Blog
Our team writes about building voice agents, including queueing maths, evaluation, fine-tuning and the lessons from shipping phone agents.
All posts
Explorations with agentic: extending the contact centre experienceMost calls end and the relationship ends with them. We're building the pattern that turns a call into an ongoing, one-to-one line to your business, in the messaging apps your customers already use.
Auto scaling for contact centresLosing half of a 4-person contact centre doesn't double wait times. It makes them infinite. The century-old queueing maths behind every meltdown, and the release valve we built for it.
Introducing warm transferYour AI agent answers the call, checks if you want it, and connects you live, or takes a message when you can't pick up.
Why your quality assurance scorecard is driving your team crazy (and how to fix it)Traditional QA scorecards review only 1-3% of calls with metrics that contribute to agent attrition. Learn how AI-powered, objective scoring with 100% coverage can transform your customer service team.- Get started with fine tuning on your MacBookHow to train specialized AI models for PII masking on consumer hardware using mlx-lm, Qwen3-0.6B-4bit, PEFT, and QLoRA
The problem with frontier modelsHow companies end up putting erratic AI geniuses on customer support- Without evals, AI development moves slowerWhy proper evaluation systems are essential for maintaining development velocity as your AI applications grow in complexity
Crossing the uncanny valley of AI voiceHow to handle function calls in voice AI applications without breaking user trust through awkward silences
Just Go LiveWhy you should skip the giant data transformation project and start building AI systems that learn from real-world feedback- Building equity in AI systemsHow companies can build lasting intellectual property in AI systems by focusing on evaluation frameworks rather than prompt engineering
Get started
Tell us what your callers ring about, and we build and test the agent for you.