SELECTED WORK

Proof, in production.

Every engagement below runs in production today. The numbers are measured against baselines our clients' finance teams signed off on.

PLACEHOLDER CASES — SWAP WITH REAL DATA
Sepia data plate: a field of account dots with high-intent leads highlighted in red, routed through a hub into a rising qualified-pipeline bar chart.
FORTUNE 100 INSURER · GTM STRATEGY

Cutting cost per qualified lead 62% — without adding headcount.

Intent-scored routing replaced static territories across a 400-seat sales org. Pipeline lift was visible in the first quarter; the CFO signed off on the second.

Read the case →
Sepia data plate: agent-run lifecycle automation across six regional lanes, red task packets flowing, beside an audit log column flagging one escalated action.
GLOBAL SAAS SCALE-UP · MARKETING AUTOMATION

11,000 hours a year, returned to a six-region marketing org.

Agent-run lifecycle campaigns with audit trails compliance could love — campaign cycle time fell from fourteen days to thirty-six hours.

Read the case →
Sepia data plate: a grid of operator dots with red ripples spreading certification outward from trained operators, beside cohort progress bars.
NATIONAL RETAILER · CORPORATE AI TRAINING

340 certified operators in 90 days — an org that ships its own automations.

A practitioner academy turned the merchandising org into daily AI users — graduating by shipping real automations, not passing quizzes.

Read the case →
← ALL WORK
CASE STUDY — FORTUNE 100 INSURER

Cutting cost per qualified lead 62% — without adding headcount.

INDUSTRY
INSURANCE
PRACTICE
GTM STRATEGY
MODEL
SPRINT → PARTNERSHIP
TIMELINE
8 WKS + ONGOING
Sepia data plate: a field of account dots with high-intent leads highlighted in red, routed through a hub into a rising qualified-pipeline bar chart.
THE CHALLENGE

A 400-seat sales org was working static territories drawn years earlier, while marketing flooded the funnel with leads that looked alike on paper. Cost per qualified lead had climbed three years straight — and adding headcount had stopped moving the number.

THE APPROACH

Two weeks of revenue archaeology found the leak: high-intent accounts were waiting in the same queue as everyone else. We built intent scoring on eight quarters of closed business — back-tested before anyone bet a dollar on it — then redesigned routing and territories around the signal, with sellers in the working group from day one.

INTENT SCORING IN CRM DYNAMIC ROUTING QBR DASHBOARD
THE OUTCOME
−62%
cost per qualified lead
3.4×
qualified-pipeline lift in two quarters
90 days
to first measurable lift

"We didn't add a single rep. We just stopped sending our best accounts to the back of the same queue as everyone else — the pipeline math changed in one quarter, and finance believed it by the second."

CHIEF REVENUE OFFICER, FORTUNE 100 INSURER
← ALL WORK
CASE STUDY — GLOBAL SAAS SCALE-UP

11,000 hours a year, returned to a six-region marketing org.

INDUSTRY
B2B SOFTWARE
PRACTICE
MARKETING AUTOMATION
MODEL
PARTNERSHIP
TIMELINE
10 WKS + ONGOING
Sepia data plate: agent-run lifecycle automation across six regional lanes, red task packets flowing, beside an audit log column flagging one escalated action.
THE CHALLENGE

Six regional teams were rebuilding the same lifecycle campaigns in parallel — fourteen-day cycle times, copy drifting off-brand, and a compliance team that had vetoed every automation proposal for two years because nothing offered an audit trail.

THE APPROACH

A workflow census costed 240 recurring workflows in hours and error rates — that baseline became the business case. Agents went live in stages: shadow mode first, then human approval, then supervised autonomy, region by region. Compliance got a live audit dashboard before the first agent sent a single email. That's why this one passed where two years of proposals had failed.

240-WORKFLOW CENSUS STAGED AUTONOMY LIVE AUDIT TRAIL
THE OUTCOME
11,000
hours per year returned
14d → 36h
campaign cycle time
100%
of agent actions logged and auditable

"Compliance had killed every automation we'd proposed for two years. This one they approved before it sent a single email — because they could watch every decision it made, in real time."

VP MARKETING OPERATIONS, GLOBAL SAAS SCALE-UP
← ALL WORK
CASE STUDY — NATIONAL RETAILER

340 certified operators in 90 days — an org that ships its own automations.

INDUSTRY
RETAIL
PRACTICE
CORPORATE AI TRAINING
MODEL
ACADEMY
TIMELINE
90 DAYS
Sepia data plate: a grid of operator dots with red ripples spreading certification outward from trained operators, beside cohort progress bars.
THE CHALLENGE

Enterprise AI licenses for thousands of seats — and 9% weekly usage a year later. The transformation program was stuck inside IT, while the merchandising org that stood to gain the most had never been in the room.

THE APPROACH

A one-week capability audit mapped fluency across the org and picked merchandising as the first cohort — the team where skill converted to value fastest. Executive briefings aligned the board on governance and ROI; then four-week practitioner cohorts trained on their own live workflows. Nobody graduated by passing a quiz — every certificate required a shipped automation, reviewed by our team.

CAPABILITY AUDIT SHIP-TO-CERTIFY COHORTS CHAMPIONS NETWORK
THE OUTCOME
340
operators certified in 90 days
92%
weekly AI usage six months later
63
automations shipped by cohorts during training

"We'd bought thousands of licenses and gotten nine percent usage. What changed this time is that nobody graduated by passing a quiz — they graduated by shipping something their own team still uses every day."

CHIEF OPERATING OFFICER, NATIONAL RETAILER

Your case study could be next.

Start a project