The challenge
Fifteen teams had to be onboarded at the same time. Email comes in from three different kinds of B2B stakeholders, each with their own questions and their own place in the process. The busiest teams receive 100 to 200 emails a day.
And as with every growing organisation, the same question could get a different answer depending on which team, or which colleague, picked it up.
Starting with a shared vision
Before building anything, operations and management set the bar together. How fast do we want to be vs. how complete in the first answer? Do we lean on self-service, or do we take work off the client’s hands? Where does our service really stand out? The outcome was a vision document, signed by both: the North Star every agent is built against.
What we built together
Routing by stakeholder
Every incoming email is read and sorted: which kind of stakeholder, which subject, which team. It lands with the right team, at the right step of the process, before anyone has to look at it.
Replies from each team’s own knowledge
The AI drafts replies based on the knowledge base of that specific team, so every answer reflects how that team actually works.
How it works
-
Trigger
An email from one of three kinds of B2B stakeholders
-
Assess
Which stakeholder, which subject, which team
-
Take action
Routed to the right team and process step
-
Communicate
A reply based on that team’s knowledge base
AI forces a choice, and that’s the point
“I’ve lost my fuel card, what now?” It sounds simple, but different teams answered it slightly differently. An AI is efficient when trained for one correct answer, so it made those differences visible and asked the organisation to choose. Each of those moments became a chance to align.
What changed
Response times dropped by 55% on average, with the same team. The time saved now goes to tasks that used to be left for less busy days.