Anonymized engagements with the numbers that survived audit.
A selection of engagements from across our practice. AI programs at top US banks. SAFe® rollouts inside insurance carriers. Government AI under federal scrutiny. Every number verified before publication.
Every case study below shipped on this arc
01
Weeks 0 to 2
Discover & scope
Sponsor outcome named. Data audit. Draft architecture.
02
Weeks 3 to 6
Pilot in production
Real data, real cohort. Evaluation harness from commit one.
03
Weeks 7 to 10
Productionize
System hardened. Runbook written. First three on-call pairs.
04
Weeks 11 to 22
90-day stabilization
Named principal in steering. Your team owns extension.
Every case study on this page follows the same seven-step structure. Built to read the way an executive sponsor would want to read it at a steering committee.
01
The challenge
The named business problem the executive sponsor cares about most. Stated in their language, not ours.
02
The constraints
The non-negotiables that shaped the work: the people who had to sign off, the systems we had to integrate with, the change posture of the organization.
03
Our approach
The architecture decision. The model and data choices. The named Rockmere Partners principal and the practitioners on the pod.
04
What we delivered
The production system. Named integrations. The runbook your team uses every morning. The artefacts the organization keeps after we exit.
05
The result
Measured outcomes the executive sponsor reported to their board. Numbers we can defend, told in their words.
06
Engagement timeline
Week-by-week shape of the work. The dates the steering committee met. What was decided in each room.
07
What survived past us
The team that owns the system today. The runbook the on-call rotation reads at 2 a.m. The patterns the next engagement reuses.
Outcomes Ledger
A log of what's landed.
Five recent engagements, told in the outcomes the executive sponsors reported to their boards. Numbers are the numbers, not the marketing version.
38% faster
Top-5 US Bank
Financial Services · 2026 Q1 · An investigation copilot used across thousands of fraud cases.
43% less manual
National payer
Healthcare · 2025 Q4 · Eligibility AI built inside the system clinicians already use.
$80M / yr
Mid-market manufacturer
Manufacturing · 2025 Q3 · A Lean operating system rolled across three plants.
87% PI Obj.
Public-sector agency
Government · 2025 Q2 · An ART launched to predictable cadence across nine teams.
27% retained
P&C carrier
Insurance · 2025 Q1 · Demand-planning AI tied to underwriting and renewals.
Anonymization Rules
What we anonymize. What we never will.
Engagement letters dictate what we can publish. We hide identifying data when the letter says so. We never hide the outcome.
We anonymize
What stays out of the public record.
Client name, where the engagement letter forbids attribution.
Individual employee names. We name our own practitioners only.
Internal product names, project codes, and integration partners (we describe the system class instead).
Dollar amounts, where the sponsor prefers percentages.
Geography below the country level.
We never anonymize
What we report exactly as it happened.
The outcome. Percentages and metrics are exactly as the sponsor reported to their board.
The standards the work was held to.
The Rockmere Partners principal who led the engagement.
The engagement duration and team size.
The tools and architecture choices we made.
Clear answers to your questions.
Are these enterprise AI case studies real?
Yes. Every case study on this page is a real engagement, documented at stabilization handover. Each metric is the number the executive sponsor reported to their board, not a marketing figure.
What outcomes can an AI consultancy deliver?
The documented outcomes here include a 38% reduction in bank fraud handle-time, 87% PI Objective achievement on an insurance SAFe® ART, 42% faster Medicaid eligibility dispositions, 11 OEE points in 90 days in manufacturing, and $40M in working capital freed through CPG demand planning.
How do you anonymize client case studies?
Engagement letters dictate what we can publish. We hide identifying detail (client name, internal product names, geography below country level) when the letter requires it, using the system class instead. We never anonymize the outcome, the standards the work was held to, or the named principal who led it.
Are the results audited or verified?
Each metric is the one the executive sponsor reported to their board at handover. We do not publish a number we cannot defend in the room with the people who lived the engagement, and we can arrange a reference conversation for serious evaluations.
Can Rockmere Partners deliver this in my industry?
If regulatory depth and senior judgment change your outcome, almost certainly. Review the industries we know best, then bring us a measurable problem and an executive sponsor and we will tell you on the first call whether we are the right firm.
Have an outcome that has to hold?
Bring a measurable outcome and an executive sponsor. We'll bring the team that delivers it.
