Cognosys pricing

About

Cognosys was an autonomous AI-agent product — pitched as an “AI employee” that could do web research, schedule, handle email, and run browser tasks toward a high-level goal. It was founded in 2023 in Vancouver by Sully Omar and Homam Malkawi, and raised a roughly $2M seed round led by GV (Google Ventures) and Untapped Capital, with an unusually strong angel list: Replit CEO Amjad Masad, Vercel CEO Guillermo Rauch, and both Cohere co-founders Ivan Zhang and Aidan Gomez.

The product did not stay still. In October 2024 it rebranded to Ottogrid after a platform overhaul, pivoting from open-ended autonomous web agents to a “smart table” research interface (an Autogrid where each cell is its own AI agent — extract data from websites into spreadsheets, enrich sales-lead lists, analyze documents at scale). In May 2025, Cohere acquired Ottogrid (terms undisclosed) and folded the product into its North enterprise application; the standalone product was sunset with a transition window for existing users.

As of this writing, cognosys.ai returns a Vercel “This deployment is temporarily paused” page — there is no live product or rate card to buy. This blueprint is a post-mortem: the pricing below is the historical consumer card, recovered from contemporaneous third-party reviews of the old /pricing page, not a live quote.

For the surviving product lineage, see Cohere.

Pricing summary : How Cognosys’s pricing worked

Cognosys ran a clean consumer freemium subscription with three public tiers plus a custom Enterprise option. Crucially, it did not price per seat — it metered two usage value metrics: messages per month and workflow executions. That made it read like a productivity app (ChatGPT-style) rather than a per-user SaaS.

• Free — $0/month: 100 messages/month, 1 concurrent integration, 1 document upload per query, 1 active workflow automation, 20 workflow executions, GPT-3.5.
• Pro — $15/month: all Free features, raised to ~1,000 messages/month, up to 10 active workflow automations, multiple document uploads per query, priority support, early access.
• Ultimate — $59/month: all Pro features, unlimited usage on most features (messages, integrations, workflow automation), priority support, early access.
• Enterprise — Custom: all Ultimate features plus custom integrations, dedicated support, advanced security; sales-quoted.

What makes this different: the value-metric choice. Most agent tools of the era either gave away unlimited chat or charged per seat. Cognosys instead gated on how much agent work you ran (messages + workflow executions), with the model itself improving as you paid up (GPT-3.5 on Free, GPT-4-class on paid). The jump from $15 Pro to $59 Ultimate is the classic “unlimited removes the meter anxiety” upsell.

Pricing by product

Sales motions across products: pure self-serve / PLG on the three published tiers — sign up, hit the Free caps, upgrade with a card. Enterprise was the only sales-led path. There was no seat negotiation on the consumer tiers.

Hidden costs : What Cognosys users actually paid

The model was refreshingly flat — the “hidden” cost was hitting the caps, not surprise overages. There was no published per-message overage rate; instead the design pushed you up a tier when you ran out of messages or workflow executions.

The real cost of Cognosys today is continuity risk, not dollars: the product is sunset. Want to model what an agent plan like this would have cost you? Use the Cognosys pricing calculator to estimate against the historical card, or the pricing calculator hub for comparable live AI-agent pricing.

Pricing evolution : Cognosys pricing history and changes

Cadence

Tracked range: 2023–present. The original $15/$59 consumer card is recovered from third-party reviews; the live page no longer renders a rate card.

Notable changes

• 2023 — Launched as Cognosys with the Free / $15 Pro / $59 Ultimate freemium card.
• 2024-10 — Rebranded to Ottogrid after a platform overhaul; the value proposition shifted from autonomous web agents to a research-focused smart table.
• 2025-05-16 — Cohere acquired Ottogrid; standalone product wound down.
• 2026-06-16 — cognosys.ai returns a paused-deployment page; pricing no longer live.

What’s unique : Cognosys’s distinctive pricing mechanics

1. Dual usage meter, not seats. Cognosys gated on messages per month AND workflow executions simultaneously — two value metrics for two different behaviors (interactive agent chat vs. scheduled automation). Most peers picked one or defaulted to per-seat. See choosing the right usage metric.

2. The model itself was a tier lever. Free ran on GPT-3.5; paid tiers unlocked GPT-4-class. That folds inference cost directly into the upgrade decision — you pay for smarter agents, not just more of them.

3. “Unlimited” as the top-tier hook. Ultimate at $59 removed the meter on most features. That’s a deliberate anxiety-killer: heavy users stop watching their message counter and just pay a flat premium — a pattern explored in bill shock and cost unpredictability.

Strengths & weaknesses

Billing UX : Cognosys billing controls and transparency

• Billing controls — Self-serve monthly subscription with a card; tier upgrades unlocked higher message and workflow-execution caps. No usage-based overage billing to manage — you simply upgraded.
• Usage visibility — Plans were defined by hard caps (100 / ~1,000 / unlimited messages; 20 / more / unlimited executions), so “usage” was effectively a quota you consumed against your tier rather than a metered running bill.
• Payment options — Standard consumer card checkout on the self-serve tiers; Enterprise was invoiced via sales.
• Today — None of this is live. The billing surface is gone; the domain serves a paused-deployment page.

Strategic wins : Why Cognosys’s pricing decisions worked

1. A consumer price the market could say yes to instantly

By publishing a $0 / $15 / $59 card with no sales gate, Cognosys made adoption a credit-card decision. That PLG motion is exactly how consumer AI tools build the early usage base that makes an acqui-hire attractive — and Cognosys was acquired inside two years. See how AI companies structure pricing.

2. Pricing the work, not the seat

Choosing messages + workflow executions as the meters tied revenue to value delivered (agent runs completed), not to org size. For a single-user productivity agent, that’s the right axis — see outcome-based pricing trends.

3. The pivot-then-exit path

When the open-ended-agent thesis got crowded, the team repriced its value (Ottogrid’s research smart-table) rather than its rate card, and exited to Cohere. The pricing stayed simple while the product found its lane — a reminder that for early agent companies, distribution and a clean entry price can matter more than rate-card sophistication. Related: usage-based pricing strategy.

Areas to improve : Gaps in Cognosys’s pricing approach

1. Two caps, unclear which bites first

Pricing on both messages and workflow executions meant users couldn’t easily predict which limit they’d hit. A single headline meter (or a clear “you’ll run out of X first”) would have made the upgrade trigger legible. See bill shock and cost unpredictability.

2. “Unlimited” that wasn’t fully unlimited

Ultimate promised “unlimited usage on most features.” The “most” is the catch — without naming which limits persisted, the top tier traded clarity for a marketing word. Naming the few remaining caps would have built more trust.

3. No mid-tier between $15 and $59

The 4x jump from Pro to Ultimate left no landing spot for a user who needed more than 1,000 messages but nowhere near “unlimited.” A usage add-on or a $29 middle tier would have captured spend that otherwise either churned or over-bought. See choosing the right usage metric.

Key takeaways

1. A clean freemium card is a growth weapon. $0 → $15 → $59 with no sales gate let Cognosys convert curiosity into card-on-file fast — the foundation of its PLG traction and eventual exit.
2. Pricing the work beats pricing the seat for single-user agents. Messages + workflow executions tied revenue to value, not headcount.
3. Dual meters create prediction risk. Two simultaneous caps obscure the upgrade trigger; pick one headline meter where you can.
4. “Unlimited” needs footnotes. Vague top tiers erode the trust that flat pricing is supposed to buy.
5. Pricing stability through pivots is fine — product trust isn’t. Cognosys kept its rate card simple while rebranding twice; for buyers, the lesson is to weigh continuity risk, which is now realized: the product is sunset.

UBP implications

1. Choose meters that map to the user’s behavior. Cognosys’s messages (interactive) + executions (automation) split shows you can meter two behaviors — but only if users can tell which one governs their bill.
2. Use the underlying model as a price lever. Gating GPT-3.5 vs GPT-4 by tier folds inference cost into the upgrade, a pattern any AI product running multiple models can borrow.
3. Flat “unlimited” top tiers cap your upside. Ultimate’s flat $59 leaves money on the table from your heaviest users; a usage-based or hybrid top tier would capture it. See introduction to usage-based pricing.

Sources

• Cognosys pricing page — now returns “This deployment is temporarily paused” (accessed 2026-06-16)
• Cohere acquires Ottogrid — Techcouver (accessed 2026-06-16)
• Vancouver AI startup Cognosys acquired by Cohere — VanTech Journal (accessed 2026-06-16)
• Cognosys pricing & features — AI Agents List — historical rate card (accessed 2026-06-16)
• Cognosys AI review — AIChief (accessed 2026-06-16)

Bottom line

Cognosys was a well-priced consumer AI-agent product — a transparent $0 / $15 / $59 freemium card metered on messages and workflow executions, not seats — that found product-market fit faster than rate-card sophistication. It rebranded to Ottogrid (Oct 2024), was acquired by Cohere (May 2025), and is now sunset; cognosys.ai serves only a paused-deployment page. Treat this as a post-mortem: the pricing lessons (meter the work, model-as-lever, footnote your “unlimited”) outlive the product. Browse the pricing blueprint for live AI-agent companies.

Want to compare against active AI-agent companies? Browse the pricing blueprint.

Pricing timeline : Major events on a vertical axis

Each milestone below corresponds to a public pricing change, product launch, or material adjustment. Major events use a filled marker; minor adjustments use a faded one.

Site paused — pricing no longer live

Jun 2026

cognosys.ai returns a Vercel 'deployment temporarily paused' page. No live rate card exists; pricing documented here is the historical consumer card recovered from contemporaneous third-party reviews.

Acquired by Cohere; standalone product sunset

May 2025

Cohere acquired Ottogrid (terms undisclosed) and folded the product into its North enterprise app. The standalone Cognosys/Ottogrid product was wound down with a transition period for existing users.

Rebranded to Ottogrid

Oct 2024

After a platform overhaul, Cognosys rebranded to Ottogrid, pivoting from open-ended web agents to a 'smart table' research interface (each cell an AI agent) aimed at sales/market-research workflows.

Launched as Cognosys — freemium AI agents

Jan 2023

Cognosys launched as a goal-driven autonomous AI agent for the web with a consumer freemium model: Free (100 messages/mo), Pro $15/mo, Ultimate $59/mo, metered on messages and workflow executions.