Understand HubSpot with AI.
Connect HubSpot to Lytically and investigate contacts, companies, deals, and pipeline through the AI tools you already use.
CRM & Sales- CRM overview
- Open pipeline
- Deal outcomes
- Contacts, companies, and deals
- Funnel activity
What you can investigate
Questions this connection can actually support.
CRM overview
See the period's headlines next to where open deals sit today.
Open pipeline
See where open deals sit by stage. This is the pipeline as it stands, not a history of how long each deal has been in a stage.
Deal outcomes
See whether closed value, win rate, or time to close changed, and whether that change sits with one owner or one pipeline.
Contacts, companies, and deals
Look up the contacts, companies, and deals in the period when a total needs the records behind it.
Funnel activity
See how many contacts entered each lifecycle stage in the period, next to deals that closed. This is activity in that window, not the same people moving from one stage to the next.
Questions you can ask
Asked from the AI tools you already use.
Illustrative answer
Why did won deal value fall while the pipeline still looks full?
01 Evidence
Deals won fell from 14 to 9. Won value fell from $184k to $112k. Open pipeline by stage was broadly unchanged. New contacts were steady.
02 Context
The company counts reported revenue when the invoice is issued. HubSpot marks a deal won when it is closed in the CRM. Those are different events. This result does not include invoice dates.
03 Reading
HubSpot shows fewer deals closed, and less closed value. A stable open pipeline does not replace those wins. An invoicing delay could change reported revenue. It does not explain why HubSpot recorded fewer won deals. That would need a note this connection does not have, such as which deals were postponed.
Sample figures, not a customer result.
Where are open deals concentrated by stage?
Did win rate or days to close change last month?
Did created deals and won value move together for a given owner?
Did new contacts rise while deals won stayed flat?
What does the open pipeline look like right now, by stage?
How the integration works
Connect the account, then ask from a compatible client.
01
Connect HubSpot in Lytically. HubSpot’s consent screen chooses the portal and asks for read access to contacts, companies, and deals.
02
That portal is stored as the selected resource. Lytically does not ask you to pick contacts or deals again. Those objects are read from the portal when a question needs them.
03
If the grant is missing read access to contacts, companies, or deals, the connection is saved in an error state instead of looking ready.
04
The HubSpot tools then appear on the MCP connection. Period reports use the portal timezone, defaulting to UTC, and end yesterday.
The same tools are available to a Routine, which repeats an investigation on a schedule you set.
Available analytical capabilities
What the tools cover, and what they leave out.
Portal
One HubSpot portal per grant, chosen on HubSpot’s consent screen. A second portal is another connection. If more than one portal is selected on the account, name the portal in the question.
Operations
CRM overview, open deal pipeline, deal performance, contact list, company list, deal list, and conversion-funnel activity.
What the funnel counts
Lifecycle metrics count contacts that entered each contact lifecycle stage in the period. Deals won and won revenue use the close date. Rates describe activity in the window, not a cohort.
Limits
Lists default to 25 rows and cap at 50. Contacts and companies can be filtered by their current lifecycle stage. Totals that have to page through search results stop after a limited number of pages, so a very large portal can be incomplete.
Freshness
Period reports end yesterday in the portal timezone. The open pipeline tool is a current snapshot and is not given a comparison window.
Access
HubSpot OAuth with read scopes for contacts, companies, and deals. Those scopes do not write records. There are no write tools, webhooks, or a local copy of the CRM. Credentials are encrypted and are not returned to the AI client. Tools run for the Lytically account on the MCP connection, and only for the portal on that connection.
Better together
More useful next to another connected source.
HubSpot can say what the CRM recorded. It cannot say which query or campaign produced a person. That evidence comes from another system, called on its own.
Google Analytics
Key events fell and new contacts did not. Analytics shows on-site events. HubSpot shows contacts created in the period. Matching them requires a definition of what an enquiry is.
Google Ads
Ads conversions rose and deals won did not. Spend and conversion actions stay in Google Ads. Created, won, and lost deals stay in HubSpot.
Google Search Console
Organic clicks increased and enquiries, counted as new contacts, did not. Search Console contributes the queries. HubSpot contributes the contact activity. The link between them is an inference, not a shared id.
Questions
Before you connect HubSpot.
Can I use this integration with ChatGPT or Claude?
Yes, once the portal is connected and the read scopes are present. The CRM tools then show up on Lytically’s MCP connection. A missing scope leaves the connection in error, so the tools are not offered against a partial grant.
Does Lytically store my HubSpot data?
Lytically stores the encrypted grant and the portal it belongs to. Contacts, companies, and deals are read from HubSpot when a tool runs. They are not mirrored into Lytically. A Routine can keep the written report from a run, and that report may include figures from the investigation. That report is not a copy of the CRM.
What permissions does Lytically require?
Read access to contacts, companies, and deals, plus the OAuth scope HubSpot uses for the grant. There is no write scope. If one of the read scopes is missing, the connection is marked as an error.
Can I connect multiple portals?
One portal is chosen per connection, on HubSpot’s consent screen. Connect again to add another portal. Lytically does not offer a second picker for objects inside the portal.
How current is the data?
Period reports end yesterday in the portal timezone. The open pipeline is the current open pipeline, not a reconstruction of last month’s stages. Deal comparisons use the previous period of the same length.
Can I combine this data with other integrations?
The assistant can call Analytics, Search Console, or Google Ads in the same investigation. HubSpot will not attribute a contact to a campaign or a query. Any link between those results is the assistant’s reading, and it should say so.
Put your HubSpot data to work.
Use a Blueprint Call to see how pipeline, deals, and contact activity would be read with the definitions your team already uses.