A practical guide to querying verified market intelligence across Identity, Fraud, and Cybersecurity directly in your AI tools.
The MCP is live for Enterprise accounts. This article is a simplified guide on how to use the Liminal MCP, covering what strategy, sales, and marketing teams can query, how to ask effectively, and where to go for more.
What the MCP gives you
One authenticated endpoint to Liminal’s verified market intelligence. Any MCP-compatible AI, including Claude, ChatGPT, Gemini, or a custom agent, connects once and queries the platform directly.
What the Liminal MCP reads is what makes it different. If you ask a general-purpose assistant about the fraud-prevention market, it will assemble an answer from blog posts and other assets available on the open web, with no way to check which parts are still true. The Liminal MCP reads from a proprietary market taxonomy and knowledge graph that are structured, verified, and maintained by domain experts. Your agent works from an expert-verified market structure and Liminal’s published research, instead of inference from the open web.
The intelligence sits in three connected layers:
- The Liminal Ontology: a structured map of the Identity, Fraud, and Cybersecurity market: vendors, buyers, capabilities, regulations, and threat vectors. All linked and updated when market events fire.
- The Liminal Market Taxonomy: the scaffolding that organizes the graph. Buyer problems and requirements on one side, solutions and product capabilities on the other, connected through use cases.
- The Expert Desk: domain experts who maintain both layers: primary research, monitor-triggered updates, and quarterly review cycles.
Every result includes source attribution when available, and your organization’s context on the platform (Team Spaces content and uploaded documents) is queryable alongside this intelligence in the same call, where your account permissions and connection type support it. Connections among vendors, capabilities, and buyers are only surfaced upon verification.
What you can ask
The Liminal MCP guide covers all available query types. The asking process is simple, and no specific commands are required. Query in your own words, and your assistant works out how to answer.
Understand a market
Name a market, a buyer problem, or a product category. The MCP returns what connects to it: buyer problems, demand signals, products, and vendors.
Query examples:
– “What problems does Liminal track in fraud, and which ones are relevant to us?”
– “What do banks actually require from an account takeover solution?”
– “How is that different for a gaming platform?”
Size up a company
Search the companies covered by Liminal and ask the assistant to cite the records it used. Compare several side by side. See which companies look most like a target, and which is moving fastest.
Query examples:
– “Which companies compete with us most directly, and how are they different from each other?”
– “How does our product compare with [competitor], and where are they moving fastest?”
Find out who sells what
Search thousands of products by name, keyword, or what they actually do. Ask which products address a specific capability area, and get the vendors behind them.
Query examples:
– “Which products do [the thing we do], and how does each one differ from ours?”
– “Who sells into [problem], and what do they lead with?”
Understand what buyers want
See the buying side as structure: who the buyers are, what they are trying to achieve, what frustrates them in the market, and the specific criteria they use to evaluate vendors.
Query examples:
– “What frustrates buyers about what is available in [category], and who runs the evaluation?”
– “What actually decides a [problem] purchase, and how should we be talking about it?”
Track what is changing
Search acquisitions, launches, partnerships, leadership changes, and other dated events. See who is buying whom and what has moved recently.
Query examples:
– “What changed for our top five competitors this quarter?”
– “Who is acquiring in [market], and who sits closest to the deal activity?”
Follow the regulations
Connect rules, standards, and legislation to the problems and requirements they affect, by country and region.
Query examples:
– “Which new rules in [country] are creating requirements we could get ahead of?”
– “What regulations apply to [problem] in [market]?
Use Liminal’s research
Search Liminal’s long-form research by topic or by full text for a specific term. Survey findings sit inside these reports. Ask the assistant to find the report and pull the figure and its context.
Query examples:
– “What does Liminal’s research say about [topic], and what are the strongest numbers in it?”
– “Search for [term] mentions across the research.”
– ”What survey data shows about why buyers in [market] chose their current vendor, and what frustrated them most about alternatives?”
Source what you write
Ask which Liminal reports and survey findings support a claim, and get the report titles and figures back.
Query examples:
– “Which Liminal reports support the claim that banks are consolidating fraud vendors? Give titles and figures.”
– “List the sources behind that answer.”
Access and query your organization’s content
Search and read documents your team uploaded to Team Spaces inside the platform, and query your GTM content (playbooks, battlecards, and account plans) all within what you already have access to.
Available content depends on your account permissions and connection type. Contact your CSM if you are unsure what your connection supports.
Query examples:
– “What do our own notes say about [topic]?”
– “Pull together everything we have internally on [account] before the call.”
– “How do we win against [competitor] in [market], and what does the battlecard say?”
How the MCP answers a complex query
The query below shows how the MCP chains three reads to answer one vendor question end-to-end: taxonomy, product graph, and demand signals. Every part of the answer was read live from the production server. When the assistant draws conclusions about vendor positioning or buyer preferences, it reflects its analysis of the records it has sourced.
Practical Example
Starting question: “Which banking use cases call for behavioral biometrics, what will buyers evaluate it on, and which vendors already ship it?”
Where it plays: Behavioral Biometrics sits in the Biometrics segment and carries 8 typed capabilities: Keystroke Analysis, Mouse and Pointer Analysis, Touch and Swipe Analysis, Behavioral Drift Analysis, Device Interaction Pattern Analysis, and others. In banking, it is one of 16 solution categories linked to Account Takeover Fraud in Banking, one of more than a dozen industry-specific ATO use cases in the taxonomy.
What buyers evaluate: The Banking ATO node carries 12 typed requirements. Behavioral biometrics competes most directly on the Standard and Guidance tier: detecting and flagging anomalous login and session behavior in real time. Regulatory requirements (SAR generation) and Business requirements (preventing fraudulent transfers from compromised accounts) define the platform integrations buyers expect around it.
The competitive set: The product graph returns products shipping behavioral biometrics as a verified capability, from identity platforms to fraud suites, each linked to the vendor behind it.
What the buyer worries about: The Banking ATO requirements include delivering explainable risk scores and fraud decisions for examiner and audit review. Based on this requirement, the assistant identifies vendors whose behavioral signals are explainable and examiner-ready as better positioned.
FAQS and Setup
Which AI tools work with the Liminal MCP?
Any client that supports the Model Context Protocol, including Claude, ChatGPT, Gemini, Perplexity, and Microsoft Copilot, as well as custom agents your team builds. There is no Liminal-specific SDK to adopt.
What does the MCP not cover?
The MCP does not support bulk export of the dataset. It is designed for querying and not for data extraction. It does not return meeting notes, internal notes, or release notes. It also does not support account administration or platform configuration. For anything the MCP cannot answer, the platform remains the right place to go.
Does the MCP replace the Liminal platform?
No. The Liminal platform includes a visual market landscape: a structured map of the Identity, Fraud, and Cybersecurity markets that lets you navigate segments, drill into use cases, and explore competitive coverage visually. The MCP does not replicate that experience. Your agent can query what sits inside the landscape (segments, buyers, vendors, capabilities), but it cannot render or navigate the map itself. For that, you need the platform directly.
Does the Liminal MCP always return outputs with the most current data?
The production server is more current than any document, and the figures in this guide reflect August 2026. Ask your assistant how large a branch of the taxonomy is today and it reads the live catalog. Some derived views refresh on a schedule, so when a figure matters, ask for its source and date.
Does querying the Liminal MCP consume credits?
No. Reading from the MCP is included in Enterprise access. Agent-to-agent queries and jobs are on the roadmap and may draw consumption credits when they ship. Reading stays included in the plan.
How do I connect the Liminal MCP?
Connection requires a 30-minute setup call with your Customer Success Manager. Once access is enabled, your CSM will provide you with the connection instructions and walk you through connecting your MCP-compatible AI to the Liminal endpoint.
I cannot add the Liminal MCP connector to my AI tool. What should I do?
This is usually an IT permissions issue, not a problem with your Liminal access. Enterprise AI tools often require administrator approval before new connectors can be added. Contact your IT team to request permission to add the Liminal MCP connector, then return to the connection instructions your CSM provided.
Access and availability
Enterprise customers: The Liminal MCP is included in your subscription. Contact your Customer Success Manager to enable access and schedule a setup call.
Standard and Professional customers: The MCP is available on the Enterprise tier. Contact your Account Manager to explore upgrading.
Prospects: Book a working session to see the MCP running against live market data in your category.
Resources
Liminal MCP Guide — Full specification: what the MCP exposes, the intelligence layers behind it, connection instructions for engineers, and the data governance model.
Product update: Liminal MCP — Introduction to the MCP and the problem it solves.
Liminal MCP Prompting Best Practices — Eight habits and role-based prompt examples to get the most out of every query, from strategy and product to sales and marketing.
August 13, 2026 Release Notes — Full changelog for this release.
