The Problem: Complex segmentation required engineering support

Mercury's lifecycle team runs campaigns through Customer.io. However, Customer.io never had the full picture. A lot of the data that mattered sat in Snowflake, and marketing couldn't access it without going through the data team, whose bandwidth was heavily contested.

In practice, that meant every deep audience question — "which orgs show readiness signals for team spend?" became a data ticket. 

Then came the launch of Mercury Personal: a cross-sell opportunity to existing business customers, exactly the kind of multi-criteria segmentation problem the old workflow would take a lot of time and effort to execute.

The Solution: Connecting Sortment to their stack

Sortment sits between Mercury's Snowflake warehouse and Customer.io. Marketing gets direct access to explore data, build audiences, and sync them into campaigns without engineering dependency by using Sortment’s agents.

"Before Sortment, every deep audience question was a data request. Now my team gets from question to synced audience in the same afternoon."

- Ellen Rockdale, Head of Lifecycle, Mercury

Use cases

1. AI segmentation for complex, multi-criteria audiences

For the Mercury Personal cross-sell use case, the team needed to identify the right people to target based on signals available in Snowflake. Sortment became a sounding board to solve this, running analyses to find leading indicators of interest in the personal banking1 product based on past data, and converting those signals to segmentation criteria. All while the lifecycle team talked to Sortment in plain natural language.

Sortment understands the underlying data schema in Snowflake. Over time, it has learned the nuances of the business, internal terminology, and team workflows. 

2. Quests: AI that works towards a goal over weeks and months

Mercury handed Sortment a goal: “grow team spend adoption.” Over a 13-day period, the Quest came back with 15 operationalizable signals: tier-2 expansion, treasury2 activation, reimbursements, and more. These signals are a set of hypotheses for the team to test through their campaigns and journeys.

3. AI analytics without the analyst queue

Campaign impact analysis that goes beyond opens and clicks and breaks it down by different audiences previously required heavy data science involvement, but now happens inside Sortment with just prompts. The team can also have AI automatically analyse campaigns, journeys, and experiments and flag issues or identify insights on a schedule, making the team more proactive than what was possible before.

4. Slack-native workflow

Tagging “@Sortment” in Slack makes it easier to access it directly from where the team collaborates. This makes talking to Sortment feel exactly like talking to a colleague.

“Sortment feels less like a tool and more like a colleague on the team. I can hand it goals, analyses, and tasks and see it contribute.”

Ellen Rockdale, Head of Lifecycle, Mercury

Explore

Strategy AI

Quests

Audiences

“Sortment feels less like a tool and more like a colleague on the team. I can hand it goals, analyses, and tasks and see it contribute.”

Ellen Rockdale, Head of Lifecycle, Mercury

Explore

Strategy AI

Quests

Audiences

“Sortment feels less like a tool and more like a colleague on the team. I can hand it goals, analyses, and tasks and see it contribute.”

Ellen Rockdale, Head of Lifecycle, Mercury

Explore

Strategy AI

Quests

Audiences