What can AI do for Coliving?

The coliving sector has spent the past two years caught in an uncomfortable tension. At every industry gathering, artificial intelligence dominates the conversation. Operators leave conferences with a nagging sense that they ought to be doing something - anything - with AI before their competitors gain an edge. Yet when it comes to implementation, many find themselves paralysed by the question of where to begin. Insights shared at the Coliving Conference 2024 in Amsterdam, Netherlands, suggest the industry is finally moving past aspirational talk toward a more grounded approach of first identifying operational pain points, before finding the technology to solve them.

Can Coliving Afford Not to Automate?

The economics of coliving create unique pressure. As Lea Mishra, Co-Founder & Chief Marketing Officer at POHA House stated, coliving is a distinctive product in the real estate market precisely because operators attempt to deliver extensive services with minimal resources. This is why rather than asking which AI tools are available and then searching for applications, operators are increasingly looking at the daily frustrations their teams face and working backward to the solution. Mishra described her own experience of feeling overwhelmed by endless "10 hottest AI tools" listicles. The breakthrough came when she stopped searching for trending applications and instead asked her operations and sales teams where they were losing time. The answers were immediate and specific - resident verification consumed hours of manual work each week, image generation required expensive photographers, and customer communication demanded constant attention.

This reality has driven POHA House to approach technology from an unexpected direction. They examined each team member's daily tasks, identified time-consuming manual work, and only then sought technological solutions. The revelation was that many operational challenges required automation rather than artificial intelligence. Data entry from housing portals, content updates across multiple platforms, and routine resident communications could be streamlined through triggers and actions before introducing intelligent systems.

The distinction matters. While automation follows predetermined rules, AI introduces adaptive capabilities to simultaneously analyse inquiries, generate contextually appropriate responses, or assess documents against multiple fraud indicators.

Trust, Fraud, and the Challenge of Verification

The rise of AI has created a parallel challenge - bad actors now possess the same powerful tools. Fake applicants and forged documents have become more sophisticated and harder to detect manually. For coliving operators managing high volumes of inquiries and short booking windows, this represents an operational risk. The sector's typically younger demographic and international resident base add complexity to verification processes that must function across multiple document types, languages, and regulatory frameworks.

Hendrick Braun, Managing Director & Co-Founder at rentcard, explained that his company's tenant vetting solution applies AI to solve problems human checkers would sometimes take 30 minutes to complete, depending on complexity. The system verifies identity, assesses payment capability, checks criminal and sanctions databases, and detects sophisticated document fraud that human eyes increasingly cannot spot.

POHA House has integrated rentcard's verification system directly into their Customer Relationship Management (CRM) workflow. When a resident reserves a flat, they automatically receive an email containing a custom verification link with data pre-filled from the CRM, before being prompted to upload documents and banking information. The operator then receives two scores - financial match and social match, utilising a traffic light system. Any team member reviewing the application sees identical information and can therefore make increasingly consistent decisions.

Mishra emphasised the objectivity this creates. In communities comprising multiple nationalities, religions, and backgrounds, a standardised assessment framework removes subjective judgment from initial screening. It also resolves a practical bottleneck. Tasks that previously consumed hours now occur automatically, freeing community managers to focus on resident engagement, rather than document review.

Braun noted that when discussing rentcard with potential clients, AI rarely features prominently in initial conversations. Operators want to understand how the solution addresses their fraud protection and efficiency problems. The underlying technology becomes relevant only when clients ask how the system achieves near-instantaneous verification. This suggests the sector has moved towards a more mature evaluation of what specific technologies can deliver, instead of using AI for its own sake.

Removing Bias: Data Quality and the Human Layer

As operators adopt AI-powered tools, questions of data quality and algorithmic fairness have emerged as critical considerations. Dorota Iskra, Senior Director of AI at DataForce by TransPerfect, brought a perspective grounded in years of building training data for AI systems. She warned that applications fail when organisations possess extensive data, but lack clarity about the problem they aim to solve with it. The temptation to deploy AI simply because data exists leads to unsuccessful implementations that do not measure outcomes or serve genuine operational needs.

Equally important is ensuring that training data represents all populations fairly. Iskra pointed to the familiar example of image search results that return homogeneous demographic representations. In coliving, where resident diversity is often a selling point, verification systems must perform equally well across nationalities, document types, and demographic groups. This requires careful manual annotation and ongoing refinement. Whilst these activities can appear expensive, Iskra argued they are essential investments that determine whether an application succeeds or introduces bias into operational decisions.

The human element remains central, even in highly automated processes, with Braun confirming that rentcard incorporates human checks into its verification workflow. Iskra emphasised that AI should handle mundane tasks while preserving space for human creativity, judgment, and correction. The quality difference between AI-generated and human-crafted communications remains clear - effective implementation uses AI as a starting point, with humans refining outputs and handling exceptions.

Making AI Work in Practice

For operators considering where to begin, Mishra recommended focusing on sales and operations - the areas where coliving shares processes with other sectors and where mature tools already exist. AI-powered guest chat systems can reduce support ticket volumes. Document verification platforms eliminate manual screening bottlenecks. Generative AI tools like ChatGPT or Gemini support content creation, email drafting, and meeting documentation. The key is integration with existing infrastructure, rather than adopting isolated tools.

For example, POHA House uses Gemini specifically because it operates within their Google Workspace environment. This allows team members to generate emails directly from their inbox, create meeting notes within their collaboration platform, and access AI capabilities without switching applications. The friction of moving between systems reduces adoption and creates inefficiency.

Iskra reinforced the importance of customisation. Off-the-shelf AI tools deliver generic results, while operators who invest in fine tuning their models to their brand voice, client base, and specific use cases see substantially better performance. This applies particularly to conversational AI. When property managers deploy AI-powered phone systems to handle resident service requests, those systems routinely perform better when trained on the operator's historical call data, common issues, and resolution patterns. Generic models lack the context to handle sector-specific inquiries effectively.

The panel also addressed a practical reality - many coliving operators have not yet fully automated their processes, let alone implemented AI. Mishra acknowledged that POHA House spent years focused on core operations before turning attention to automation. This sequential approach may actually prove advantageous. Understanding workflows deeply before introducing intelligent systems makes it easier to identify where AI adds value and where simpler automation suffices.

The Path Forward

The next wave of AI applications in coliving is beginning to take shape. Braun described a recent development in property management where an AI-powered voice system handled tenant service calls in real time. The system answered more than 80% of inquiries without human intervention, drawing on data from past calls to diagnose problems and trigger solutions. As a result of this, tenants received immediate service instead of waiting in a queue, and property managers gained capacity to focus on complex issues requiring human judgment.

Mishra identified three areas where POHA House sees the greatest potential - creating visual content including floor plans and renderings, customer service where the expectation of around-the-clock availability creates staffing challenges, and using AI for sustainability, such as to optimise energy consumption based on occupancy patterns and weather forecasts.

Iskra highlighted the importance of fine-tuning AI models to the specific needs of each operator. Off-the-shelf solutions provide a starting point, but performance improves dramatically when the system is trained on an operator's own data, learning the language residents use and the brand voice the operator wants to project.

The coliving sector has moved beyond the anxiety of needing to "do something" with AI and toward a more strategic stance. The path forward is less about keeping up with the newest trends, but more about solving real problems in ways that free people to do what they do best - create spaces where residents feel at home.

‍