →Add a HubSpot source in Dataddo with Deals, Deal Pipelines and Owners, then attach it to a Google Gemini destination or AI Model. In ~/.gemini/settings.json, add a dataddo server with httpUrl https://headless.dataddo.com/mcp-data, restart Gemini CLI and run /mcp auth dataddo to sign in. Gemini Code Assist in VS Code reads the same file.
With the url key, Gemini CLI falls back to SSE, which Dataddo does not support. Use httpUrl and leave out headers and any oauth block; the sign-in runs through /mcp auth dataddo.
Partly. Gemini Code Assist in VS Code reads the same settings.json as Gemini CLI. In IntelliJ, Gemini Code Assist uses its own mcp.json, so add the Dataddo server there as well.
No. Once signed in, Gemini reaches every flow or Context attached to an AI destination or AI Model in that Dataddo account, not only the Google Gemini ones. Keep unrelated flows off AI destinations if that matters.
As fresh as the last run. Deals and Line Items pull every current record on each run, while Properties History reads a relative window. HubSpot history cannot be backfilled, so append runs with the extraction timestamp if you need trends.
No. Gemini asks the semantic layer for entities and fields by name. Dataddo never accepts SQL from the model, and nothing is written back to HubSpot.
Run /mcp auth dataddo again and sign in with your Dataddo email and password, with MFA if prompted. Gemini CLI opens the sign-in page in a browser on the same machine.
