About this role
This role is about agent outcomes, not just features. You’ll define what a good outcome looks like, measure it rigorously and keep improving it. You’ll be a senior individual contributor who sets the standard for AI product thinking across the team, working with AI engineering, design and the commercial teams who sell and support Mim.
This role is for someone who has already done it. You have shipped an LLM-based agent to production and owned how its quality was measured. This isn’t a role for learning AI on the job.
- You’re the authoritative voice on the direction and quality of your problem area, and you own its near-term roadmap with Design and AI/ML Engineering
- You define what “good” looks like before anything is built, and validate it against real customer conversations
- You drive outcomes through metrics such as verified resolution rate and conversation quality
- You read real conversation traces to work out why an agent did what it did, then turn the findings into prioritised improvements
- You own your area’s evaluation strategy: golden datasets, scoring rubrics, LLM-as-judge and human review
- You set the quality gates that decide whether a prompt, model or knowledge change is safe to release
- You weigh quality, latency and cost per resolution against each other, including which models to use, and you understand how the definition of a resolution affects revenue, margin and customer trust under outcome-based pricing
- You design how agents behave end to end: workflows, actions, tone, safe failure, and handoffs to human agents that carry full context so customers never repeat themselves
- You make the agent easy for non-technical CX teams to configure, test and understand, and you surface knowledge gaps and failure patterns to customer admins
- You work closely with commercial and customer-facing teams to stay close to how customers deploy and judge the agent
How success is measured
- Verified resolution rate in your area, our primary outcome metric
- Quality of AI-handled conversations: CSAT, accuracy and policy adherence, and hallucination rate, measured through evals
- Escalation quality: timely handoffs that carry full context, with few unnecessary or missed escalations
- Release quality: eval pass rates and regressions caught before release
- Efficiency: cost per resolution and response latency within agreed targets
- Customer time-to-value: how quickly customers configure the agent and reach target performance
- You have 5-10 years in product management and have owned outcomes in a defined problem area or customer segment
- You’ve owned a production LLM-based product or agent for 1-2+ years, including how its quality was measured
- You’ve designed evals and can explain why traditional QA falls short for LLM systems
- You’re fluent in prompting and context engineering, retrieval (RAG), tool calling and agent orchestration, and can challenge architectural trade-offs with engineering (you don’t need to be an engineer yourself)
- You’ve worked in B2B SaaS serving several personas, such as end customers, human agents, admins and buyers
- You use AI tools yourself to prototype, analyse data and move faster
- You work from hypotheses and data, and drive outcomes through a framework such as TARS (target, activated, retained, satisfied) or an equivalent
- You can explain priorities clearly to technical and non-technical colleagues at every level
- Customer service, CX or conversational AI experience
- Experience with outcome-based or usage-based pricing
- Experience running experiments on agent behaviour, such as A/B tests and staged rollouts
- You think independently and stay composed in ambiguity, shipping under uncertainty with guardrails, staged rollouts and monitoring rather than waiting for certainty
- You raise the standard of AI product craft across the PM team, through eval practices, prompt and context reviews and AI-assisted prototyping
- You build bridges between teams that move at different speeds toward different goals
- You take accountability for outcomes, including in high-stakes situations
Who you’ll work with: You’ll report to our CPO, and work closely with Engineering, Product Design, Engineering, other Product Managers and our commercial and customer-facing teams. Dixa has pioneered the concept of Customer Friendship, and this is also something that shapes how we connect and engage at Dixa. We are a friendly bunch, we trust and support one another and we like to have fun together!
We offer:
- Competitive salary and 5% pension on top
- 25 days’ paid annual leave + 5 Dixa Wellbeing days for recharging
- Health care insurance
- Generous paid parental leave for work-life balance
- Delicious subsidized lunch
Apply to Dixa
Clera fills in Dixa's application for you.
How it works
Open the role in Clera
Sign in, and this role opens on your Clera job board, ready to apply.
Clera fills in the application
Clera fills the form from your profile and resume, and asks what's missing.
Track it in Clera
Each application you send shows up in your Clera tracker, sorted by stage.
Ask your AI about this role
It knows you. Clera gives it the role.
First time? Connect Clera once, about a minute.
https://mcp.getclera.comSign in with your Clera account. MCP docs
About Dixa
Dixa is an agentic customer service platform that helps leading e-commerce brands increase operational efficiency and deliver faster and better customer experiences.
Trusted by leading consumer brands of the world, Dixa supports over 30 million conversations annually.
Founded by four friends determined to make a difference in customer service, we're now a fast-scaling company with backing from some of the most reputable investors globally.
- Team size
- 51-200 employees people
- Founded
- 2015
- Industry
- Software Development
- Workplace
- On-site
- Locations
- Copenhagen, Capital Region of Denmark, Denmark
Benefits
- Competitive salary
- Pension
- 25 days paid annual leave
- 5 Dixa Wellbeing days
- Health care insurance
- Paid parental leave
- Subsidized lunch