How a Compliance SaaS Provider Automated 41% of Support Tickets Without Extra Headcount
The Challenge
Sticos, a Norwegian compliance software provider, faced a scaling crisis in customer support. Between 30–40% of tickets arrived via email — each requiring manual customer identification before an agent could even begin. Around 20% of all tickets were simple, repetitive questions: password resets, login issues, basic how-to guidance. When the business absorbed a new product line bringing 25 additional staff and 15,000 extra annual support tickets, the team had no path to absorb that volume without significant hiring. The problem was structural: support demand was growing faster than headcount could follow.
The Solution
Deployed HubSpot Service Hub with the Customer Agent AI feature. Built a structured knowledge base spanning all product lines and embedded the agent across product chat, knowledge base, and email channels. Created custom workflow triggers to route specific email phrases through automation before human review. Treated the AI agent as a coached team member with detailed behavioural guidelines. Added multi-language support across Norwegian, English, Danish, and Swedish. Introduced a custom "AI Potential" ticket property to systematically identify further deflection opportunities over time.
The Outcome
The Customer Agent handled 41% of all incoming support volume, exceeding the team's 40% target. Chat deflection reached 91%, with 75% of those chats fully resolved without human intervention. The team absorbed the full 15,000+ ticket increase from the product acquisition without proportional hiring. Support staff shifted from reactive ticket-clearing to proactive customer success work.
Strategic Takeaway
Sticos illustrates the inflection point many growing SaaS businesses hit: support volume compounds faster than headcount can scale, and hiring one-to-one becomes economically untenable. The resolution is not more agents — it is building a knowledge infrastructure that makes repetitive queries self-serve. When AI deflection is trained on structured, product-specific content rather than generic responses, resolution rates can reach 75%+ in chat. The key design decision was treating the AI agent as a coachable team member with explicit guidelines, not a bolt-on chatbot.
- Support scaling crises are knowledge problems disguised as headcount problems — the fix is structure first, automation second.
- A 91% deflection rate requires investment in knowledge base depth, not just AI tooling — the content is the product.
- Tracking 'AI Potential' as a ticket property turns deflection from a one-time project into a continuous improvement system.
See how HubSpot can help your business address similar challenges.
Affiliate link — we may earn a commission at no cost to you.