Argus · First Seed Investor Meeting Guide

Novai LTD (HK) is the proposed fundraising entity for this round, and Argus is the project / product. The company’s legal name and the relationship between its operations and IP ownership are [TO FILL: verified relationship]. If there are related entities or any ownership formalities that are not yet complete, we will explain that accurately and provide documents during a controlled review later.

We improve the quality and verifiability of factual inputs: facts are returned together with their sources, timestamps, and quality information. Downstream models can still misinterpret or reason incorrectly, and source disclosures can also be wrong, so we cannot claim to eliminate hallucinations. We can provide [TO FILL: task-level evaluation] to show where we help.

We plan to start with [TO FILL: segment and task of customers already contacted], then convert a small, deliverable pilot into a paid engagement rather than relying first on the long procurement cycles of large banks. Our actual pipeline is [TO FILL: distinguish contacts / pilots / commercial discussions / signed agreements]. If we have no orders, we will say clearly that nothing has been signed.

The scope that is currently usable and authorized is [TO FILL: markets / entities / historical depth / fields]. Update latency is [TO FILL: listed by source]. A tool definition does not mean that a vendor or production data is in place. We disclose gaps and latency by source, and do not describe daily or delayed data as real time.

The current mechanisms retain fact sources and the time when information became available. Historical queries should exclude information that became available only after the query timestamp; conflicts and revisions need to be recorded explicitly. Evidence of reliability requires [TO FILL: sample size, benchmark, human review, error rate, and date]. If we have no production evaluation, we will say directly that this is still being validated.

If the scope, authorization, and time allow, we can validate this in a separate follow-up technical session using a sandbox. Today, we’ll first show the standard request / response format so an unplanned configuration does not take over the introductory meeting. We will not ask you to provide production credentials or share our own production credentials.

Actual feedback is [TO FILL: feedback that can be quoted and its date]; the most common obstacle is [TO FILL]. Based on this feedback, we have made [TO FILL: completed improvement] and have yet to resolve [TO FILL]. If we have no user feedback, we will not invent any; we will explain that validation is being arranged.

The observed sales cycle is [TO FILL: sample and range]; reasons for churn or for stopping a pilot include [TO FILL]. The share of revenue from our largest customers is [TO FILL]. If the sample is too small, we will say the figure is unstable. Next, we plan to reduce dependency through [TO FILL: channel and product measures], without presenting projections as results already achieved.

Costs for a typical task should include data licenses, scraping and synchronization, parsing and model usage, storage, compute, operations, and human review; the actual sample is [TO FILL]. Caching, batch processing, and deterministic processing may reduce costs, subject to licensing and update requirements. If we do not yet have production measurements, we will not report unsupported gross margins and will first prepare a budget using clearly stated assumptions.

Key dependencies are [TO FILL: sources, vendors, models, cloud]. We will describe exclusivity only if there is an actual contract and we are permitted to disclose it. Our substitution and migration plan is [TO FILL]. Source shutdowns, price increases, or changes in authorization could affect the service; we will document the impact on coverage and our response for each dependency.

We are starting with a deliverable factual layer and clearly defined customer tasks. Over time, we may expand to more authorized data domains and enterprise workflows, but each expansion must be validated through payment, quality, and cost. We will not present our vision as an established platform or claim achieved scale effects.

The actual reason for establishing Novai LTD (HK) is [TO FILL: genuine business reason]. The locations of operations and employees, customer contracting entity, and IP ownership are [TO FILL: verified facts]. If your fund has requirements for the investment entity or structure, we are willing to understand them and have our advisors assess them. Today, we will not promise any restructuring or tax outcome that has not been reviewed.

We are currently considering [TO FILL: preference for equity / convertible instrument], with a target for terms of [TO FILL: what can be shared externally]. If we have not decided yet, we will say so. Our need for a lead investor is [TO FILL]. We will describe the status of other investors separately as contacted, evaluating, written terms received, signed, or funded; verbal interest is not a commitment. Applicable Hong Kong documents and corporate approvals must be checked separately.

We can operate for [TO FILL] months without new funding. We can prioritize [TO FILL: key deliverables] and defer [TO FILL: non-core expenses], focusing on [TO FILL: validation goal]. Our gaps and risks are [TO FILL]. Fundraising would accelerate validation, but it cannot replace customer demand.

Argus itself provides factual services and does not generate investment judgments, recommendations, target prices, portfolio instructions, or orders; customer Agents may independently analyze data they are legally permitted to access, subject to authentication, tenant, permission, and data license requirements. Actual legal and regulatory requirements must be assessed based on business activities, customers, data, and jurisdiction. Completed legal assessment: [TO FILL]; if it has not been completed, state that truthfully.

The project documentation describes a unified core boundary that handles authentication context, tenants, permissions, licensing, structure, source validation, and auditing; external source text is treated as data. Actual evidence of production configuration, threat assessment, security testing, and incident procedures is [TO FILL]. We will not claim zero risk or certification based on documentation alone.

We distinguish deterministic field processing from model-assisted steps by task; actual usage is [TO FILL: validated workflow]. Model-assisted facts need to retain their sources, confidence, and review status. External processing must also comply with data permissions. A model change requires reevaluation; we cannot assume all models are equivalent.

We will describe coverage, quality, latency, and anomalies, and use source tracing to help locate errors. Our specific service commitments, remediation process, and contractual responsibilities are [TO FILL: actual terms or proposed approach]. We will not promise unlimited liability in an introductory meeting or present quality labels as a guarantee.

The status of public repositories or code access is [TO FILL]. The licenses, third-party dependencies, and basis for Novai's ownership of the source code and data processing assets are [TO FILL]. We will check the chain of assignment or licensing for work contributed by founders, employees, and contractors; if the paperwork is incomplete, we will describe the gap. Public interfaces and private implementations should be treated separately.

In the first meeting, we can share a high-level architecture and redacted results. During technical due diligence, we can confirm the scope, recipients, confidentiality arrangements, and third-party restrictions before considering read-only materials or an on-site review. We will not provide production database access, complete customer datasets, or production credentials.

At an appropriate stage, we can arrange validation with authorized customers or suppliers. First, we will clarify the assessment purpose and questions, then coordinate consent and timing. Today, we can provide redacted evidence and avoid contacting customers without permission or exposing commercial relationships.

We aim to build sustainable revenue around [TO FILL: clearly defined customer task], then expand coverage and workflow value. The return logic depends on actual customer growth, retention, and unit economics. Potential exits can be discussed as long-term scenarios, but there is no identified buyer or guaranteed return. The current priority is to validate whether this business works.

We can understand your request, but these matters affect corporate governance and other fundraising arrangements. Please first provide proposed terms and a timeframe in writing. We will have authorized decision-makers and advisors review them before responding; we will not make an unauthorized commitment today.

We saw that your website lists a Global Tech portfolio that includes Aether AI, PuppyGraph, and MorphMind. We would like to understand whether your experience in AI infrastructure and enterprise adoption could apply to customer validation for Argus. The fit remains to be discussed by both sides; we are not treating portfolio names as an indication of investment interest or a commitment of resources.

SWC Global should answer. Suggested response for Novai: We can share non-sensitive introductory materials. If relevant portfolio companies, external experts, recording or transcription, or further distribution are involved, please confirm the recipients and purpose first. We will provide core information separately under controlled access.

SWC Global should give specific examples. Suggested response for Novai: The most useful support for us would be [TO FILL: needs related to customer validation]. We would like to understand what support can be put into practice. These resources are not committed; we will rely on actual arrangements.

Respond: Understood. For your fund, what would most change your assessment: paid usage, retention, coverage / licensing, or another metric? If there is a specific threshold, we can update you as we make actual progress. If the stage or direction is not a fit, we can end this process.

Respond: Thank you for your confidence. Please clarify the investment entity, proposed amount, terms, and next steps. We will verify company authorization and arrange the next steps. We will record today's verbal indication accurately; formal documents, closing conditions, and receipt of funds still need to be completed, so we cannot immediately announce that the round has closed.