Selected work

Customer Results

Construction · Small business

Multi-agent QA that turned 20 years of unstructured data into defensible business intelligence.

QA cycle time
2 hrs → 15 min
Data restructured
20 yrs
Project margin impact
+30%

Context

A small construction business facing compliance pressure from financial partners needed to extract defensible quality-assurance data from two decades of unstructured project records, without adding headcount or standing up a data team.

Approach

On-site operational discovery to map the legacy workflow, then a multi-agent workflow that organized the unstructured history into a structured technical writing and asset-quality guide. QA checks were automated against the new source of truth, with the human owner kept in the loop for final sign-off.

Outcome

QA turnaround compressed from two hours to fifteen minutes per review, with cleaner, more consistent documentation going to partners. Lower liability risk, and the team freed for higher-value project work.

Read the full story

CPA and financial advisory · 150 people

A skills calibration agent that gave HR two weeks back and the C-Suite a real succession picture.

Annual calibration cycle
2 wks → days
HR capacity returned
160 hrs/yr
Staff calibrated end to end
150

Context

A family-owned CPA and financial advisory firm runs an annual skills assessment and talent calibration across its 150 accountants and advisors. The exercise informs succession, workforce planning, and development decisions for the C-Suite. But the work itself was eating two full weeks of two HR staff every year, organizing inputs, chasing managers, and hand-assembling spreadsheets and PowerPoint slides for executive review.

Approach

We built three things in one engagement: a structured calibration and skills assessment tool the managers would actually use, a simple database that pushed clean records into their existing BambooHR tenant, and an HR agent that handled the manager follow-ups and assembled talent summary reports for the C-Suite. The HR team shifted from running the process by hand to managing exceptions and spending the recovered time on real talent work.

Outcome

The annual two-week scramble compressed into a few days of HR oversight. Manager completion improved because the agent chased on a cadence HR set, instead of two staff playing email triage. The C-Suite now gets calibrated talent reports for succession and workforce planning, generated in days instead of weeks. The firm avoided roughly $50K of annual cost: HR time recovered plus the calibration and succession product add-on they no longer needed to license.

Non-profit · ~12 employees

Readiness assessment and a tiered roadmap that put fundamentals before the AI buzz.

Sequenced recommendations
3 tiers
Active initiatives at a time
1–2
Measurement plan
4 KPIs

Context

A small non-profit Executive Director was getting pulled in several AI directions at once (self-service IT, a chatbot, process automation) with a 12-person team, one IT resource, and limited clarity on where to start.

Approach

Two discovery sessions plus independent analysis to assess data, documentation, tooling, and licensing readiness. Recommendations were tiered from "Zero AI required" foundational work through knowledge-base automation and operational automation, each rationalized by business value and pacing for a small team.

Outcome

A clear "Do Now / Do Next" roadmap focused on fundamentals first (SharePoint content organization, self-service password reset, an AI usage policy) with specific downstream use cases and tooling recommendations sized to the organization’s license level and team capacity.

Get in touch.

Send a short note about what you are working on. No pitch, no pressure. If it looks like a fit, we will set up a time to talk.

Email Joe