Aurascape

• A secure-web or data-loss platform could add inline AI discovery and prompt inspection to controls security teams already buy, absorbing the employee-use half as a bundled feature.
• Every customer figure in the public record is Aurascape's own, so disclosed adoption could prove smaller or shallower than the case studies imply.
• The reviewed sources record no SOC 2 or ISO 27001 report for Aurascape, and regulated buyers can require one before signing.
• The gateway half of the agent controls enforces policy on tool calls routed through it, so enterprises moving away from the Model Context Protocol would blunt that half.
• The application catalog needs daily engineering to stay current, so a slowdown in that work would erode the coverage the product is sold on.
• Splitting effort across employee AI use and the agent build side could leave each half matched by a more focused competitor.

Aurascape targets the gap between how fast enterprises adopt AI and how little their security teams can see of it. Employees paste sensitive material into chatbots and copilots nobody sanctioned, and developers wire agents into internal systems through tools and APIs that older controls were never built to read. The company sells to the security and compliance leaders who carry that exposure.

SecurityWeek echoes the problem framing in its own voice, describing unsanctioned third-party AI applications that slip past traditional security controls and a sizable enterprise market for tools that mitigate them. SiliconANGLE adds quantification, attributing to McKinsey a finding that employees use generative AI three times more than leaders expect and to Deloitte a finding that one in three employees pay for an AI application themselves. Both figures also sit in Aurascape's own launch release.

Aurascape scopes the problem in two halves on one platform. One half is employee AI use across apps, browsers, and copilots. The other is the agents and applications a company's own teams build and run, which its March 2026 platform expansion moved to the center of the pitch. Covering both is the company's positioning bet, and it asks a governance owner and an engineering owner to buy the same product. [s5, s4, s15, s1]

The platform inspects AI activity while it happens rather than after the fact. Menlo Ventures describes a cloud-native inline proxy paired with an endpoint agent that decodes more than 1,500 applications and captures activity across text, audio, image, and code, and Khan told BankInfoSecurity the endpoint client redirects only AI-related traffic to the cloud platform. Sitting in that path is what lets the product classify sensitive data inside prompts and responses and stop risky activity in real time.

Aurascape's own engineering post puts the current coverage at more than 20,000 catalogued AI applications as of June 2026, each carrying a decoder implementation and a risk characterization, with new applications added daily and a 48-hour target the company scopes to the vast majority of AI applications. The catalog is the working asset behind the discovery and policy claims, and keeping it current is continuous engineering rather than a shipped feature.

The build side extends the same position to agents. The product maps MCP servers, tool connections, and data flows, stress-tests agents against simulated prompt injection and jailbreak attempts before release, and enforces policy on the model conversation and on gateway-routed tool calls.

Depth otherwise rests on the company's own account. The one third-party technical write-up in the record, an NSFOCUS analysis of the vendor's pages, concludes that real-world validation is still needed. A buyer would have to confirm the decode and enforcement claims hands-on. [s7, s1, s12, s23, s13]

Aurascape competes on covering both AI use and AI building on one platform. Its product pages document discovery and policy across employee AI activity, and a separate agent side that maps MCP servers, tests agents before release, and governs tool calls at runtime. The one survey of competing coverage in the reviewed sources is Aurascape's own market page, which sorts the market into four vendor groups and places Aurascape across all of them, so the breadth claim rests on the vendor's account of its rivals rather than on evidence about them.

The sharper pressure comes from the platforms whose budget line Aurascape shares. By its own market page, Zscaler delivers AI security through its cloud proxy, discovers AI applications using its existing shadow-IT framework, applies inline threat detection and data-loss controls, and announced an MCP gateway in early 2026. The same page positions Aurascape as additive to an existing secure-web deployment rather than a replacement, which lowers the bar to adoption and leaves the network position with the vendors already in that path.

Aurascape's counterweight is the application catalog and the agent controls, which take more work to reproduce than a policy toggle. Whether that depth embeds the product in customer workflows deeply enough to resist bundling is not answerable from the reviewed record, which carries no independent account of how any customer runs it. [s1, s17, s12, s5]

Aurascape's disclosed traction grew substantially over the past year, and the company published all of it. The homepage carries four testimonials whose speakers are identified by title alone, plus one quote attributed to Vineet Arora, CTO of WinWire Technologies. The homepage publishes image alt text naming eighteen organizations, among them two credit unions, a transit agency, and a university, and one of those names is Auradine, the portfolio company Mayfield describes Aurascape as spinning out of, so at least one published customer name is the company Aurascape came out of.

Aurascape publishes four case studies. One names The Police Credit Union and records a two-phase deployment aligned to NCUA guidance and the NIST AI Risk Management Framework, with a projected 27 percent productivity gain and a projected 83 percent reduction in AI-based risk. Three describe unnamed Fortune 500 enterprises, at more than 30,000 users at an insurer, more than 15,000 developers at a financial services firm, and more than 60,000 users at a healthcare technology company. Aurascape publishes every one of those figures, and the credit-union percentages are that customer's own projections.

The independent record covers the money and stops. SecurityWeek, SiliconANGLE, and BankInfoSecurity all reported the April 2025 launch and the 50 million dollars behind it, RSA Conference's finalist announcement records a 5 million dollar investment awarded to each of the ten 2025 finalists, and BankInfoSecurity recorded a near-term goal of 25 enterprise customers in finance, healthcare, and retail plus a recurring-revenue target of 10 to 15 million dollars within two years. No reviewed source outside Aurascape reports a customer, a deployment, or a revenue figure since. In August 2026 the company named Jim Walsh chief revenue officer, an appointment it frames as following a year of growth. [s3, s10, s16, s19, s20, s22, s6, s14]

Aurascape names its founders publicly. The company page lists Moinul Khan as CEO, Patrick Xu as CTO, Viswesh in product management, and Liang Li in engineering under leadership, adds a marketing lead and a sales lead beside them, and separately lists Rajiv Khemani as a co-founder and board member. Aurascape announced Jim Walsh as chief revenue officer on August 3, 2026.

The senior experience the independent record documents belongs to the chief executive. BankInfoSecurity reports that Khan spent nearly six years at Zscaler, ending with three years leading its SSE and data protection business, after four and a half years in product management at Palo Alto Networks and 18 months at Netskope. It names no product he built. Mayfield credits the six people it names as the team it partnered with, saying they built products that generated over 10 billion dollars in revenue, and it names none of those products.

The backers are the louder credibility signal, and they are backers rather than builders here. Former Palo Alto Networks CEO Mark McLaughlin sits as a board observer, and the launch release names him alongside former Symantec CEO Greg Clark and former Zscaler strategy chief Manoj Apte among the investors. The company page also lists CISOs at AT&T and Mastercard and a Databricks CIO as advisors, which is advisory standing rather than customer proof. [s2, s6, s22, s14, s15]

Aurascape publishes no attestation a security buyer can read. A probe of its public surfaces found the trust and security subdomains failing DNS against a nonsense control subdomain that failed the same way, the /trust and /pricing paths returning 404, the /security path serving an image file, and no compliance badge filename in the homepage markup. The reviewed sources record no SOC 2 or ISO 27001 report for the company.

That gap carries weight for this particular product. The platform reads every prompt and response an employee sends, and the buyers its case studies name are examined by the National Credit Union Administration or bound by state insurance data security laws. Those case studies are built around assembling examiner-ready evidence with the product, while the reviewed record documents nothing about Aurascape's own controls.

The company's visible security credibility is its research instead. Its AuraLabs team published a coordinated disclosure of a zero-click prompt-injection takeover in Meta's Manus agent, reported to that vendor in September 2025 and published in February 2026. That shows engineering depth and does not speak to how Aurascape handles customer traffic. [s11, s10, s16, s21]

Add analyzed competitors to compare them side by side with Aurascape.

Aurascape sells to enterprise security and compliance leaders carrying two exposures at once, employees sending sensitive material into AI apps nobody sanctioned and internal teams wiring AI agents into business systems. At launch the company told BankInfoSecurity its near-term goal was 25 enterprise customers across finance, healthcare, and retail, on multi-year subscriptions. The buyers it has since published match that aim.

Those buyers are regulated and data-sensitive. A case study documents The Police Credit Union governing AI use against NCUA guidance and the NIST AI Risk Management Framework, and three more describe Fortune 500 enterprises in insurance, financial services, and healthcare technology. The homepage publishes image alt text naming eighteen organizations, among them two credit unions, a transit agency, and a university.

The segmentation bet is scope. Aurascape addresses both the employees who use AI and the teams who build it, which widens the addressable budget and asks a governance owner and an engineering owner to buy one platform.

The core capability is decoding AI traffic while it moves. Menlo Ventures describes a cloud-native inline proxy paired with an endpoint agent that decodes more than 1,500 applications and captures activity across text, audio, image, and code, and Khan told BankInfoSecurity that the endpoint client redirects only AI-related traffic to the cloud platform for real-time inspection and enforcement.

Aurascape's own engineering post puts current coverage at more than 20,000 catalogued AI applications as of June 2026, each carrying a decoder implementation and a risk characterization, extended daily by automated discovery bots that crawl the web and pull API feeds from app catalog sites, under a stated 48-hour target that the company scopes to the vast majority of AI applications and not to all of them. That catalog is the working asset behind every discovery and policy claim, and keeping it current is a permanent engineering cost rather than a shipped feature.

The agent side extends the same position to what agents do. The platform maps MCP servers, tool connections, and data flows, stress-tests agents with simulated prompt injection and jailbreak attempts before release, and enforces policy on the model conversation and on gateway-routed tool calls. The company also runs its own offensive research, with its AuraLabs team disclosing a zero-click prompt-injection takeover in Meta's Manus agent through a coordinated window that closed in February 2026.

Two limits sit on the agent claim. Aurascape names the control point the Zero-Bypass MCP Gateway, while its own March 2026 announcement describes that gateway combined with its AI proxy as identifying risky MCP activity visible to the platform and reducing bypass risk across agent interactions, which is a narrower promise than the name carries. Depth otherwise rests on the company's own account, since the one third-party technical write-up in the record, an NSFOCUS analysis of the vendor's pages, concludes that real-world validation is still needed.

The motion is direct enterprise sales funded ahead of any disclosed revenue. No free tier or self-serve sign-up appears on the reviewed pages, and Khan told BankInfoSecurity the near-term plan was 25 enterprise customers in finance, healthcare, and retail on multi-year subscriptions, and a recurring-revenue target of 10 to 15 million dollars within two years.

The published customer record is now substantial and it is entirely Aurascape's. Four case studies describe deployments, one naming The Police Credit Union and three describing unnamed Fortune 500 enterprises at more than 30,000 users, more than 15,000 developers, and more than 60,000 users. The homepage publishes image alt text naming eighteen organizations, one of which is Auradine, the portfolio company Mayfield describes Aurascape as spinning out of, so one published customer name is the company Aurascape came out of. The homepage testimonials pair speakers with titles rather than employers apart from the WinWire Technologies quote.

Awareness leans on backing and recognition rather than reported adoption. RSA Conference named Aurascape one of ten Innovation Sandbox finalists in April 2025, its analyst-coverage page reports a Gartner listing as a sample vendor in AI usage control, and the launch drew coverage in SecurityWeek, SiliconANGLE, and BankInfoSecurity. None of those sources reports a customer or a deployment, and the chief revenue officer appointed in August 2026 is the company's own account of a year of growth.

Aurascape publishes no pricing. The /pricing path returns 404, and the reviewed sources do not state the unit the company charges by, whether per user, per traffic volume, or per agent.

The one pricing signal in the record is structural. Launch coverage records that the company set retention targets for multi-year subscriptions, so revenue is recurring. Nothing public indicates how Aurascape prices agent-side usage, where consumption varies with agent activity rather than headcount.

Delivery is a cloud service with an endpoint component. Aurascape pairs a cloud-native inline proxy with an endpoint client that redirects only AI-related traffic to its cloud platform, so a deployment touches both network routing and device fleets. A financial services case study records the product deployed alongside the firm's existing secure-service-edge stack, carrying AI traffic only.

The heavier operating load sits with the vendor. Decoding more than 20,000 AI applications works only while the decoders track those applications as they change, which Aurascape handles with automated discovery bots and a daily addition of new applications with their decoders and risk characterizations. For the customer, the visible work is routing AI traffic through the proxy, rolling out the endpoint client, and tuning policy to roles and data.

Aurascape's public trust collateral lags what it sells. A probe of its public surfaces found the trust and security subdomains failing DNS against a nonsense control subdomain that failed the same way, the /trust and /pricing paths returning 404, the /security path serving an image file, and no compliance badge filename in the homepage markup. The reviewed sources record no SOC 2 or ISO 27001 report for the company.

The company's visible security credibility is its research instead. Its AuraLabs team published a coordinated disclosure of a zero-click agent takeover in Meta's Manus agent, reported to that vendor in September 2025 and published in February 2026, which shows engineering depth and does not speak to how Aurascape handles customer traffic.

On the buyer-facing side, trust is part of the pitch. The Police Credit Union case study is built around examiner-ready evidence and controls mapped to NCUA guidance and the NIST AI Risk Management Framework, so the product sells compliance readiness to its customers while the reviewed record documents none of the company's own certifications.

Aurascape positions itself beside the incumbents rather than against them. Its own market page states the product runs alongside an existing secure-service-edge platform such as Zscaler rather than replacing it, and describes itself as additive to existing secure-web, cloud-access, and data-loss tools with no rip-and-replace. That lowers adoption friction and leaves the network position with vendors already in the path.

The platform hooks it is building are agent-era control points. The Zero-Bypass MCP Gateway sits on the tool-call channel and an AI proxy on the model channel, and if agent traffic settles on the Model Context Protocol, holding a policy-enforcing checkpoint there is a real position. The same market page records the threat to it, noting that Zscaler discovers AI applications through its existing shadow-IT framework and announced an MCP gateway in early 2026.

The founding team is public and senior. The company page names Moinul Khan as CEO, Patrick Xu as CTO, Viswesh in product management, and Liang Li in engineering under leadership, lists Rajiv Khemani separately as a co-founder and board member, and Mayfield's seed announcement names six people it partnered with, including Rajiv Khemani, and describes Aurascape as a spin-out from its portfolio company Auradine.

The senior experience the independent record documents belongs to the chief executive. BankInfoSecurity reports that Khan spent nearly six years at Zscaler, ending with three years leading its SSE and data protection business, after four and a half years in product management at Palo Alto Networks and 18 months at Netskope. It names no product he built. Mayfield credits the six people it names as the team it partnered with, saying they built products that generated over 10 billion dollars in revenue, and it names none of those products.

The company page lists a marketing lead and a sales lead beside the founders, and Aurascape announced Jim Walsh as chief revenue officer on August 3, 2026. Former Palo Alto Networks CEO Mark McLaughlin sits as a board observer, and the company page lists CISOs at AT&T and Mastercard and a Databricks CIO as advisors, which is advisory standing rather than operating leadership or customer proof.