How Alien Road Uses HubSpot Breeze in Client Campaigns

Breeze is HubSpot’s name for its AI features across the Marketing, Sales, Service, Content and Operations hubs. HubSpot introduced it in 2024 and has expanded it since. It has three main parts: Breeze Copilot, an assistant available across HubSpot; Breeze agents, which perform specific tasks such as content creation, prospecting, customer service and social posting; and Breeze Intelligence, which provides data enrichment and buyer intent signals.

Which features you get depends on your HubSpot subscription and the credits included, and HubSpot changes packaging regularly. Check your account’s settings and HubSpot’s current documentation at the time of writing.

Context. Breeze works on the data already in HubSpot: contact properties, deal history, emails, tickets and website activity. A separate chatbot would need that data copied in. Keeping the work inside the CRM also keeps a record of what was done, by whom, against which contact or deal.

That said, Breeze is not always the best tool for every job. For long-form strategy documents or complex analysis we may use other models. The rule is simple: use Breeze when the task depends on CRM data or ends in a CRM action.

Copilot appears in the HubSpot interface and in the mobile app. Our account managers and the client’s team use it for small, frequent tasks.

• Summarising a contact or company record before a call, including recent emails and deals
• Drafting follow-up emails based on the last meeting notes and deal stage
• Writing first drafts of marketing emails and landing page sections in HubSpot’s editors
• Building lists and reports by describing them in plain language
• Explaining unfamiliar properties or workflows in a client’s portal

Copilot output is always read before use. For emails sent to prospects or customers, the sender edits and approves the text personally.

Agents take on larger, defined jobs. We use them where the job is repeated and the output can be checked.

Content agent

The content agent drafts blog posts, landing pages, case study outlines and podcast or video scripts from a brief. Our content team uses it for first drafts on topics where the client’s brand voice is already well documented in HubSpot. Every draft goes through an editor who checks facts, adds expertise the agent cannot know, and fixes tone.

Prospecting agent

For sales teams, the prospecting agent researches target accounts and drafts personalised outreach. We configure it with the client’s ideal customer profile and messaging, start with a small set of accounts, and have a sales rep review each email before sending. Volume only increases after the reps agree the drafts are good.

Customer agent

For service teams, the customer agent answers questions from the client’s knowledge base on chat and other channels. We prepare the knowledge base first, set handoff rules to human agents, and review conversations weekly in the early weeks.

Poor data makes every AI feature worse. Breeze Intelligence can enrich contact and company records with firmographic data such as industry, size and location, and can show buying intent based on website visits by companies. We use enrichment to fill gaps in segmentation fields that the client’s workflows depend on.

• Audit which properties drive lists, workflows and lead scoring.
• Measure how many records have those properties empty.
• Enrich priority segments first, since enrichment uses credits.
• Check a sample of enriched records for accuracy before trusting them in automation.
• Document which fields are enriched and which are entered by people.

Clear rules keep AI use safe in a system that holds customer data. We agree these with the client before switching on new features.

• Only users who need a Breeze feature get access to it, managed through HubSpot permissions.
• No AI-drafted message is sent to an external contact without human review.
• Enriched data is labelled, and personal data handling follows GDPR, KVKK or other applicable law.
• Agent configurations, prompts and knowledge sources are documented with an owner.
• Credit usage is monitored monthly so costs do not surprise the client.

We measure at task level and at outcome level. At task level, we log time spent on recurring jobs before and after, and the number of edits drafts needed. At outcome level, we look at HubSpot reports the client already uses: email reply rates for prospecting, content publishing frequency and organic traffic, ticket resolution time and customer satisfaction for service, and pipeline created from enriched segments.

We present those side by side with credit spend. If a feature uses credits without improving a metric, we switch it off or change how it is used.