Senior Agent Architect
About the job
- 3+ years of experience in enterprise customer-facing roles, with proven expertise in solution architecture, systems architecture, or enterprise integration.
- Strong systems thinking, with a track record of solving complex technical problems across multiple systems and data sources.
- Experience with complex SaaS implementations at scale (CRMs, ERPs, contact center platforms, ticketing systems).
- Proficiency in REST API design, JSON/XML data structures, API testing tools (e.g., Postman), and API authentication patterns (OAuth, API keys, JWT).
- Experience reading and consuming API documentation across enterprise SaaS platforms, including data transformation and field mapping logic.
- Analytical and critical thinker with expertise in risk assessment and problem-solving.
- Strong project and stakeholder management skills, with a passion for meeting milestones and effective communication.
- Entrepreneurial mindset with initiative, results orientation, and ability to identify opportunities for impact.
- Ability to analyze customer requirements and translate them into LLM prompts and conversational flows, with attention to subtle conversational nuance and A/B testing to assess agent performance.
- Prior conversation design, prompt engineering (e.g., chain-of-thought, few-shot learning), or A/B testing experience is a plus, not a requirement - we will train this.
- Fluency in French and English is required
- Design end-to-end solution architectures for AI Agent deployments, mapping how Parloa's platform integrates into customers' enterprise systems (CRMs, ERPs, ticketing systems, contact center platforms).
- Build and configure integrations using integration platforms and APIs, applying systems thinking to solve complex, multi-system data flows.
- Collaborate with Forward Deployed Engineers on complex or custom integration scenarios that go beyond standard integration capabilities.
- Apply structured testing approaches to validate architecture, integration reliability, and AI agent performance under real-world conditions.
- Identify and resolve blockers together with other departments at Parloa (e.g., Product, Solution Engineering, Sales) and the customer.
- Scope AI Agent deployments, providing strategic architecture advice and execution support to customers and partners.
- Use your knowledge of LLM internals (e.g., embeddings) to analyze customer requirements and design prompts for reliable, user-aligned agent behavior.
- Translate complex customer workflows into structured conversational components that LLMs can execute reliably.
- Fine-tune conversational flows and voice output (e.g., SSML, Lexicons, Regex) to align with customer brand standards.
- Document best practices, how-to guides, and product capabilities for internal and external audiences, representing the expertise of the Agent Architect team.
- Prior conversation design, prompt engineering (e.g., chain-of-thought, few-shot learning), or A/B testing experience is a plus, not a requirement - we will train this.
About the company
Parloa builds an AI Agent Management Platform for enterprise customer experience across voice, chat, and messaging. The platform lets organizations design, deploy, and manage AI-powered conversational agents and route interactions across channels, with CRM/CCaaS integrations to connect data and workflows. It coordinates multi-channel conversations through a centralized management layer, enabling training, testing, and deployment with connections to enterprise stacks for handoffs between AI and human agents. Its goal is to help large organizations deliver consistent, data-connected customer experiences by coordinating AI agents across channels and integrating with CRM/CCaaS ecosystems.
Company Size
501-1,000
Company Stage
Series D
Total Funding
$564.3M
Headquarters
Berlin, Germany
Founded
2017
Get referred to Parloa
See people who can refer or advise you
Simplify's Take
What believers are saying
- September 2026 WAVE Berlin showcases new products to 700 enterprise buyers.
- June 2026 Alvaria integration opens compliant outbound CX in regulated industries.
- January 2026 Series D funded London, France, Spain, Milan, and New York expansion.
What critics are saying
- Zalando partnership remains unannounced, underscoring sales claims without contractual proof.
- Salesforce Agentforce, Google, Zendesk, and Microsoft absorb agent-management features by 2027.
- OpenAI and Microsoft infrastructure dependence creates platform risk if pricing or policies shift.
What makes Parloa unique
- SAP Service Cloud integration and strategic SAP investment lock enterprise distribution.
- Parloa runs voice, chat, WhatsApp, and Teams across 140 languages.
- Its Agent Management Platform emphasizes simulation, evaluation, and deployment controls for enterprise CX.
Help us improve and share your feedback! Did you find this helpful?
Benefits
Company Equity
Growth & Insights and Company News
6 month growth
0%1 year growth
-2%2 year growth
0%Berlin-based AI startup Parloa has become a unicorn, with a valuation of more than $3 billion. It recently completed a funding round and has nearly 600 employees. Founded in 2018, Parloa builds "agent-based AI" that powers the hotline at Berlin's BER Airport. The system analyses airport operations to answer complex or vague queries precisely. The company faces strict regulation, a risk-averse culture, and limited domestic capital for scaling. Scaling often requires funding from international investors. Parloa is part of the "Made for Germany" initiative, which aims to foster local innovation. Founder Stefan Ostwald says a more integrated ecosystem is needed to unlock the potential of German startups.
Parloa on reinventing the customer experience with AI Agents. Parloa was recently recognized as an AI-Powered Enterprise Agent Solution of the Year for its AI Agent Management Platform, AMP. Tech Ascension Awards sat down with Malte Kosub, CEO and Co-Founder of Parloa, to discuss why agentic AI represents a fundamental shift beyond traditional chatbots, what enterprises need to rethink before deploying AI agents at scale, and how autonomous systems could reshape the relationship between companies and their customers. Parloa's AI Agent Management Platform (AMP) was named AI-powered Enterprise Agent Solution of the Year. What does this recognition say about where the market is headed, and what makes AMP's approach to agent management different from the field it beat? Winning AI-powered Enterprise Agent Solution of the Year confirmed the market has moved past proving AI Agents can hold a conversation and into proving they can run as production infrastructure. AMP was built to go beyond voice AI layers and deliver a full agent lifecycle at global scale. AMP unifies the entire lifecycle, design, testing, deployment, and optimization, inside one environment, while most competitors split those stages across separate tools. It also runs an evaluation-first architecture, testing every agent through thousands of synthetic conversations with LLM-as-a-judge scoring before it ever reaches a customer, on a foundation built compliant with GDPR, ISO 27001, and SOC 2. And its Agent Composition feature lets a company build an agent once and deploy it across more than 130 languages and 70 countries without rebuilding for every new market, something competitors still require. That's the difference maker for its customers: one platform, one build, full confidence. Parloa's own research found that 99% of companies failed your first agentic-readiness evaluation. How does AMP close that gap? Its 2026 research on the State of Agentic CX evaluated thousands or websites, chats and phone trees. The findings were bleak: 43% of company websites gave customers no clear path to support, fewer than 10% of chat conversations actually reached the customer's goal, and the vast majority of voice experiences still ran on decades-old automation or none at all. The companies that failed its evaluation came down to two consistent markers. Companies are overestimating their own AI-readiness, discovering mid-deployment that they were lacking the APIs, MCP connections and general architecture needed for success. And more critically, many companies misunderstand how to apply agentic AI in the first place, treating it as a conversational version of the same IVR tree they've been using for years rather than an opportunity to redesign the customer journey. AMP solves for both these by running a readiness assessment before any agent hits the field, and surfacing data gaps up front. How should enterprise leaders think about agentic CX differently from the conversational AI or chatbot investments they may have already made? Most leaders are still evaluating agentic CX with the same mental model they used for chatbots, and that is the first mistake. A chatbot answers a question and forgets you the moment the session ends. An agent remembers your last issue, your preferences, and your history, and it carries that context across the app, the web, and the phone. That is not a feature upgrade. It is a completely different relationship with the customer. The old generation of automation ran on fixed branch logic: if this, then that, and it never veered off script. Agentic systems reason. They pick up on tone, pace, and frustration, and they take action instead of just routing a request. What I keep telling leaders is that most agentic pilots look great in isolation, one channel, one use case, and then fall apart the moment agents have to coordinate across backend systems or hand off between channels. That breakdown is an orchestration and infrastructure problem, not a model problem, and no one solves it by buying a bigger chatbot. Why does that matter? If you measure agentic CX the way you measured chatbots, by tickets deflected or average handle time, you will optimize for the wrong outcome. The smartest leaders are instead asking whether they made life easier for the customer, and whether that drove loyalty or revenue. If a brand's real goal is pure cost cutting, a high quality AI solution is probably not the right fit for them, and that choice will cost them customer loyalty in the long run. The organizations that treat this as a bigger chatbot budget line end up cutting labor costs without ever changing the actual customer relationship, and they leave the real value on the table. As enterprises move from AI pilots to production, what will separate the organizations that successfully scale agentic CX from those that remain stuck in experimentation? The companies stuck in pilot purgatory almost always made the same early decision: they picked a low complexity use case, something like an account balance lookup, because it was easy to prove out. The problem is that proving an easy use case does not tell you anything about how to scale, and it does not force the internal process change that agentic AI actually requires. It just confirms the technology works on something simple. The organizations that break through, do the opposite. They start by mapping what the end to end customer journey should look like, and only then apply AI to that redesigned journey, instead of layering AI on top of a process that was never built for it. They also stop evaluating vendors on demos. If you hand three or four vendors the same narrow use case, they will all give you a similar answer, and that tells you nothing about which platform can actually deliver value at scale, across languages, load, and architecture, over the long term. The last piece is governance, and it has to be built in from day one, not bolted on after volume increases. Can the agent explain its own decision? Can it escalate with context? Do your teams have real time visibility into what it is doing and why? If the answer is no, you are not ready to scale, no matter how good the pilot looked. What's next for Parloa and AMP as enterprise adoption of AI agents continues to grow? The phase of ultra cheap, subsidized AI is ending. Compute costs are rising, models are getting more complex, and enterprises expect security, reliability, and customization at scale. That is exactly where Tech Ascension Awards is putting its energy: continuing to build the evaluation and governance layer that lets enterprises trust an agent before it ever reaches a real customer, not just after something breaks. Tech Ascension Awards also believe the next era of customer experience is personal. Every customer should have an agent that recognizes them instantly, whether they are in the app, on the web, or on the phone, and that agent should already know their history instead of starting over with a stranger every time. That is one of the key drivers Tech Ascension Awards is building toward. And Tech Ascension Awards is going to keep doing the research that got Tech Ascension Awards here. Its State of Agentic CX work showed Tech Ascension Awards that the market is still early, only about one percent of CX systems today can successfully manage agent to agent interaction. That gap is the opportunity, and closing it for its customers is what comes next for AMP.
Parloa to unveil what's next in agentic customer experience at WAVE 2026 Berlin. Sep 24, 2026, 09:19 ET New product innovations, live demos and insights from leaders at OpenAI, Microsoft, SAP, IKEA, HealthEquity and more will take center stage at Parloa's flagship AI + CX event NEW YORK and BERLIN, Sept. 24, 2026 /PRNewswire/ - Parloa, the leading agentic CX company for enterprise customer service, today announced new details for WAVE 2026, its flagship AI and customer experience conference taking place October 7-8, 2026 at Wilhelm Studios in Berlin. At WAVE 2026, Parloa will discuss emerging trends in how AI and CX converge, giving attendees an early look at new capabilities designed to improve agent performance, enable more intelligent customer interactions and help enterprises scale agentic CX with greater control. More than 700 enterprise leaders will gather in Berlin for two days of keynotes, customer stories, panels, master classes and live demos featuring executives and experts from the world's most customer-centric brands in every industry. The Next Wave of Innovation from Parloa As enterprises move from experimenting with AI to putting it into production, the focus is shifting from what AI agents can do to how reliably they can perform, improve and deliver impact at scale. At WAVE, Parloa will unveil innovations across its AI Agent Management Platform focused on: * A clearer view of agent performance: New ways for enterprises to measure quality, evaluate outcomes, benchmark performance and understand where AI agents are succeeding and where they need refinement. * Agents that get better over time: Advances in testing, simulation, evaluation and optimization designed to help enterprises continuously improve agent performance. * More natural customer interactions: Richer experiences across voice, chat and other modes of communication that give customers more flexible ways to engage with AI agents. * More intelligence in every interaction: Deeper context designed to help AI agents better understand the customer, the conversation and the situation at hand. * Greater enterprise control: Capabilities designed to help global organizations meet security, data and deployment requirements while scaling AI across markets and use cases. Live demos will give attendees a closer look at how these capabilities can be applied across enterprise customer experience. Parloa CEO and co-founder Malte Kosub, Chief Marketing Officer Latané Conant and other Parloa leaders and builders will also share perspectives on the future of agentic customer experience and what its evolution means for enterprises, brands and customer relationships. Hear From the Enterprises Putting Agentic CX to Work More than 30 executives and experts from across Parloa's customer and partner ecosystem will share firsthand perspectives on how agentic AI is reshaping customer experience, operations and enterprise strategy. Speakers and participants include leaders from OpenAI, Microsoft, SAP, IKEA, HealthEquity, Volkswagen Financial Services, OBI, Booking.com, AUTO1, Deloitte and more. The program will also feature former professional footballer and World Champion, Bastian Schweinsteiger. WAVE will tackle the decisions enterprises face as AI moves deeper into production: where to invest, how to scale, how to govern AI systems and how to prove business impact. What Attendees Can Expect at WAVE 2026 Across keynotes, customer stories, panels, master classes and live demos, attendees will explore: * How leading brands are scaling agentic CX * How to prove business value as AI moves deeper into CX * How AI is reshaping CX teams, roles and KPIs * What reliable AI agents require at enterprise scale * What's next for AI agents, regulation and data sovereignty Be Part of WAVE 2026 in Berlin General admission tickets are €799 for WAVE 2026, taking place October 7-8 at Wilhelm Studios in Berlin. Following Berlin, Parloa will bring WAVE stateside with WAVE Summit in New York on November 4-5, convening a select group of CX and business leaders for a more intimate conversation about where agentic AI is headed and what enterprises need to prepare for next. About Parloa Parloa empowers global enterprises to build, train and manage agentic AI solutions for premier customer experiences. Founded by Malte Kosub and Stefan Ostwald, Parloa began with the belief that every great conversation is the start of a relationship, a principle that still guides how the company builds technology today. Leading global brands use Parloa's advanced AI agents to improve service at scale, increase customer loyalty, and unlock new revenue. Parloa employs more than 600 people across eight offices throughout the US and Europe. SOURCE Parloa
Parloa says it is partnering with Zalando on AI customer conversations. Malte Kosub announced the partnership himself with no scope, no markets and no timetable, seven weeks after Article 50 of the AI Act began requiring that people be told when they are talking to a machine September 22, 2026 - 4:13 pm Parloa co-founder and chief executive Malte Kosub said Zalando will use the Berlin company's AI for customer conversations across Europe, though neither company has published an announcement and Zalando has not commented. Article 50 of the AI Act, applicable since 2 August, puts the duty to disclose that a user is talking to an AI on the provider rather than the retailer. Zalando will use the Berlin company Parloa's AI for customer conversations across Europe, co-founder and chief executive Malte Kosub said in a post on Monday. Neither company has published an announcement. Zalando has not commented. That is one half of a deal. Kosub gave no scope, no markets, no timetable and no indication of whether anything is running yet. Parloa raised $350M in January at a $3B valuation, on $50M of annual recurring revenue and 150% net revenue retention, taking total capital past $560M. Zalando is a large place to land. It counted 62 million active customers across 29 markets and more than 7,000 brands when it launched its own AI assistant in November 2023, co-created with customers rather than bought in, who rated its suggestions with a thumbs up or down. Whether Parloa replaces that assistant, sits beside it or takes the phones is unanswered. The timing is the part that matters. Article 50 of the AI Act became applicable on 2 August. It requires that systems built to interact directly with people are designed so those people are informed they are dealing with an AI, and the duty sits with the provider rather than the shop. European retail has been arguing about that article. The sector asked Brussels in June to exempt AI-generated advertising from the same article's labelling rules, calling the burden disproportionate to the risk. Breaching it runs to EUR 15M or 3% of global turnover. The Commission published a voluntary code of practice the same month. Parloa is European. Its stack is not. The platform runs on Microsoft Azure and uses OpenAI models, and its named customers so far are Allianz, Booking.com and HealthEquity. SAP has invested in the company and built it into its Service Cloud. A German retailer choosing a German vendor still routes the conversation through American infrastructure. That is the question worth putting to both. Zalando's customers will be talking about orders, returns and refunds, in 29 markets and as many languages. Who processes that, where it is processed and on what legal basis is not in a post on a social network. It is in a contract neither company has described.
WAVE 2026: Meet the minds behind AI and CX transformation. Julie Goelles VP Marketing Home > blog > Article September 11, 2026 - 4 mins Next month, over 30 speakers and 700 AI and CX innovators will take over Wilhelm Studios in Berlin for WAVE 2026, Parloa's flagship conference that explores the intersection of AI and CX. From October 7-8, leaders in enterprise customer experience, data, and artificial intelligence will collaborate on where agentic CX is headed through workshops, panels, and networking. The WAVE 2026 preview covered the six major themes shaping the event: the AI reality check, new rules for AI-era customer service, moving from pilot to production, the CX business case, governance, and what's next in agentic CX. Now, it's time to introduce the people taking the stage. Senior leaders from OpenAI, Volkswagen Financial Services, Allianz Partners, SAP, Microsoft, Booking.com, OBI, and more will share what they've found to work (and not work) with agentic CX, and where they're looking next: Day 1: build with Parloa experts. No slides. Just Parloa experts and partners working alongside you to build AI agents, observe them, and organize teams to work with them. Marco badorrek, director of solution engineering, Parloa. During Build your own voice AI agent, Marco will walk participants through customizing a pre-built Parloa agent with Parloa Navigator's no-code workflow. Andrey shilin, principal Product manager | Ni xiong, engineering manager, Parloa. During See what your agents are actually doing with Parloa Lens, Andrey and Ni will demonstrate how AI-powered performance monitoring flags issues with agentic interactions in real time and unpack the scoring methodology behind it. Ziv baum, principal Product manager | Rômulo luzia, lead agent architect for strategic projects, Parloa. In Self-improving agents: Building your optimization loop, Ziv and Rômulo will demonstrate how elements of Parloa Lens and Navigator work together to put expert-level optimization in the hands of business teams. Frauke von polier, AI transformation board member, & former CPO at zalando & SAP | Corina taban, CEO, 934 advisors. Technology is only half the transformation. In The hybrid organisation: A practical blueprint for human + AI teams, Frauke and Corina will break down the human side of redesigning an organization around AI. Deloitte brings a client's-eye view to scaling AI in customer service. During What successful companies do differently, Hans, Thomas, and Nicholas will unpack the four dimensions of AI transformation, from business value to customer journeys, governance and trust, and the operating models that separate AI initiatives that scale from those that stall. Day 2: Production Stories from those who have automated already. Following a keynote address from Parloa CEO and Co-Founder, Malte Kosub, including major innovation updates presented by Chief Product Officer, Csaba Tamas, and VP of Product Marketing, Kevin Boyer, Day 2 will highlight stories of scale, learning, and how teams are preparing for what's coming next. Laura Modiano, Founder Experience, OpenAI will join Malte on stage to talk about why the future of customer experience is now, and what's coming next. Pieter Viljoen, Chief Data & AI Officer, Allianz Partners will provide a candid look at what it takes to scale complex AI integrations across markets, business units, and high-impact use cases. Tobias babic, head of customer process & data management, Volkswagen Financial Services | Ben ellermann, managing partner at MUUUH! In a panel, Tobias and Ben will unpack what it takes to treat strategy, build, and enablement as one motion, where the line between external expertise and internal ownership belongs, what's changed for the people building the agents over the last two years, and what they'd decide differently today if they were just starting out. Marco casalaina, VP products, coreai and AI futurist, Microsoft | Stefan ostwald, Co-Founder & CAIO, Parloa. Most voice AI failures trace back to a handful of preventable design mistakes. Marco will break down the four cardinal sins of building a voice AI agent, and Stefan will demo the fixes live, turning theory into a playbook you can take away. Bastian schweinsteiger, football world champion. After lunch, the conversation shifts from technology to leadership. Bastian will talk virtually about how to lead under pressure. Prof. Dr. Philipp hacker, professor of law and ethics of the digital society | dr. Carolin gütschow, director of legal, Parloa. Philipp and Carolin will bring a regulatory angle to the conversation, grounding the practical, industry-specific realities of deploying AI agents in the strategic role of governance. Further insights from industry leaders. Senior leaders from HealthEquity, IKEA, OBI, Booking.com, AUTO1 Group, Transcom, Advanzia Bank, OBjektkultur, and assona will share how they've deployed AI agents in business-critical use cases first and then scaled them across countries and languages. They'll reveal what made their pilots successful and what they'd recommend to companies just starting out. Be part of what's next. More speakers are being added every day. Check out the full lineup and get the full agenda here. WAVE 2026 runs October 7-8 at Wilhelm Studios, Berlin. See you there.