We don’t do off-the-shelf. Every engagement here started with a problem that standard tools and generic approaches couldn’t solve, the kind that had already survived a round of consultants, a shelf of reports, and more than one failed implementation. We came in where others had stopped, and we stayed until it actually worked.
What we build isn’t handed over and forgotten. Every solution is designed from the start to keep delivering, adapting as your data grows, your team evolves, and the problem itself shifts shape. Months after the project closes, the work is still running, still learning, still earning its place.
That’s not a promise we make upfront. It’s a pattern we’ve earned over time, engagement after engagement, with clients who came to us because nothing else had worked and stayed because something finally did.
A client needed to migrate 15,000+ projects and 40,000+ related records from a legacy PSA platform into a new system with a fundamentally different data structure without losing relationships, business logic, or five years of reporting history, delivering 100% data integrity, zero downtime, and a 90% reduction in the cost of future migrations. Another client running quarterly employee surveys needed more than a spreadsheet, so we built an automated pipeline to clean, restructure, and analyze their data end-to-end, with results delivered through a branded Looker dashboard, fully operationalizing their executive reporting, surfacing equity and belonging insights for action planning, and leaving a repeatable system in place for every cycle after.
The Customer Science team at HireVue ran structured job analysis surveys for every new client engagement, each requiring manual setup, hours of data wrangling, and spreadsheet-based reporting before anything reached the client, eating a significant chunk of consultant time on entirely repeatable work. Colin and Brittany built an end-to-end system that handles the whole workflow without manual intervention, letting team members configure and launch surveys in minutes through a simple Google Sheets tool, pulling responses automatically, merging them with historical records, and keeping everything in a clean unified dataset, where a short internal form triggers the rest, processing the data, building a formatted Excel report, and emailing it to the requester usually within 15 minutes. Outcome: 800 consultant hours reclaimed per year, human error eliminated from the most critical parts of the process, and a team freed up to focus on the work that actually requires judgment.
Every year, producing the benchmark dataset meant weeks of manual work restructuring and aggregating survey data across hundreds of organizations into a format that standard tools simply weren't built to handle, until we wrote a purpose-built transformation script that automated the entire process end-to-end, producing clean, benchmark-ready and trend-ready outputs on a repeatable, hands-off basis, dropping manual production time from 40+ hours to zero and moving benchmark delivery from roughly a month into the year to the first business day.
Fractional Insights runs live AI transformation workshops for executive and C-Suite audiences where the experience itself is the product, and they needed something purpose-built for an executive room, fully on-brand, and capable of turning participant responses into live insights with no delay or manual work in between, so Broccoli built a custom web application that serves as the backbone of the whole workshop, where participants complete a digital assessment from their phones or laptops, each submission is instantly classified into one of four AI Archetypes with a personalized results screen, while a presenter dashboard updates in real time with a live quadrant chart, a word cloud from open-ended responses, and readiness scores across leadership dimensions, and every participant receives a personalized summary email at the close of each session. Outcome: a level of polish and interactivity most consultants can't match, and the ability to show a room full of C-Suite leaders their own data, live, and use it to drive the conversation in real time.
Running four client engagements simultaneously means the administrative layer never stops, and for Colin the evening routine alone was taking 30 to 45 minutes of scattered effort every night sorting tasks, flagging emails, prepping for tomorrow, and protecting time for deep work, so Broccoli built a custom automation layer on top of Claude Cowork that acts as a true executive assistant, connecting TickTick, Obsidian, Gmail, Google Calendar, Google Drive, Asana, and ClickUp to give Claude a full picture of the workday across all four business contexts, where every evening a single trigger kicks off an end-of-day routine that reviews unfinished tasks interactively, syncs and preps tomorrow's meetings with context from past conversations and decisions, blocks deep work time before the calendar fills up, and writes a polished evening brief into the knowledge base ready to read the next morning. What makes it feel like a real assistant rather than clever automation is how deeply it has been shaped by Colin's working style, with a voice profile that captures how he actually writes and a memory file that tracks patterns over time, so when the system drafts an email or preps a brief it produces something that sounds like him. Outcome: several hours reclaimed per week, meetings consistently prepped instead of cold-started, deep work protected by default, and a workday that ends with confidence that nothing is lost and tomorrow is already set up.
What would you do with the time you're losing to broken processes?
Once a month we share what we are learning from real client work, what is new on the blog, and practical thinking on data and AI for businesses. Subscribe to our newsletter so you won’t miss it.
