LangProtect at GISEC Global 2026: Building a Secure Future Run by AI and Agents
DUBAI, UAE, 29 September 2026 - LangProtect, the AI security platform built by Quokka Labs, exhibited at GISEC Global 2026, the largest cybersecurity event in the Middle East and Africa, held from 16 to 18 September at Dubai Exhibition Centre, Expo City Dubai. Over three days at Booth H1-SP58, the LangProtect team met security leaders, enterprises, managed security service providers (MSSPs) and technology partners from across the region.
The team came to Dubai with one message: AI risk now sits in three places, and each one needs its own controls. The first is the workforce, where employees use AI tools such as ChatGPT, Claude, Gemini and Copilot every day, often without security teams knowing. The second is the AI applications organisations build and ship, from customer-facing assistants to internal copilots and RAG-based systems. The third, and fastest-moving, is AI agents: software that can call tools, reach APIs and act on data without a human approving each step.
At the booth, visitors explored LangProtect's three products, each built for one of those layers: Guardia for employee AI governance, Armor for AI application security, and Vector for agent, MCP and non-human identity security. Across the three days, the conversations kept returning to one question: how do we build a secure future run by AI and agents? That question is the reason LangProtect exists.
GISEC Global 2026 at a Glance
GISEC Global is where governments, national cyber authorities, enterprise CISOs, technology providers and security practitioners from across the Middle East, Africa and beyond meet each year to set the regional cybersecurity agenda. The 2026 edition ran under the theme "Cyber First: The New Digital Order," and its scale reflected the region's growing weight in global cybersecurity.
AI security was one of the event's defining themes, alongside critical infrastructure protection, post-quantum readiness and cyber resilience. That focus reflects a wider shift across the industry. The World Economic Forum's Global Cybersecurity Outlook 2026 found that 94% of organisations expect AI to be the biggest driver of cybersecurity change in the coming year, and that 87% identify AI-related vulnerabilities as the fastest-growing cyber risk.
For LangProtect, a company focused entirely on securing AI adoption, GISEC Global offered something rare: the region's security decision-makers, channel partners and technology leaders in one room, at the same time, working on the same problem.
A Message from Our Co-Founder, Live from the Show Floor
On the opening day of GISEC Global 2026, LangProtect Co-Founder Suny Choudhary stepped away from the booth for a few minutes to share a short message from the show floor. Posted on LangProtect's LinkedIn page while the event was running, the video captures his first impressions of GISEC and what the team came to Dubai to showcase.
For Suny, GISEC was a chance to take LangProtect's view of AI security directly to the people responsible for it: CISOs, security architects, government cyber teams and the channel partners who serve them.
What Happens When AI Makes the Wrong Move?
That question was the title of the short film that played at Booth H1-SP58 throughout the three days. It was also the question behind almost every conversation the team had.
AI no longer sits quietly in a chat window. It reads, writes, decides and increasingly acts, and that changes what a mistake looks like.
Where a Wrong Move Can Happen
A wrong move looks different depending on where AI sits in the organisation. The examples below are illustrative. They show the kinds of risk LangProtect's three products are designed to address.
Each of these starts as a small, ordinary action. By the time it reaches the user, the model or the downstream system, it can already be a data exposure or an unauthorised action. Here is how LangProtect addresses each layer.
One Platform. Three Layers of AI Risk.
LangProtect brought a single platform to GISEC, built around a simple idea: AI risk does not live in one place. Employees use AI tools. Teams build AI applications. Agents act on systems and data. Each layer creates different exposures, needs different controls and is usually owned by a different team. Here is what visitors explored at Booth H1-SP58.
Guardia: Employee AI Governance
Secure AI use across your workforce.
Most organisations cannot say with confidence which AI tools their employees use, or what data goes into them. LangProtect Guardia gives security teams visibility into employee AI use, protects sensitive data and applies policy control across AI interactions. Employees can keep using AI productively while the organisation stays in control of its data.
- Discover Shadow AI. See which AI tools employees are actually using, including tools security teams have never reviewed or approved.
- Protect sensitive data. Detect personally identifiable information (PII), protected health information (PHI), payment card data (PCI), credentials, source code and other sensitive content in prompts before it leaves the organisation.
- Govern AI by intent. Move beyond keyword lists and regular expressions to context-aware, intent-based policies that reflect what a user is actually trying to do.
- See AI risk across teams. Track users, tools, risk levels and usage patterns from one place.
- Stay audit-ready. Maintain real-time alerts, activity logs and evidence of AI policy enforcement.
Learn more about shadow AI detection.
Armor: AI Application Security
Protect the AI applications you build and ship.
When an organisation puts a large language model (LLM) behind a customer chatbot, an internal copilot or a retrieval-augmented generation (RAG) system, the application becomes a new attack surface. LangProtect Armor secures what goes into these applications and what comes out of them.
- Stop AI threats. Detect prompt injection, jailbreaks and malicious inputs aimed at your AI applications.
- Protect sensitive data. Prevent PII, secrets, confidential and proprietary data from being exposed or propagated through AI responses.
- Protect context and RAG. Secure the data and context your AI uses to reason and respond.
- Control AI outputs. Detect and restrict unsafe, non-compliant or unintended responses before they reach users or downstream systems.
- Enforce policies at runtime. Allow, restrict or block interactions based on your security policies.
- See and audit every decision. Keep a clear record of AI security decisions for investigation and compliance.
- Test before attackers do. Use AI red teaming to surface threats before they become an attack surface, and Evals to assess how your AI agents perform with customers and employees.
Learn more about LLM application security.
Vector: Agent, MCP & NHI Security
Secure what your AI agents can see, access and do.
AI agents do more than generate text. They call tools, query databases, reach APIs and connect to Model Context Protocol (MCP) servers, often with credentials and permissions no one has reviewed. LangProtect Vector brings discovery, permission control and auditability to agents and the non-human identities (NHIs) they run on.
- Discover. Find AI agents, MCP servers and non-human identities across your environment.
- Control. Define granular permissions for what agents can discover, read, write and execute.
- Enforce. Prevent over-privileged agents from accessing tools, APIs or data beyond their intended purpose.
- Secure every tool call. Inspect agent actions and evaluate intent before execution.
- Stop rogue actions. Allow, block or redact unauthorised or risky agent actions.
- Protect against agent attacks. Detect direct and indirect prompt injection, malicious tool use and sensitive data exfiltration.
- Secure MCP connections. Monitor and control the agent-to-MCP layer, including Shadow MCP servers.
- Audit every agent decision. Get a human-readable trail of what an agent tried to do, why it was allowed or blocked, and which policy applied.
Going Further with LangProtect
Beyond the three core products, visitors also explored capabilities that extend the platform.
Three Days at GISEC Global 2026
The LangProtect team documented the event day by day on LinkedIn. Here is how the three days unfolded.
Day 1, 16 September: The Booth Goes Live
GISEC Global 2026 opened, and the LangProtect team was ready at Booth H1-SP58 in the Dubai Exhibition Centre. The first day was about introductions: meeting the cybersecurity community, taking visitors through the platform and hearing how organisations across the region are approaching AI adoption. Co-Founder Suny Choudhary also recorded his message from the show floor that day.
Day 2, 17 September: Conversations with Enterprises and Partners
Day 2 was the busiest day of the event. It brought more conversations, sharper questions and strong interest in what LangProtect is building for AI security. The team met representatives from global consultancies, enterprises, managed security service providers (MSSPs) and technology companies from across the region. Many visitors arrived with a specific problem to solve, whether that was controlling employee use of AI tools, securing an AI application already in production, or understanding what their agents can access.
Day 3, 18 September: Closing Strong
The final day kept the momentum going, with more conversations, new connections and follow-up meetings lined up. When GISEC Global 2026 closed, the team left Dubai with a list of demos, technical discussions and partner conversations to continue. The post below includes glimpses from Days 2 and 3.
What Security Leaders Asked Us About
The product demos drew people in, but the conversations were where the event proved most valuable. Over three days, security leaders, enterprise teams and channel partners came to Booth H1-SP58 with real questions about how AI is being adopted inside their organisations and where the risk is showing up. Three themes came up again and again.
AI Red Teaming Is Becoming a Priority
Several conversations, particularly with distribution partners, pointed to AI red teaming as a clear and growing need. As organisations move AI applications and agents from pilots into production, they want to know how those systems behave under attack before launch, not after an incident. The questions were practical: can our assistant be manipulated through prompt injection, will it leak data it has access to, and what happens when someone deliberately tries to break it? Adversarial testing gives teams evidence instead of assumptions. Mapped against frameworks such as the OWASP Top 10 for LLM Applications, it also produces findings that can become enforceable controls once the application is live.
Agents Are the Next Security Frontier
As AI moves from answering questions to taking actions, visitors wanted to understand what that means for control and accountability. They asked what agents can access, which tools and MCP servers they connect to, whether their permissions match their purpose, and how every agent decision can be traced afterwards. Many organisations are already experimenting with agents, often faster than their security teams can review them. The gap between agent adoption and agent governance was one of the clearest signals from the event, and it is why AI agents increase security risk in ways traditional controls were not designed for.
The Channel Is Ready for AI Security
MSSPs, system integrators and distribution partners showed strong interest in bringing AI security to their own customers. Many of these partners already serve enterprises that are adopting AI tools and building AI applications. Those customers are now asking their trusted security providers how to govern AI use. For partners, AI security is quickly becoming part of the conversation they are expected to lead.
"GISEC gave us the chance to have interesting conversations with global leaders on one question: how can we build a secure future run by AI and agents? That is exactly our mission at LangProtect. Thank you to everyone who stopped by our booth and shared their perspective."
— Suny Choudhary, Co-Founder, LangProtect
5 AI Security Lessons from GISEC 2026
Three days on the GISEC floor produce more than leads and business cards. They produce a clearer picture of how organisations are actually adopting AI, where security teams feel exposed, and what they need to move forward with confidence.
After the event, LangProtect Co-Founder Suny Choudhary set out the five lessons he took away from GISEC Global 2026. They draw on conversations at Booth H1-SP58 with CISOs, enterprise security teams, MSSPs and distribution partners from across the region. The piece looks at what security leaders are prioritising and how AI adoption is changing the job of securing an organisation. It also covers why the gap between how fast AI is being deployed and how fast it is being governed matters more than ever.
It is a practitioner's view of the event, written for security and technology leaders deciding how to secure AI in their own organisations.
Thank You, GISEC
Three busy days, countless conversations and many exciting discussions about securing AI. GISEC Global 2026 gave the LangProtect team exactly what it came to Dubai for: direct conversations with the people responsible for making AI adoption safe.
A big thank you to everyone who visited Booth H1-SP58 and took the time to share their questions, challenges and perspectives with us. Thank you to the GISEC Global organisers for bringing the region's cybersecurity community together under one roof, and to the LangProtect team on the ground in Dubai: Suny, Sanchit, Pooja, Harsha and Dhawal.
We are heading home with plenty to think about and even more to build on.
What's Next for LangProtect
GISEC Global 2026 may be over, but the conversations that started at Booth H1-SP58 are only beginning. The LangProtect team is now following up with the enterprises, MSSPs and technology partners it met in Dubai through product demos, meetings and deeper technical discussions about their AI security priorities.
If you visited us at GISEC, expect to hear from the team soon. If you missed us in Dubai, the conversation is still open.
Get Your 30-Day Free Pilot
See LangProtect working in your own environment. Run it on your own AI and get reporting on what your employees share with AI tools, what your AI applications expose, and what your AI agents can reach.
About LangProtect
LangProtect is an AI security platform built by Quokka Labs to help organisations secure AI agents, applications and workforce. The platform covers the three layers where AI risk now sits:
- Guardia governs how employees use AI tools, with visibility into Shadow AI, sensitive data protection and intent-based policies.
- Armor protects the AI applications organisations build and ship, from prompt injection and jailbreaks to unsafe outputs and data exposure.
- Vector secures AI agents, MCP connections and non-human identities, controlling what agents can see, access and do.
Together, they give security teams the visibility, control and audit evidence they need to adopt AI with confidence. Learn more at www.langprotect.com.
About GISEC Global
GISEC Global is the Middle East and Africa's largest cybersecurity event. It brings together governments, national cyber authorities, enterprise CISOs, critical infrastructure operators, technology innovators and cybersecurity professionals from around the world. Organised by inD, the 2026 edition was held from 16 to 18 September at Dubai Exhibition Centre, Expo City Dubai, under the theme "Cyber First: The New Digital Order." Learn more at gisec.ae.
LangProtect Team
Email: teams@langprotect.com
Phone: +91 95016 20330
Web: www.langprotect.com