All case studies

Marketplace customer experience

Marketplace CX found $4.2M+ in annual savings in 17 days.

In 17 days, 8Flow found the automatable friction in high-volume contact workflows and put a dollar figure on fixing it.

Large-scale CX operation spanning BPO sites, admin tools, CRM, knowledge content, and AI assistant usage.

$4.2M+
annual savings identified
from roughly one minute of automatable friction per contact
23
opportunities found
across CRM and admin workflows in the first 17 days
50K+
manual transfers
observed in two weeks in the highest-volume copy/paste pattern
115 hrs
recoverable search time
top of the modeled range across about 9,400 knowledge-assisted cases in two weeks

The starting point

Before 8Flow

  • An earlier proof of concept showed contacts took too many clicks and systems, but not where to invest.

  • Users juggled CRM, admin tools, knowledge sources, and an AI assistant during live contacts.

  • Leaders had no ROI evidence to sequence automation, tooling, and content work.

The findings

What 8Flow surfaced.

Every finding below comes straight from measured workflow data.

Manual movement

The biggest pattern: copying data between systems mid-contact.

Users pulled data from admin tools and documents, then pasted it into call and case workflows, over and over.

  • 50,000+ manual transfers in two weeks.
  • 70+ hours of data movement in one call-handling flow alone.
  • The fix: auto-fill inside the live support surface.
50K+
manual transfers
over a two-week window

Knowledge base

Knowledge search added time to every contact.

A single lookup added 2 to 2.5x handle time across channels and contact types.

  • The average search took about 42 seconds per contact.
  • About 9,400 knowledge-assisted cases in two weeks.
  • Auto-surfacing articles could close 60 to 70% of the gap.
115 hrs
recoverable search time
top of the modeled range over two weeks

AI assistant

Users opened the AI assistant. Few really used it.

Most users opened the assistant, but typed usage stayed shallow and app switching went up during sessions.

  • 71% of users opened the assistant.
  • Only 6% of opens became typed queries.
  • App switching rose 34% during assistant sessions.