Caspr Review: The $15 AI That Settles Which Number to Trust

Two credible reports give you two different numbers for the same market, and both look well sourced. That is the exact moment most research stalls: you either pick the figure that supports the decision you already want, or you spend weeks tracing why the sources disagree. Caspr, a new analytical AI launched on November 4, 2026, is built for that moment. Instead of handing you another summary, it says it digs underneath each figure, explains why the sources differ, and recommends which number fits your specific question.

The idea is easy to grasp and easy to oversell, so this review separates what Caspr actually announced from what it has not yet shown. The company, based in Gurgaon, India, says its service is fully live at caspr.ai, with more than 1,600 signups at launch. Pricing starts at $15 per report. What follows is what that money buys, who it suits, and the gaps a careful buyer should notice before relying on a Caspr report in a boardroom.

What Caspr is, in plain terms

Caspr calls itself an Analytical AI company, and the distinction matters. Generative AI systems are trained mainly to understand language and produce fluent text. Caspr says it was built for a different design problem: sourcing data, assessing where sources agree and disagree, and concluding with a position that holds up when questioned. In its own framing, language was a problem it did not need to solve from scratch. Analysis was the work left to do.

The workflow starts with a business question, not a prompt for a generic essay. You describe what you need and can attach your own data. Caspr then asks a few clarifying questions, a step the company says usually takes about two minutes and that users say sharpens their own thinking. The output is a finished report, not a chat transcript: editable in the tool and exportable as a PDF or PowerPoint deck, in Caspr's template or your own.

Each report is drawn, according to the launch announcement, from more than 25 million curated public and proprietary sources, refreshed with the latest information every time a report runs. That source count is a company figure, not an independently audited one, and the announcement does not publish a source list, a methodology paper, or an accuracy benchmark. Treat it as a claim about scale, not proof of correctness.

The example that explains the product

Caspr's launch example is worth reading closely, because it shows the kind of disagreement the tool targets. Published estimates of the UK ready meals market ran from $5.75 billion to $8.70 billion, a gap wide enough to change an investment case. Caspr found the gap was mostly definitional: publishers counted pizza, meal kits, and direct-to-consumer packaged meals differently.

The useful part was not the range itself. It was the next step: setting out which definition suits which decision, and recommending the mid-range figure for investment use. That is the product in miniature. A range tells you sources disagree. A recommendation, with the reasoning attached, tells you what to do about it. Whether Caspr does this reliably across topics is the open question, since the company has published one worked example, not a test set.

Caspr pricing: what $15, $80, and $300 buy

Caspr does not sell by the page, and it does not describe a monthly subscription in its launch announcement. You pay per report tier, and a report runs as long as the topic needs. The three Market Research tiers are:

GBP and CAD figures are approximations at November 5, 2026 market rates, verified at publication (about £0.74 and CAD 1.39 per US dollar). Caspr lists its prices in US dollars, and the actual amount you pay in another currency will depend on your payment provider's rate and any fees.

The fine print matters here. First, the $300 Intelligence tier arrives later in 2026, according to the announcement, so it cannot be bought yet. Second, depth depends on the topic: a Brief on a thinly covered niche will not magically gain sources that do not exist. Third, the announcement mentions no free trial, no revision policy, and no refund terms. If a report misses the mark, the published material does not say what happens next. Ask before you buy a Study or an Intelligence report, not after.

For context, Caspr contrasts its prices with traditional analysis of this kind, which it puts at $5,000 to $50,000 per engagement and several weeks of work. Those are Caspr's own comparison figures, not a surveyed market rate, but the direction is plausible: a $15 Brief is roughly 0.3% of the bottom of that traditional range, and an $80 Study is 1.6%. The saving is real only if the report is good enough to cite. A cheap report you cannot defend costs more than an expensive one you can.

Who Caspr is built for

The target audience is specific: consulting and research firms, strategy and corporate development teams, investors, agencies, startups raising capital, category managers, and graduate researchers. In other words, people whose job includes defending a number to someone else, whether that is a partner, a board, an investor, or an examiner.

That focus explains the export options. A PDF or PowerPoint deck in your own template is a report meant to be forwarded, not a chat log meant to be retyped. It also explains the enterprise notes: Caspr says it supports data-sovereignty requirements, including on-premise deployment for enterprise customers, and that it holds ISO 27001:2022 certification and is GDPR compliant. Those are company statements from the launch announcement. Enterprises should still confirm the scope of the certification and the deployment terms in writing, since the announcement does not publish them.

How to use Caspr, step by step

Based on the official launch description, the process runs like this:

1. Describe your business question in plain terms, and attach your own data if the question depends on it.
2. Answer Caspr's clarifying questions, which the company says usually take about two minutes.
3. Choose a tier: Brief for a single-level answer, Study for a deeper cut, Intelligence later in 2026 for a multi-approach report.
4. Wait for the report: under 15 minutes for a Brief, typically 15 to 20 minutes for a Study.
5. Edit the report in the tool, then export it as a PDF or PowerPoint deck in Caspr's template or your own.

The strongest use is a question with a decision attached: which market size to use for an investment memo, which pricing benchmark to cite in a strategy deck, which definition a category plan should adopt.

What Caspr has not proven yet

A launch announcement is not a review, and Caspr's has the usual limits. There are no independent reviews of its reports yet, only syndicated copies of the company's own announcement. There is no published accuracy test, no side-by-side comparison against a human research firm on the same question, and no public sample report to inspect. The 1,600-plus signups show interest, not satisfaction, and signups are not paying customers.

Three gaps stand out. The first is source transparency: a report you can cite needs visible evidence, and while Caspr says the evidence behind each figure is available to anyone who wants it, no sample shows how complete that evidence trail is. The second is handling of disagreement inside its own source base: the announcement does not explain what Caspr does when sources conflict on facts rather than definitions, or when the best answer is that no reliable figure exists. The third is coverage: 25 million sources sounds large, but for a niche market or a non-English topic, the relevant sources may be few, dated, or paywalled. Caspr does not say how it handles paywalled or proprietary data it cannot access.

None of this makes Caspr a poor product. It makes it an unproven one, and the honest way to try it is on a question where you already know the answer, or where a wrong recommendation costs little. Compare its Brief against your own quick research before trusting a Study in a live deal.

Common mistakes when trying a tool like this

The first mistake is asking for a number without naming the decision. "What is the market size?" has several correct answers, depending on definition. "Which market size should we use for an investment memo, and why?" is a question Caspr can actually answer.

The second is treating speed as quality. A Brief in under 15 minutes is impressive, but a fast report that cites a definition unsuited to your decision is worse than a slow one that gets the definition right. Read the reasoning, not just the recommended figure.

The third is skipping the evidence check. If a figure will appear in a board deck or a fundraising document, open the cited sources yourself, at least for the headline number. That check takes minutes and is the difference between citing research and repeating a claim.

Who should try it, and who should wait

Try a $15 Brief if you regularly reconcile conflicting figures, if you need a citable starting point for a memo, or if you want to test whether analytical AI fits your workflow at low cost. The price is low enough that a single useful report justifies the experiment, and a poor one teaches you the tool's limits cheaply.

Wait if you need the Intelligence tier, which is not out yet, or if your work requires audited sources, a named methodology, or a guaranteed revision process. Those buyers should ask Caspr for a sample report, the certification scope, and the revision terms before committing to an $80 Study, let alone an enterprise deployment. Readers weighing other agent-style tools may also want to compare approaches: NorthCUA takes the opposite route, a free open-source agent that operates your computer directly, while Together AI's polymorphic inference shows how the infrastructure behind AI tools is itself being tuned for cost and speed.

Frequently asked questions

Is Caspr free?

No free tier is mentioned in the launch announcement. Market Research reports start at $15 for a Brief and $80 for a Study, with the $300 Intelligence tier arriving later in 2026. Enterprise pricing, including on-premise deployment, is not published.

How long does a Caspr report take?

A Brief arrives in under 15 minutes and a Study typically takes 15 to 20 minutes, according to the company. Timing for the Intelligence tier is not published, since that tier is not yet available.

What makes Caspr different from a chatbot with web search?

Caspr says it is built to source, assess, and conclude, rather than to produce fluent text. Its stated difference is the recommendation: when sources disagree, it explains why and says which figure fits your question, instead of listing conflicting numbers and leaving the choice to you.

Can I cite a Caspr report in a board deck or investment memo?

Caspr positions its reports as citable, and says the evidence behind each figure is available. Before citing one in a high-stakes document, verify the headline figure against its cited sources yourself, since no independent review of Caspr's reports exists yet.

Who founded Caspr?

Caspr was founded by Joy Sharma, who previously spent five years at McKinsey and started EZ in 2017, and co-founded with Jayant Jha, who joined EZ in 2019 and moved to Caspr full-time after being invited to co-found it in 2025.

Caspr is a narrow tool with a clear job: settle which number to use when sources disagree. At $15, the Brief is a cheap test of that idea, and the worked ready-meals example shows the reasoning style at its best. The product's real test comes next, when independent users compare its recommendations against figures they already trust. Until then, use it as a fast, citable starting point, and keep the final judgment where it belongs: with you. This is general information, not financial or legal advice.

Sources

• Caspr launch announcement, ACCESS Newswire, November 4, 2026: Caspr Launches Analytical AI That Tells You Which Number to Use
• Caspr official site: caspr.ai
• Exchange-rate references for November 5, 2026 approximations: USD to GBP, about 0.7408 and USD to CAD, about 1.3935