Find the useful work
We identify repeated tasks, decision bottlenecks and information gaps worth exploring.
We help you put Breeze to work on specific jobs, with a clear purpose, controlled usage and a way to judge the result.
Quick answer
Breeze AI implementation starts with one defined job, suitable data and a person accountable for the result. Project36 checks feasibility, prepares knowledge sources, configures the use case and tests responses or actions with your team. Human review, permissions and running costs remain part of the design.
Key takeaway: Use AI when it improves a real task. A simpler workflow or existing feature may be the better answer.
| Your situation | A useful next step | What to keep in mind |
|---|---|---|
| The task needs help interpreting or drafting information | Assess a focused Breeze use case | Outputs need testing against real examples and a clear escalation path; don't assume every answer is correct. |
| The task follows consistent rules and needs a predictable outcome | Review the existing process and automation | A workflow may be more predictable than AI, but still needs clean data, ownership and testing. |
| You can't explain the current credit consumption or expected cost | Review usage before expanding AI | Availability and consumption depend on your HubSpot subscription and the actions used. Check your current account and quote. |
The useful question isn’t how many AI features you can switch on. It’s where assistance, automation or better access to information can improve a real task without weakening trust or control.
We prepare data and knowledge sources, configure agents, test responses and actions, set human handovers and train your team. Then we keep improving how they perform.
GuideWhy marketing AI keeps producing generic work, and how to give it your business context.
Read morePrioritised Breeze use cases
Less time spent on repetitive work
Clear human review points
Better prepared data and knowledge
Visible usage and running costs
We identify repeated tasks, decision bottlenecks and information gaps worth exploring.
We examine data, permissions, platform capability, risk and expected benefit.
We configure a focused use case and test it with the people responsible for the outcome.
We document boundaries, review points and the practical way to use it.
For Artisan, behaviour-driven education helped reduce three-month break-clause activations by 65% and increased feature adoption 2.3 times.
Read the client storyAI and AI search
Your team spends more time briefing and correcting AI than it saves. The problem isn't the AI. It's context.
ReadAI and AI search
Tempted to use AI to build custom sales tools? It's a trap. Creating data silos is the opposite of what your growing business needs from HubSpot.
ReadAI and AI search
Xero's new AI integrations are a sign of things to come. For founders, it's a prompt to ask who owns your data when you connect HubSpot to third-party tools.
ReadUseful preparation
No. A well-defined task can be a sensible place to start, provided the data, risk and expected benefit are understood.
Yes. We can handle use-case design, data preparation, configuration, testing, training and ongoing improvement.
Yes. If a simpler process or existing feature is the better answer, we’ll say so.
We find the friction, underused capability and avoidable admin holding your team back, then turn the best opportunities into practical improvements.
Find out moreWe connect HubSpot to the tools your business depends on and build custom integrations when a suitable connection doesn’t already exist.
Find out moreWe train people using your setup, your processes and the work they actually do, so learning turns into better everyday habits.
Find out moreTell us about the time-consuming task, disconnected system or idea you haven’t quite worked out yet.
AI needs a defined job, appropriate data access and someone accountable for the result. Explore use cases alongside governance and consumption, not in isolation.