ntelio Portfolio in Practice
Practice  ·  ntelio.ai  ·  MMXXVI
Companion document to the ntelio investment thesis Issued 2026-05-08

The thesis — in practice

The shape every product followed. The discipline we never break.

The thesis says what we will and will not invest in. This page shows how that runs — the six-step engagement shape, the origin of every product we’ve shipped, what we will not touch, and why the resulting portfolio composition compounds.

Read the shape → Back to the thesis →
01 — The shape

Every product followed the same six steps.

No exceptions. The shape is what makes ntelio a studio rather than an opportunistic consultancy.

  1. 01

    Real, paying vertical pain.

    Surfaced by an enterprise partner with a budget line item, or rooted in our own core experience and validated by a real market. Not a hypothesis. Off-the-shelf frontier AI cannot solve it on the terms the work requires.

  2. 02

    PRISM engagement.

    We turn “we should do something with AI” into a scoped, measurable, build-ready vertical opportunity. Five phases. See PRISM →

  3. 03

    Vertical build on the platform.

    Paid revenue today. Market signal for tomorrow. Co-built with the partner; deployed; in production.

  4. 04

    Bucket check.

    Does the build land in A, B, C or D? If not, we deliver the engagement and stop. We do not invest beyond it.

  5. 05

    Generalisation.

    The patterns we find generalise back into the platform. The platform makes the next product cheaper from an engineering, operational complexity and costs perspectives. It’s a compounded acceleration edge.

  6. 06

    Commercialisation, if the vertical generalises.

    Cartaja, DocGenie, ArtBound and TheMeerkat all came out of this final step. The portfolio is the residue of the engine.

02 — The origins

Where each product came from.

Concrete, not theoretical. Two products were born from inside the team — the category was visible from our own platform and vertical work before any one partner needed it. Two were partner-led — the engagement revealed the category. Never a theme picked off a market map.

Cartaja

Internal A B C
OriginInternal product. A decade of conversational-commerce work the team had already done across MENA SMEs and major venues made the messaging-commerce long-tail visible from inside the platform.
PainMerchants running a business on WhatsApp DMs with no website, no app, no developer; orders, catalogue, checkout and support all collapsed into a single chat thread.
What we builtA system of record on the channel where commerce already happens. Catalogue, cart, checkout, fulfilment, multi-channel operational moat.
Why it became a productWe saw the category before any single partner needed it built. The platform and the team's experience were the precondition for Cartaja, not a design customer.

DocGenie

Partner-led D C
Origin engagementsThree sovereign deals: a top-2 global logistics company (proposal generation), a major Abu Dhabi public administration (private-AI document workflow), and an urban-planning department (50-page feasibility studies).
PainDocument workflows that cannot leave the perimeter. Cloud frontier AI is structurally disqualified by privacy, audit and repeatability requirements.
What we built100% on-premise / sovereign-cloud document generation, structured-and-repeatable output, audit-ready by construction.
Why it became a productThe category — sovereign, audit-ready, repeatable doc generation — only became visible after the third engagement. The pattern was the product.

TheMeerkat

Internal A B C
OriginInternal product. ntelio's own conviction about the agent-era information landscape: frontier AI floods the internet with synthetic claims, and the load-bearing primitive for trusted signal becomes cross-source corroboration.
PainSingle-source claims, opaque platform scoring, no cross-platform synthesis. The trust layer was missing for OSINT analysts, journalists, and intelligence teams.
What we builtTheMeerkat — a consumer and prosumer news diet built on an in-house corroboration engine (ingest, claim decomposition, source graph) we developed for this product.
Why it became a productAn experience-led bet on the verification layer. ntelio's first consumer product, built without a single design partner — the category was visible from inside our platform work.

ArtBound

Partner-led A C D
Origin engagementMark Hachem and other partner galleries — cultural authorities running auction businesses without auction infrastructure that fit their economics or channel.
PainNo auction system that respects how galleries actually run: gallery-set commissions (not platform-set), the gallery's own collector address book kept private, and bidding on the channel collectors already use.
What we builtAn art auction system delivered over WhatsApp, with a gallery-friendly commission structure on each sale and an onboarding fee per gallery. Auctions run against the gallery's exclusive collector address book.
Why it became a productThe true moat is the exclusive gallery relationships and the address book each gallery brings — both stay with the gallery, never the platform. Revenue is recurring: commissions on every sale, plus onboarding fees as galleries come online.
03 — What we will not invest in

Stated as plainly as the affirmative thesis.

Because it is just as load-bearing.

  • Horizontal SaaS dashboards.

    CRM, ticketing, project management, HR, generic admin panels. Frontier agents are absorbing these surfaces. We will not build them, fund them, or extend a vertical product into them.

  • “Better chatbot” plays.

    Building a smarter human-facing chatbot as the core product. The buyer or end-user will increasingly ask their own agent. We invest in the storefront the agent calls, not the chatbot the human opens.

  • Frontier-model competition.

    We do not train frontier foundation models. We do not compete on raw model capability. We compose, deploy and operate frontier models inside vertical workloads where the substrate, surface or sovereignty wraps the model.

  • Non-defensible verticals.

    Workloads that are pure prompt-and-API away from being a feature inside ChatGPT or Claude. If our only moat is a system prompt, we do not invest.

  • Pre-revenue research bets without either a paying partner or a strong root in our own experience.

    Cartaja, DocGenie, ArtBound and TheMeerkat each exist because the engagement was real before code was written — surfaced by a paying partner, or rooted in our own platform and vertical work. We do not break that discipline.

04 — Why the portfolio composition works

Three reinforcing properties.

Each tied directly to the thesis.

  1. 01

    Two-axis diversification.

    The portfolio is balanced across the infrastructure axis (Cartaja, TheMeerkat, the ntelio platform — the things agents call) and the sovereign / specialised-vertical axis (DocGenie, ArtBound — the things agents cannot enter). If frontier capability accelerates faster than we expect, the infrastructure side compounds. If regulation, sovereignty and creator economics dominate the next cycle, the vertical side compounds. Both are bets on the same thesis from opposite ends.

  2. 02

    Studio mechanics, not project mechanics.

    Every enterprise engagement is both paid revenue today and a market signal for tomorrow. Patterns generalise into the ntelio platform. The platform makes the next product cheaper to build. When a vertical generalises, we commercialise it. Cartaja, DocGenie, ArtBound and TheMeerkat all came out of this loop.

  3. 03

    Pricing surfaces match the layers.

    Merchant subscriptions on the SaaS side; transaction take-rate and metered MCP access on the agent-addressable side; sovereign-deployment licences on the on-prem side; commissions on every auction sale plus per-gallery onboarding fees on the curatorial side. The revenue model tracks the layer-split — we charge per layer that survives, not per seat in a dashboard.

05 — In summary

The bet, restated.

  • Frontier AI capability jumps clarify ntelio’s portfolio. They do not threaten it.
  • We invest in four categories and no others: infrastructure for frontier AI, agent-callable surfaces and trust layers, specialised creation surfaces, and regulated / sovereign verticals.
  • Cartaja and TheMeerkat are infrastructure plays — the substrate the agent era runs on.
  • DocGenie and ArtBound are vertical plays — the workloads frontier AI cannot enter, by regulation or by the nature of the relationship.
  • The ntelio platform is the substrate beneath all of the above. PRISM is how we make sure every new engagement lands in the right bucket.
  • The risk is not that frontier AI gets stronger. The risk is mis-positioning a product as something frontier AI replaces, instead of something it runs on or cannot reach.
  • We have made our bet on the second category, twice on each axis, and we do not intend to make any other kind of bet.

Return to the thesis →

06 — Speak to us

For investors and partners.

You back applied AI companies with revenue and IP across a portfolio designed around the layer-split.

We’ll send the long-form thesis, an updated portfolio brief, and an introduction to the founder best placed to answer your specific questions.

investors@ntelio.ai → partners@ntelio.ai →