SumizAI vs Caitlyn

What Caitlyn is

What Caitlyn is

Caitlyn is a private, secure AI solution that transforms how teams and customers access information. Trained on your business's own documents, policies, and data, not the internet, Caitlyn gives instant, accurate answers, 24/7. Instead of relying on the general knowledge of a model trained on the public internet, a company gets an assistant that knows only its own internal materials, eliminating the risk of hallucinations based on outdated or company-irrelevant information.

A paid-only model suggests Caitlyn targets companies willing to invest in a dedicated, secure AI solution for customer service and internal support, rather than relying on generic chatbots not tailored to a specific organization's specifics. A New Zealand origin places Caitlyn among startups from the Oceania region building enterprise tools competitive with solutions from larger tech markets.

Training exclusively on a company's documents, policies, and data, rather than the entire internet, radically increases the accuracy and reliability of answers in business contexts, where a precise answer aligned with internal procedures matters more than broad but imprecise general knowledge.

A customer service department using Caitlyn to give instant answers to customer questions about company policies gradually notices which types of questions come up most often and which of Caitlyn's answers need further clarification from a staff member. Without systematically documenting those observations, it's hard to improve the knowledge base Caitlyn is trained on for the next update.

SumizAI complements Caitlyn exactly at that point: observations about the most common customer questions and typical shortcomings in Caitlyn's answers can be saved as ai notes in a local Markdown file, building a list of improvements to implement at the next company knowledge-base update. A customer service manager ends up with ai notes forming a ready improvement plan, prepared based on real, observed customer interactions.

These tools complement each other: Caitlyn delivers instant, accurate answers based solely on a company's internal documents, while SumizAI ensures the observations from daily use of Caitlyn survive as ai notes ready to use when planning the next round of company knowledge-base improvements.

It's worth adding that such a cycle of observation and improvement, documented in these notes, means Caitlyn gets increasingly accurate and better matched to customers' actual needs over time, instead of remaining a static tool deployed once and left without further optimization.

A company providing B2B services to multiple corporate clients at once, using Caitlyn to handle each client's queries individually, especially benefits from keeping a separate these notes per client, documenting the specific questions and expectations tied to that relationship. An account manager, with access to such a note during the next client relationship review, can quickly verify whether Caitlyn is actually answering questions according to the current contract terms, rather than terms that became outdated months earlier, which could expose the company to a dispute with the client over imprecise information. Regularly comparing Caitlyn's answers against the actual current state of a given contract, documented in a note, becomes part of the company's standard customer-service quality-management procedure, carried out systematically rather than only in reaction to unhappy customers' complaints after the fact.

That is our reading of the record. For AlternativeTo's own wording, see its page for Caitlyn — the facts above were read on 2026-08-16.

Caitlyn in depth

A knowledge assistant that refuses to answer without a source

Caitlyn is an enterprise generative-AI platform whose defining constraint is negative rather than positive: the system is architected so that a response cannot be produced unless supporting material exists in the customer's own document set. That single design decision explains most of what follows — the deployment model, the pricing bracket, the industries it targets and the kinds of buyer it deliberately ignores. It is not a general assistant with a document upload feature; it is a retrieval layer for organisations whose reputational exposure makes a plausible-sounding wrong answer unacceptable.

The product is delivered as a cloud service and AlternativeTo records Online as the only platform, with English as the supported interface language. The listing classifies it under AI Tools & Services and Business & Commerce and tags it with artificial intelligence and generative AI. The licence is proprietary and commercial — there is no free tier, no trial listed on the directory and no self-service signup path on the vendor's site, which instead offers a demo booking and a downloadable seventeen-page information pack.

Who builds it

Caitlyn comes from Custom D, a software company based in Christchurch, New Zealand, which describes roughly two decades of custom development work preceding the product. The company presents Caitlyn as a response to a pattern it kept encountering in client engagements: off-the-shelf AI tools that could not deliver the accuracy, control or system integration the organisations actually needed. The leadership named publicly consists of Julie Ryan as chief executive, Sam Sehnert as chief technology officer, Josh Smith on growth and partnerships, Matthew Rhodes as chief marketing officer and Tim Ackroyd as product lead — an unusually specific disclosure for a product in this category, and one that makes the vendor easier to assess than most.

Note that the AlternativeTo entry, last updated in July 2025, points at getcaitlyn.ai. That domain now redirects to caitlyn.ai, so directory links still resolve, but the current site is the better source for anything time-sensitive.

Ingestion: what it will actually read

The vendor's claim on intake is breadth without preprocessing. Documents, tables, policies, reports, images, audio and video are all accepted, and the platform segments and interprets the content itself rather than requiring the customer to normalise files first. Two properties are singled out. The system reads visual elements alongside text, meaning a figure or a chart is not simply skipped. And it preserves the structural relationships inside a document — hierarchy, which section belongs under which, how one table relates to the surrounding argument — rather than flattening everything into an undifferentiated pile of text chunks. For an archive of scientific reports where the substance often lives in a results table rather than the prose, that distinction is the whole ballgame.

Reasoning over a knowledge graph, not a keyword index

Caitlyn's retrieval is built on knowledge-graph reasoning with ontology alignment, and the vendor illustrates the point with a domain example: the word ryegrass carries a different meaning in a cropping trial than in a pasture system, and a practitioner's question may legitimately span three separate reports published two decades apart. Keyword search collapses under that; so, frequently, does naive vector retrieval. The stated aim is multi-step questions where a generic assistant produces something fluent and wrong. The vendor puts accuracy on complex research tasks at ninety-five percent, though it does not publish the benchmark or methodology behind that figure, so treat it as a vendor claim rather than an independently verified result.

Every answer carries deep-linked citations that point to a specific table, figure or paragraph rather than merely naming a document. Combined with the rule that responses must be source-supported before generation, this makes verification a normal part of using the tool rather than an afterthought. The architecture description also mentions dual evaluation as part of the intelligence layer.

Sovereignty, permissions and where the data sits

Deployment is into the customer's own AWS account. Caitlyn describes four layers: a delivery layer offering APIs, white-label embedding and multi-channel distribution; an intelligence layer holding the graph reasoning, document intelligence, ontology alignment, dual evaluation and permission-aware governance; an AWS foundation using Amazon Bedrock to orchestrate underlying language models; and the customer's own environment underneath all of it. The company is an AWS Advanced Tier Services Partner with a Strategic Collaboration Agreement, a Generative AI Competency and several security specialisations, and the product is available through AWS Marketplace.

The consequence for the buyer is that research and intellectual property stay inside the organisation's boundary and are not exposed to public models or third-party infrastructure. Access control is enforced at content level and is permission-aware in a specific sense: a user can learn that relevant material exists without being shown content they are not cleared to read. That is a meaningfully harder problem than simply filtering results, and it matters in regulated settings where the existence of a document can itself be informative.

Because the deployment lives in the customer's account, the question of what happens after cancellation has a straightforward answer that most SaaS products cannot give: the underlying corpus was never anywhere else. AlternativeTo also lists ad-free operation, real-time collaboration, cloud synchronisation and no-code setup among the product's characteristics.

Branding, analytics and how customers deploy it

Caitlyn is typically fielded under the customer's own name rather than its own, and the published case studies show this clearly. Birchip Cropping Group runs it as ChatBCG, turning more than thirty years of independent local trial work into evidence-based answers for growers. Beef and Lamb New Zealand deploys it as Bella, an assistant that opens up a research portfolio the vendor values at over one hundred million dollars. The Foundation for Arable Research uses it to make decades of agronomic work reachable by growers and advisors, and Forest Growers Research applies it to forestry material. The customer base is described as organisations across New Zealand and Australia.

A less obvious feature is the analytics layer. Because every question is logged, the platform reports what practitioners are actually asking and, more usefully, where the knowledge base has no answer. That converts routine usage into a map of research gaps and can feed into decisions about where to direct future investment — a genuine secondary value that does not depend on the assistant being right every time.

Pricing, fit and the caveats

AlternativeTo records a subscription in the range of 600 to 3,000 US dollars per month. The vendor does not publish pricing itself; the site routes prospects to a demo or an information pack, so the real figure for any given deployment will depend on corpus size, environment and support. At that bracket the product is plainly aimed at organisations, not individuals, and the absence of a free tier is a deliberate filter rather than an oversight.

Sector coverage extends beyond farming — insurance, legal, education, financial services and healthcare all appear in the company's positioning — but the current site and every published case study lean heavily toward agricultural and primary-industry research. A buyer outside that space should expect less proven ground and should probe the ontology work required for their own domain.

The honest limitations are these. There is no independent user feedback to consult: the AlternativeTo listing has no likes, no comments and no reviews, and only eight alternatives suggested against it, so everything above rests on vendor material. The interface is English-only. Deployment requires an AWS account and the organisational willingness to operate infrastructure, which rules out teams wanting a sign-up-and-go tool. And the accuracy figure, the flagship number in the marketing, is unaudited. What Caitlyn does offer that most competitors do not is a clear, checkable account of where data lives and a citation discipline that lets a sceptical reader verify any claim the system makes.

For anyone comparing this against SumizAI: Caitlyn is an infrastructure-grade deployment for institutional knowledge bases with a corresponding price and setup burden, whereas SumizAI addresses a lighter, individual-scale need. They overlap only in the broadest sense of working with documents.

Product page: caitlyn.ai

What SumizAI is

A note-taking application built around a conversation with an AI model. You ask a question, the answer streams back, and the answers worth keeping become Markdown notes — filed into a vault that is a folder on your own disk.

A vault is a directory holding base.md, a generated table of contents up to six levels deep, and a notes/ folder with one .md file per note. Before a note is written it is checked against the ones already there, so the fourth note about the same idea gets merged instead of added. Links between notes are ordinary Markdown links to files that exist.

The model is never ours: you bring your own API key to one of seven providers — Anthropic, OpenAI, Gemini, Groq, OpenRouter, a local Ollama or your own server — and pay that provider directly. A question sends the table of contents plus at most five relevant notes within a 24,000-character budget, and the app shows you which five it used.

Every answer kept this way becomes one of these ai notes, ready to search again months later without hunting back through old chats.

Side by side

Where they differ

• Both mention AI. The difference is whose key it runs on: SumizAI never resells inference — you connect one of seven providers (Anthropic, OpenAI, Gemini, Groq, OpenRouter, a local Ollama or your own server) with your own API key and pay that provider directly. Check what Caitlyn does with the model bill before comparing prices.
• Caitlyn is paid software. Compare the two numbers directly: SumizAI is one dollar a month, one plan, and your AI provider bills you separately for the model you chose.
• AlternativeTo's description of Caitlyn does not mention Markdown, so check what format your notes end up in before you fill it up. SumizAI writes standard .md files to a folder you picked, and an unpaid account can still export all of them.
• Caitlyn is described as something more than one person uses at once. SumizAI is not: there are no shared vaults, no comments, no permissions and no sync between devices. If the work is a team's, that is a reason to pick Caitlyn over SumizAI.
• Caitlyn runs in the browser. SumizAI is a desktop and mobile application, and on the desktop the vault is a folder on your own disk rather than a document in someone's cloud.

Which one to pick