We recently joined HubSpot’s sales team in Sydney to speak on their Partner Perspectives AI panel. The audience was a room full of HubSpot salespeople looking to understand how real businesses are navigating AI adoption.
As an Elite partner working across mid-market and enterprise clients, we see the patterns up close—what’s working, what’s not, and what’s slowing teams down.
We wanted to share some of the key trends and advice we shared at the event, starting at the day-to-day level:
- Leaders are excited and want to move fast.
- Teams responsible for delivery often feel overwhelmed.
- There’s a growing gap between ambition and execution.
- Budget, resources, and readiness often don’t match the vision.
This tension isn’t a technical problem—it’s an organisational one. And it’s where most AI initiatives either stall or fail to move beyond experimentation.
The two biggest blockers we see
On stage, we broke down the two issues that consistently get in the way of progress:
- Uncertainty and discomfort: Teams aren’t sure what AI should be used for, how far to trust outputs, or where it actually fits in their workflow. There is often uncertainty on a company level - are they allowed to use it? Will they be perceive as lazy for using it?
- AI washing: Organisations pitch big, flashy ideas without the underlying process, data structure, or operational maturity to support them.
Both create unrealistic expectations and lead to stalled projects.
Why process matters more than features
A recurring question from HubSpot’s sales team was how we help clients decide where to start. We explained that our approach is never “tool first”. It’s process-first, because the value of AI is unlocked in the workflow—not in the interface.
When we step into an organisation, we document five core areas:
- Existing workflow friction
Understanding where work slows down, bottlenecks form, or teams rely on manual patch-jobs shows us exactly where AI can create immediate relief. Without mapping this friction, businesses risk applying AI to the wrong places and reinforcing bad processes instead of improving them. - Manual tasks that can be safely automated
Not everything should be automated. Some tasks need human oversight, contextual judgment, or emotional intelligence. Others—like repetitive triage, formatting, routing, or transcription—are perfect candidates for automation. By documenting which activities fall into which category, we help teams adopt AI without creating operational or quality risk. - Data availability and quality
AI is only as good as the data it’s fed. If key systems aren’t connected, if content is outdated, or if CRM data is inconsistent, the output will always fall short. Identifying gaps early prevents rework later and gives us a clear roadmap for what needs to be cleaned, centralised, or restructured before automation can scale. - Change readiness across teams
Even the best AI use case will fail if the people responsible for it aren’t prepared, trained, or confident. Documenting where each team sits—excited, skeptical, overwhelmed, or confused—helps us tailor the rollout. Change management is an operational requirement, not an afterthought. - Expected value vs. effort
We map every opportunity against the return it can deliver relative to the effort required to implement it. This prevents organisations from chasing high-effort, low-impact ideas. It also highlights the fast wins that build momentum and trust.
The principle we emphasised was simple: start with “what hurts today?”, not with “what can this tool do?”.
That’s where the highest value lives and where confidence in AI adoption grows the fastest.
Data quality still makes or breaks every outcome
We emphasised that AI only works as well as the data it has access to. Across our projects, this shows up in:
- Outdated knowledge bases
- Disconnected systems
- Messy CRM architecture
- Legacy content the business didn’t realise was still live
One example provided was a client whose AI agent surfaced old web pages that hadn’t been updated in years—something the business wasn’t even aware existed. It’s a perfect illustration of why data readiness must come before AI scale.
The use cases we shared on the panel weren’t the big, shiny AI projects people expect. They were the practical wins that deliver real value quickly. For example, customer service agents taking on low-value enquiries so teams can focus on the work that actually moves the needle. An example of this is a custom assistant built for executives, pulling investor history and key notes instantly—simple, high-impact time savers.
AI creates value when it replaces inefficient processes, not people.
We also highlighted how confidence grows once a team experiences its first win. After that moment, ideas surface more naturally, teams begin identifying opportunities themselves, resistance drops, and alignment across departments increases. The challenge isn’t convincing organisations to use AI—it’s helping them introduce it in a way that builds momentum rather than overwhelm.
AI isn’t a silver bullet, and it’s not something you switch on and walk away from. It’s iterative, it requires solid foundations, and it rewards businesses that take a structured approach. When that happens, the outcomes become very real: time saved, better customer experiences, and teams who feel empowered rather than displaced.
If you’d like support evaluating use cases or understanding how AI can support your internal operations, get in touch through the form below.