Our Work
Real transformations.
Not proofs of concept.
We don't talk about what AI could do. Here's what we've built — across 7 engagements and counting.
From one-line call logs to AI-powered clinical intelligence.
Before
Call center agents logged incidents as one-line notes. "Client called for assistance. Complained about difficulty getting groceries." Critical health signals — medication risks, social isolation, fall danger — disappeared into flat text.
What We Built
- AI-powered call analysis that extracts medical concerns from natural conversation
- Automatic Social Determinants of Health detection — surfacing transportation barriers, caregiver gaps, isolation risks
- Tiered intervention recommendations prioritized by severity
- Visual dashboards giving supervisors instant population-level visibility
After KLM
The same call that produced one sentence now generates a multi-layered clinical report — catching risks before they become emergencies. Agents developed sharper clinical instincts. The AI works with them, making every person on the team more effective.
From weeks of video production to hours — with AI avatars.
Before
Creating professional video content meant coordinating talent, studios, and editing timelines. A single explainer video took weeks and thousands of dollars. Teams that needed video for training, onboarding, or marketing simply couldn't keep up with demand.
What We Built
- AI avatar-based video pipelines using Synthesia and D-ID for lifelike presenter videos
- Scripted content workflows that take a brief to finished video in under a day
- Multi-language avatar generation for global teams — same video, 12+ languages
- Branded template systems so teams produce on-brand video without design support
After KLM
Video production dropped from weeks to hours. Teams now produce training content, product walkthroughs, and marketing videos independently — no studios, no scheduling, no bottleneck. Output went from 2-3 videos per quarter to 20+ per month.
Custom microsites that convert — built and deployed in days, not months.
Before
Partners needed dedicated landing experiences — co-branded pages, tailored messaging, conversion-optimized flows. But the internal team was backed up for weeks, and agencies quoted 6-8 week timelines. Partners were going live with generic one-pagers that didn't convert.
What We Built
- A rapid microsite framework that produces custom, high-converting partner pages in 2-3 days
- AI-assisted copy generation tailored to each partner's audience and value proposition
- Integrated analytics and conversion tracking from day one
- Reusable component library that maintains brand consistency while allowing partner customization
After KLM
Partner microsites now launch in days instead of months. Conversion rates improved 3x over the generic pages they replaced. The framework is reusable — each new partner site takes less time than the last, and the team can spin up new experiences without engineering support.
We build the tools that make other builders faster.
Before
Every team we worked with hit the same bottleneck. Documents everywhere. Structured data nowhere. PDFs, contracts, medical records, onboarding docs — every project reinventing the parsing wheel.
What We Built
- AI-powered extraction from complex, unstructured documents via LlamaParse
- Natural language querying with cited answers
- Template-based analysis with reusable placeholders
- A block editor for building custom parsing logic without code
- Agentic pipelines that reason across document sections and produce structured output
After KLM
We use DataLoom internally. We deploy it for clients. We've used it to help development teams accelerate their delivery workflows. This is infrastructure we built because the problem was universal.
From manual spreadsheet chaos to automated operational intelligence.
Before
Operations teams were living in Smartsheet — tracking projects, managing timelines, coordinating cross-functional work. But everything was manual. Status updates required chasing people down. Reports took hours to compile. Data lived in silos, and leadership had no real-time visibility into project health.
What We Built
- Bi-directional Smartsheet API integrations that sync project data across CRM, ticketing, and communication tools
- Automated status rollups and health scoring — no more manual weekly reports
- AI-powered anomaly detection that flags at-risk projects before deadlines slip
- Custom dashboards pulling live Smartsheet data into executive-ready views
After KLM
Weekly reporting that took 4+ hours now happens automatically. Project health is visible in real time. At-risk work gets flagged before it becomes a fire. The ops team reclaimed over 15 hours per week in manual coordination — and leadership finally trusts the data.
Production-grade voice AI — shipped, not prototyped.
Before
Voice AI sat on every product roadmap but never shipped. Too complex, too many moving parts, too far from what teams knew how to build.
What We Built
- Realtime AI personas with emotional intelligence, contextual memory, and natural conversation
- ElevenLabs-powered voice synthesis with precise control over tone, character, and emotional depth
- AI-powered image and video processing for production workflows
- A sandbox environment where teams explore and validate voice AI use cases before deployment
After KLM
Teams moved from "we should look into voice AI" to "it's live and we know how to extend it." The technical foundation is in place and the capability keeps compounding.
One embedded analyst. Four functions transformed in six months.
Before
A mid-market services company had tried three different AI tools across different teams. Nothing stuck. Ops was still doing manual data entry. Engineering had an unused Copilot license. Product was making decisions on gut feel. CS was drowning in tickets with no prioritization.
What We Built
- Deployed AI-assisted triage and routing in Linear — reducing ticket response time by 60%
- Built Slack workflows that surface CRM signals and summarize async threads automatically
- Integrated Claude-powered analysis into the product team's weekly prioritization process
- Created AI-assisted customer health scoring in Attio with proactive churn alerting
After KLM
Within six months, all four functions were operating with AI as a daily tool — not an experiment. The team built and maintained their own automations. The AI Business Analyst didn't just deploy systems — they created internal champions who carried the capability forward across the organization.
Have a challenge that looks like one of these?
Talk to Us