In a major recognition of our vision to automate complex business processes with AI, we are honored to be selected to join Meta‘s inaugural Llama Incubator Program Singapore 2025. The selection places us in a six-month, hands-on programme built to help startups and SMEs integrate open-source large language models into real-world products and operations.
This is more than a badge. For Gleematic, which already automates workflows across finance, HR and operations, access to Llama’s tooling and Meta’s technical mentorship means the company can push beyond rules-based automation into generative, language-aware automation — the kind that understands context, drafts responses, extracts meaning from messy documents, and negotiates multi-step tasks with fewer human handoffs.
About Meta Llama Incubator Program 2025
13 March 2025 – Meta launched the Llama Incubator as part of a broader effort to accelerate responsible, open-source AI innovation in Singapore. This program is in partnership with the Singapore Government and key agencies including IMDA, GovTech, Enterprise SG, AI Singapore, and SGInnovate. The programme brought together government agencies, industry partners, and local AI players to give startups and SMEs practical access to Llama models, safety toolkits, mentoring and (in some cases) funding. Over 100 organisations participated in early workshops; 40 digitally-ready teams were selected for the six-month accelerator where they would receive technical and business mentoring, training on safety and deployment, and support to take Llama-powered prototypes to market.
The incubator deliberately mixes technical enablement with trust and safety training. Participants were taught to use Llama’s protection tools alongside local safety toolkits such as IMDA’s AI starter kit and open-source testing frameworks — an approach that aims to produce useful systems that are also auditable, secure, and aligned with local regulatory expectations. The cohort work culminated in a Demo Day where teams showcased Llama-powered solutions and several projects won cash awards.
The Programme Structure: Practical Help, Safety Training, and Ecosystem Support
Over the six months, the selected cohort of 40 startups and SMEs received an intensive blend of support.
The Llama Incubator was set up with two complementary tracks. The Startup Track focused on helping newer companies refine their MVPs, validate prototypes, and build go-to-market strategies with investor readiness. The SME Track emphasised hands-on customization and operational deployment so established businesses could embed Llama models into internal processes. Across both tracks, participants received:
- technical mentorship on model fine-tuning and API usage,
- business mentoring on go-to-market and scaling,
- training on safety frameworks and testing, and
- credits/support from cloud partners to help ship prototypes.
Ecosystem support was a major part of the program as well. Meta partnered with AWS Singapore to provide eligible teams with cloud credits and technical advisory, helping them accelerate their product development.
The incubator also connected teams to regional partners and potential customers, reflecting Singapore’s play to become a permissive, well-governed testbed for practical AI that scales across ASEAN and beyond. That mix of local governance know-how and global technical aid was a deliberate design point — to spur innovation while managing risk.
All of this hard work culminated in the Demo Day on 15 October 2025, held at Meta’s Singapore office. At the event, participants showcased more than 30 Llama-powered solutions across sectors like finance, public services, education, and healthcare.
Meta Llama Incubator Program: Why Gleematic’s Inclusion is a Smart Fit
Gleematic’s core product is automation that lets “everyday” business users create and run bots without heavy engineering. That no-code posture plus experience in process automation makes us a natural adopter of LLMs: where rule engines struggle with ambiguity or messy language inputs, generative models can supply flexible understanding, summarisation, document extraction, and natural language interfaces.
Gleematic can combine its existing automation capabilities with Llama’s strengths in text, reasoning, and dialogue to build higher-value features such as adaptive document processing that handles diverse invoice formats, emails, contracts, and unstructured notes, and smart exception handling where bots use natural language to ask clarifying questions, summarize problems, and suggest next steps.
Beyond these, the integration of LLMs unlocks even more powerful capabilities:
- Autonomous workflow orchestration: Bots can plan and adjust multi-step processes based on context, not just fixed rules — enabling smarter routing, follow-ups, and automated task sequencing.
- Knowledge-aware automation (RAG-enhanced): By retrieving information from internal SOPs, manuals, and policy documents, LLM-powered bots can validate decisions, check compliance, and ensure processes align with the latest guidelines.
- Intent-to-action automation: Llama can understand emails, tickets, and messages, classify intent, extract key details, and automatically trigger the correct workflow — whether that means updating a system, generating a document, or escalating a case.
- Generative form filling & data transformation: From incomplete or messy texts, the system can infer missing fields, normalize data formats, and convert human language into structured information that downstream systems can use.
- Automated narrative reports: Instead of producing raw logs, bots can generate readable summaries, audit trails, and compliance reports — complete with explanations, highlights, and reasoning grounded in workflow data.
- Policy and risk assessment automation: LLMs can interpret regulations and compare them against submitted documents, helping bots flag inconsistencies, assess risks, and draft recommendations for reviewers.
- Natural-language automation-building: Users can simply describe what they want — “extract invoice data, validate it, and email suppliers for missing items” — and the system can map those instructions into automation steps.
- Intelligent monitoring and predictive alerts: By analyzing patterns in system logs and workflow activity, bots can surface anomalies, explain why errors happen, recommend fixes, and even take corrective action.
- Customer-facing conversational agents tied to back-end automation: LLM-powered assistants can talk to users, gather information, and instantly trigger automated processes behind the scenes, closing the gap between conversation and action.
Meta Llama Incubator Demo Day: Diverse Solutions, Real-World Use Cases
When Meta held its inaugural Llama Incubator Demo Day on 15 October 2025, the event delivered much more than just pitch presentations — it crystallized a powerful narrative about how open-source LLMs can fuel real, operational impact across sectors like finance, healthcare, education, and public services.
Over the course of the programme, the 40 teams developed more than 30 Llama-powered projects. At Demo Day, these teams showcased their work in an exhibition-style format, highlighting how Llama models are not just for chatbots — but can be deeply embedded into business processes, predictive systems, personalization tools, and anomaly-detection flows.
A Showcase of Practical, Sector-Wide Innovation at the Llama Incubator Demo Day
The Llama Incubator Demo Day brought together one of the most diverse cross-sections of AI innovation seen in Singapore in recent years. Startups, SMEs, and public-sector teams filled the exhibition floor with prototypes and applications built on Llama — each demonstrating how open-source generative AI can solve real operational challenges rather than simply power chat interfaces. It was a snapshot of how quickly LLM technology is moving from experimentation to meaningful, industry-shaping deployment.
Instead of futuristic concepts, teams showcased systems designed to automate daily workflows, support frontline operations, enhance internal decision-making, and reduce manual effort. From productivity tools to city operations, the projects reflected a shared belief that generative AI’s biggest impact will come from enabling organisations to work smarter and faster. Many prototypes blended text understanding, multi-step reasoning, data retrieval, and workflow execution — a sign of the rising shift from conversational AI to action-oriented AI.
Another defining feature of the Demo Day was the wide representation of sectors. Exhibitors spanned business services, infrastructure, public administration, operations, education, and more — demonstrating how Llama can flexibly adapt to different industry requirements. These teams showed how generative AI can interpret documents, summarise complex information, guide users through decisions, or augment routine processes. The diversity of use cases reinforced Llama’s role as a foundational technology capable of supporting both high-stakes environments and everyday workplace efficiency.
How Gleematic’s Productivity Showcase Fit into a Broader Exhibition of Llama-Powered Innovation
At the Llama Incubator Demo Day Exhibition on 15 October 2025, Gleematic stood alongside some of the most forward-leaning AI teams in Singapore — each demonstrating how open-source large language models can solve real operational challenges.
We took place under “AI is Optimizing Work” category, with our Gleematic AI Agents – supercharged with Llama – to deliver smarter, more context-aware digital workers.
Visitors to the booth were introduced to our AI Agents, designed to handle multi-step processes with reasoning, decision-making, and natural-language understanding. These agents could read documents, ask clarifying questions, summarize issues, and act across systems — showing how Llama enhances Gleematic’s existing automation engine with adaptability and contextual judgment.
We also showcased Gia, our AI Virtual Avatar Assistant, a human-like interface that brings automation to life. The avatar conversed naturally with visitors, gathering information, answering questions, and triggering backend workflows through Gleematic’s orchestration engine. This demonstration highlighted the future of human–machine collaboration: AI that is not just functional, but intuitive and approachable for everyday business users.
Our centerpiece was the integration of Llama into Gleematic’s signature capabilities Instead of relying solely on rules, Gleematic’s Llama-enhanced workflows could understand unstructured text, interpret ambiguous scenarios, and guide users through next steps. For many attendees, this bridged a longstanding gap between conversational AI and real operational automation.
Although the Demo Day featured many sectors, Gleematic’s showcase had a unique focus: practical productivity gains. While other exhibitors explored areas such as public services, sustainability, or cybersecurity, Gleematic zeroed in on a problem that cuts across all industries — the need to reduce manual work, streamline processes, and empower teams with AI that actually performs tasks, not just explains them.
By demonstrating both AI agents and the virtual avatar assistant, Gleematic illustrated what enterprise automation could look like in the Llama era: intelligent, conversational, autonomous, and deeply integrated with real business workflows.
The response from attendees reflected this shift. Many recognized that Gleematic was not merely adding generative AI as a feature — it was redefining how automation itself works. The showcase made clear that when Llama’s reasoning and language strengths combine with Gleematic’s orchestration and action capabilities, organizations can move from simple task automation to true digital co-workers that boost productivity end-to-end.
Institutional Endorsement & Strategic Alignment
The strong presence of government and industry partners at Demo Day underscored institutional buy-in. From IMDAto GovTech, from Enterprise SG to SGInnovate, multiple agencies backed these efforts — signaling that LLM-based innovation is now central to Singapore’s AI strategy.
In her opening address, Minister Josephine Teo reflected on how AI has moved from “the fringe to the mainstream” in Singapore, and praised the incubator’s role in promoting not just innovation, but trustworthy AI. She emphasized that these solutions are not shallow experiments, but important building blocks for real-world applications: “planning, prediction, personalisation, automation, and anomaly detection” — exactly the kinds of use cases incubator teams tackled.
Beyond Prizes: Validating a Bigger Thesis
More than the money, Demo Day validated a core thesis of the Llama Incubator: open-source LLMs can drive operational innovation, not just consumer chat features. By applying Llama models to document extraction, predictive analytics, agentic workflows, and public-sector risk tools, the participants showed that large language models are capable of tackling real business pain points.
That is particularly meaningful for Singapore’s long-term AI ambitions. Meta’s collaboration with government institutions sends a signal that generative AI is not just a Silicon Valley toy — but a serious component of a national innovation strategy.
Looking Forward: What These Outcomes Imply
- For Startups & SMEs: Smaller firms can build generative AI tools with direct, practical value — especially when combined with domain expertise (sustainability, compliance, sales).
- For the Public Sector: Government agencies are already using LLMs not just for aesthetics or novelty, but to address regulatory and operational challenges (e.g., risk reporting, document analysis).
- For Singapore’s AI Ecosystem: The event reinforces Singapore’s role as a regional hub for responsible, open-source AI. The incubator’s structure, which emphasizes safety and scalability, reflects a long-term belief in AI as a foundational technology, not just hype.
- For Meta: By positioning itself as an “ecosystem partner” rather than just a model provider, Meta is helping to build a sustainable pipeline of Llama-powered innovation — reinforcing open-source as a business strategy, not just a research trend.
A Meaningful Milestone: Gleematic’s Appreciation and Vision Beyond the Llama Incubator
The Demo Day made one thing unmistakably clear: generative AI is no longer confined to labs or tech experiments — it is becoming a practical engine of productivity, governance, and operational transformation across Singapore. Startups and SMEs are proving that domain expertise paired with open-source models can produce high-value solutions; public agencies are already applying LLMs to real regulatory and operational challenges; and the ecosystem as a whole is strengthening Singapore’s position as a regional leader in responsible, scalable AI adoption. Meta’s role as an ecosystem partner — not just a model provider — has laid the groundwork for a sustainable wave of Llama-powered innovation.
Against this backdrop, we are deeply grateful to have been selected for the Meta Llama Incubator 2025. The experience has given us not only technical and product guidance, but also the opportunity to contribute to a vibrant community that believes in building AI that is practical, safe, and impactful. As we move forward, we remain committed to creating solutions that elevate productivity and bring real, tangible benefits to organisations across the region — and we look forward to continuing this journey with the broader Llama ecosystem.
Media references:
Meta accelerates AI Innovation in Singapore with Llama Incubator Program Demo Day
The first Llama Incubator programme launched for startups and SMEs in Singapore
