Agent seen as a gimmick
Teams click Copilot once, get a mediocre answer and switch it off. No second attempt, no context investigation. The agent disappears from the routine.
TRAINING AND ADOPTION
Working with agents in HubSpot: Breeze, Copilot and the Customer Agent.
HubSpot has been an agentic customer platform since 2026. Breeze, Copilot, Customer Agent and Sales Agent run alongside your team. Some teams are already using them, others are not yet sure how or why. We train sales, marketing and service on an agentic mindset, prompt design, governance and the AI Act context. On your own environment, with your own data, in your own workflows.
What do you deliver?
-
01
Working deployment of Breeze, Copilot and Customer Agent On your own HubSpot tenant with real cases
-
02
Prompt design suited to commercial cases A team library you maintain yourself
-
03
Governance agreements plus audit trail approach Which data, which actions, which escalations
The observation
HubSpot adds agents, commercial teams fall behind.
HubSpot announced Breeze in 2025 and has been an agentic customer platform since 2026. At the same time we see that sales, marketing and service teams barely know how to work with it. An agent is not a button, it is a new way of working that requires a well-considered CRM architecture. These are the four patterns we encounter.
No prompt discipline
Everyone types something different for the same question, with inconsistent output. No team library of good prompts, so no learning effect.
Governance undiscussed
Which data may the Customer Agent show a client, which actions may the Sales Agent take on its own. No one has discussed it, so everything is either on or off.
AI Act stuck with legal
Sales and marketing hear "AI Act" and push it to legal or compliance. Yet the practical impact on transparency and data use is commercially relevant.
Approach
From button to colleague.
Four sessions on-site or online, with homework on your own environment. We do this differently from a generic AI course: your own data, your own Breeze, your own workflows. No abstract cases from a textbook, but a real place within our training offer.
01
Week 1
Agentic mindset
How an agent thinks, what it can and cannot do, and why a sales rep needs to adapt their way of working. With live demos on Breeze in your own HubSpot tenant.
Wat je krijgt
mindset document plus a first set of working prompts for your role
02
Week 2
Prompt design and cases
Prompt discipline for commercial cases: lead research, deal update, ticket categorisation, customer summary. With a team library you maintain yourself, also after the CRM training.
Wat je krijgt
prompt library for your team in a shared channel
03
Week 3
Governance and audit
Which data, which actions, which escalations. Plus how to set up audit trails so management can see what an agent does for your team. Not all agents on, not everything off.
Wat je krijgt
governance document plus audit trail setup in HubSpot
04
Week 4
AI Act for commercial teams
Practical impact of the AI Act on transparency, client communication and data use. Not a legal lecture, but concrete guidance for sales, marketing and service routines.
Wat je krijgt
commercial AI Act checklist plus communication template for clients
Which agents do we cover
The agents commercial teams work with.
Not all Breeze agents are relevant for every team. Based on the intake we choose which agents to cover in depth, and which to address only briefly. Below are the six we encounter most often.
Breeze Copilot
The assistant in every HubSpot view: contact context, deal summary, content suggestions. The entry-level agent for almost every team.
Customer Agent
Automating client conversations in live chat and email, with escalation paths for service teams. Including tone of voice and boundaries.
Sales Agent
Lead research, initial outreach, deal update. For sales teams that want to lighten their BDR workload without losing quality.
Content Agent
Blog elaboration, email variations, landing page copy. For marketing teams producing at volume without it becoming AI soup.
Prospecting Agent
Account research and first-touch sequences. For SDR teams and account managers who want to accelerate their pipeline coverage.
Knowledge Agent
Access to your internal knowledge base via natural language. For service teams that want to quickly build client context before a conversation.
What your team can do afterwards
Agents as colleague, not as gimmick.
Concrete actions for your commercial team, not just an understanding of the concept. A sales rep uses the Sales Agent in their first-touch routine, a service agent lets the Customer Agent co-write a complex reply.
What your team can do afterwards
- Build a working prompt for your own case
- Judge when an agent does and does not help
- Assess Copilot suggestions on quality
- Take over a Customer Agent conversation at the right moment
- Recognise and respect governance boundaries
- Explain AI Act context to a client
Who delivers the training
Yoni designs, Kim delivers on the agents.
For AI training you work with Yoni and Kim. Yoni translates the agentic mindset to your operating model, Kim runs the hands-on sessions on Breeze agents in your own HubSpot environment. For the broader adoption approach, see also our adoption programme.
Case
·
B2B mid-market
Read more cases
→
Agents as colleague, not as gimmick
Concrete actions for your commercial team, not just an understanding of the concept. A sales rep uses the Sales Agent in their first-touch routine, a service agent lets the Customer Agent co-write a complex reply, all within your HubSpot environment.
After four sessions your team can build a working prompt for their own case, judge when an agent does and does not help, assess Copilot suggestions on quality, take over a Customer Agent conversation at the right moment, respect governance boundaries and explain AI Act context to a client.
B2B mid-market
Training Lead, ~80 employees
·
B2B client
4 wks
training block
6
Breeze agents in scope
2 wks
parallel support after sessions
Questions
What clients usually ask.
Honest answers to the questions we hear before every AI training. For the technical side we refer to our AI readiness scan.
Who is the AI training for? +
Sales, marketing and service staff who are going to work with HubSpot Breeze, Copilot, Customer Agent or Sales Agent. Plus team leads who want to understand governance and prompt design before releasing the agents. Not a developer training, but an agentic mindset for commercial teams. Not sure where you stand? Take our maturity scan first.
What is an agentic mindset? +
The mindset with which you deploy an agent as a colleague rather than a button. You give it context, an assignment and boundaries. You assess its output as you would with a new colleague, and you adapt your way of working to what the agent does and does not do well. For commercial teams, that is a significant shift.
Does this training also work if we do not use Breeze? +
The principles (agentic mindset, prompt design, governance) are vendor-neutral. We often train teams that use ChatGPT, Claude or Copilot for Microsoft 365 alongside Breeze. The main focus remains commercial context within your CRM, not a generic AI course.
How deep do you go on governance and the AI Act? +
Practical and commercial, not a legal lecture. We cover which data you do and do not give to an agent, how to set up audit trails for agent actions and what the AI Act concretely means for commercial teams. For a real legal review we refer to an AI Act expert.
How long does a programme take? +
A standard programme is 4 to 6 sessions spread over 4 weeks. Online sessions are 90 to 120 minutes, on-site sessions are a half-day. Then two weeks of parallel support while your team works with the agents and we monitor real cases alongside them. For ongoing guidance clients look at RevOps as a Service.
What does AI training cost? +
Our services are bespoke. An AI training for a commercial team is typically between 6,500 and 14,000 euros depending on team size and number of sessions. Schedule a call for a quote tailored to your situation.
Can this also be in English for international teams? +
Yes. We train in Dutch or English, and mix where that is useful. We deliver materials in the language your team requests.
What is the difference with the AI readiness scan? +
The AI readiness scan looks at your data model and architecture to check if it is ready for agentic use. The AI training looks at your team: are the people ready to work with agents. Many clients combine them: first the scan for the technical side, then the training for the team.
Ready to start?
Ready to get your commercial team working with agents?
Schedule a short intake. We listen, outline which agents are the priority for your team and give an honest price.