To drive AI adoption, senior leaders must explicitly give their teams permission to experiment and push boundaries. A key leadership function is to absorb risk by saying, "Blame me if it all goes wrong," unblocking hesitant engineers.
To get an engineering team "AI pilled," a powerful strategy is to give them a month to fix everything they dislike about the codebase using AI tools. This provides a tangible, motivating win and demonstrates the power of AI on familiar problems.
Intercom noticed AI-generated pull request descriptions were poor. Instead of a wiki, they built a mandatory "Create PR" skill that enforces high-quality, intent-focused descriptions, turning a cultural standard into an automated process.
The transition from basic AI code completion to advanced models means the tool is no longer the limiting factor. The real challenge for engineers is now expanding their imagination to conceive of what's possible, rather than massaging the tool to get a result.
Intercom monitors its internal AI skills with Honeycomb for usage and analyzes session transcripts stored in S3. This product-centric approach provides insights to improve tools, identify user struggles, and offer personalized feedback to engineers.
Facing an AI bill that looks like their velocity chart, Intercom deliberately absorbs the cost. They encourage universal use of the most powerful models, viewing the immediate gains in speed and innovation as an investment that outweighs near-term cost concerns.
AI makes achieving a zero-item backlog a feasible reality. The ability to quickly resolve tech debt, perform migrations, and tackle long-standing "wish list" items means teams no longer have to choose between maintenance and new features.
Contrary to fears that AI creates low-quality "slop," Intercom found their code quality improved. AI compresses the cost of fixing tech debt, flaky tests, and other internal projects, making it easier for the business to invest in them.
An Intercom AI skill for fixing flaky tests goes beyond a simple script. It updates its own internal checklist when it encounters a new type of fix and then proactively searches the codebase for similar problems, creating a 100x impact.
Intercom's CTO set a goal to 2x R&D throughput, using pull requests as a simple, albeit crude, metric. In a high-trust environment, this focused the team on adopting AI tools to increase output, leading to measurable success.
AI agents often default to "build it yourself" because SaaS products aren't designed for them. To stay relevant, SaaS companies must create agent-friendly CLIs, APIs, and even add hints in help text to guide agents through complex workflows.