Microsoft calls this the “Transformation Paradox”: employees feel pressure to adopt AI quickly, “but the systems around them … continue to reinforce the old way of working” (same source). That sentence is worth reading twice in any leadership meeting.
Why does the gap open?
Personal AI use is cheap to start. One person, one tool, one task. Organisational AI use is not. It needs shared data, changed processes, clear permissions and someone accountable for results. In our work across the region, four blockers come up again and again.
- The work is personal, the systems are shared. A rep can draft a better email in seconds, but the CRM still holds duplicate accounts and missing contacts, so the email goes to the wrong person.
- Incentives reward the old process. Targets count activities the AI now does, such as emails sent, rather than outcomes.
- Nobody owns the change. AI sits with IT, innovation or a task force, not with the leaders who own revenue and service.
- Risk says no by default. Without clear rules, the safest answer is to block tools, so people use them anyway, out of sight.
Why is the gap a cost, not just a curiosity?
Because the value stays personal. One person saves an hour; the business sees nothing, because the hour disappears into the next task. Worse, work done in personal tools leaves no record. Customer conversations, research and drafts sit outside the CRM, where they cannot be reused, measured or protected. Under Malaysia's amended data protection law, that is also a compliance question, which we cover in PDPA amendments for CRM teams.
How do leaders close the gap?
Not with more training alone. Training helps people who are already ahead go further. Closing the gap means moving the organisation. Five moves work.
- Align the leadership team on three use cases. Pick the processes where AI changes a revenue or service number, and say so publicly. Our method for ranking AI use cases by value fits on one page.
- Put the work where the data lives. Move AI from personal tools into the shared platform, so drafts, research and conversations land on the customer record. In HubSpot, that means agents and the Breeze Assistant working on the CRM, not beside it.
- Change what you reward. Replace activity targets with outcome measures: meetings held, cases resolved, revenue retained. Recognise people who redesign a process, not only those who use a tool.
- Give every agent and workflow an owner. Make a business leader accountable for each AI use case, with written limits and a review rhythm. We set this out in who owns the agent.
- Write the rules once. A short policy on data, tools and approvals lets risk say yes with conditions instead of no by default.
What does the gap look like inside a revenue team?
Take a typical regional sales team. Reps use a personal AI assistant to research accounts and draft emails, and the drafts are good. But the research lives in a chat window, the email is sent from a personal inbox, and the CRM records nothing beyond a manually logged activity, if that. The manager's pipeline review still runs on a spreadsheet exported on Friday. Leadership sees no change in win rate and concludes that AI is overhyped.
Nothing about the people failed. The organisation never connected their work to the shared system. The same research done by an agent on the CRM would sit on the account record, inform the next rep who picks up the account, and show up in the numbers leadership already watches.
Where should you start?
Start with your frontier people. Every organisation already has some. Find the people who are furthest ahead, ask what they do and where the systems stop them, and make their way of working the standard for one team. Then fix the data and process that blocked them, and scale it to the next team, and the one after that.
Expect the first team to take longest. Each later team inherits the data, rules and habits already in place.
To find the team to start with, our AI readiness assessment measures both sides of the gap.
