Rethinking escalation management for the AI era

Most organizations only have two speeds when it comes to escalations — "all hands on deck” or “please hold.” This has been the status quo for most teams because there simply weren’t enough hours in the day for support and engineering teams to attend to every case and issue in the backlog. In fact, developers cite lack of context, time for deep work, clear direction, understaffing and context switching as the key barriers to productive work (Atlassian State of DX Report 2024). 

Time spent maintaining software and time spent participating in the software maintenance process robs teams of time and focus that could be spent developing the next breakthrough feature.

We believe there is a better way - where support and engineering teams outsource the drudgery of researching, comparing, authoring and reporting to an AI Customer Support Engineer so that they can focus on more impactful work.

That’s why we raised $3.1M led by 42CAP and Dreamcraft Ventures, to resolve escalations before they become emergencies. That’s been the case for Kpler, one of our early customers, who cut escalation resolution time in half without growing their Support team, while the company grew from 150 to 700 employees and added 2 new product lines.. 

Early adopters of Waypoint AI like ClickHouse understand that large language model (LLM)-based agents offer a lever to reimagine escalation management to slash resolution time, increase CSAT and free up capacity in support and engineering.

No one is happy with escalations today.

When bugs spiral into emergencies the cost of moving slowly is churn, so organizations over-allocate engineering resources. Afterwards and rarely, engineers and support teams spend even more time discussing and filling in postmortem templates. By the time teams complete the paperwork, we’ve moved on to another fire drill and our good intentions to address the root cause fall by the wayside.

Meanwhile, Revenue teams are still watching escalations from the sidelines even though 73% of users say they switch to a competitor after multiple bad escalation experiences (Zendesk).

Escalations spiral because of knowledge asymmetry.

Customers know when the product isn’t working as they expected but they don’t know if it’s a bug, a feature, or user error. 

Support can try to verify if it’s a bug, but they often rely on predefined runbooks for the most common recurring issues or rely on their past experience. 

Engineers assigned to unfamiliar bugs spend hours searching for clues buried in code commits, chat threads, ticket comments or old postmortems.. 

Leaders struggle to establish or measure metrics for measuring escalation efficiency

Waypoint AI brings information to the game: it tells you the very next move - ask, route, investigate, update the customer -  no senior ‘bug whisperer’ required. No need to pull in the veteran engineer to get up to speed.

How Waypoint AI works

Waypoint AI embeds in the tools that support and engineering teams already use to proactively accelerate escalations from intake to resolution:

• Customer Support: Salesforce, Jira Service Desk, Zendesk
• Engineering: Jira, GitHub, GitLab
• Enterprise Knowledge & Communications: Slack, Teams, Confluence, Knowledge Bases

As soon as an issue lands, Waypoint crafts a Runbook for support — step-by-step, source-linked, auto-updating — so the first responder always knows the next move. These Runbooks can cite standard operations procedures, but they also reflect best practices and learnings from previous escalations, Slack threads, documentation and other sources.

The engineering team recently changed their tribe scopes again. Before I could even tell our agents about these new scopes, Waypoint AI was already recommending routing to new tribes.

— Fraser Fairbairn, Head of Support at Kpler

Waypoint also automatically extracts and tags key events on our Timeline. We track time spent per function, impacted components, touch points with the customer and flag stalled escalations.

When escalations resolve, Waypoint AI converts that Timeline into an editable Postmortem, tags Root Causes and generates recommendations to prevent similars escalations.

Our Commander View gives you a overview of escalations across the organization enabling quick action on stalled escalations and synthesized learnings from recent postmortems.

Where we are going

Waypoint AI can operate both as a copilot and an autonomous agent. With early customers reporting that over 95% of escalations are routed to the right team the first time, we are ready to autonomously triage and route escalations. If you’re looking for ways to unlock 10-15% of engineering and support productivity this year, deploy Waypoint AI today and see gains within a sprint. Book a quick demo today.

If you’re not ready, sign up to get our forthcoming series Supporting Engineering delivered straight to your inbox, where you’ll hear over a dozen experts share how they manage escalations in world-class organizations.