Same mission. Better ship.

Why Hatz 2.0

Here's something we hear from almost every MSP we talk to: your clients are already using AI. They're pasting customer data into free chatbots, expensing random subscriptions, and nobody in the building knows exactly what's leaving or where it's going. You didn't sell them that AI. But you're the one they'll call when it goes sideways.

That's the gap Hatz was built to close. We bring the world's best AI models, built-in security, and smart cost routing into one place, delivered through the MSPs your clients already trust, so AI stops being a loose end and starts being something you manage, sell, and grow with.

Hatz 1.0 did that job well. It put real AI power in your clients' hands, and it taught us a lot about how small businesses actually use it. Users chose their own mode, Lite, Performance, or Turbo, and picked their own balance of cost, output, and speed. That choice mattered, and our partners and their clients got real value from it.

But we keep evolving with the times, and we listen. You told us where you wanted us to level up next: "Show me what I'm being billed for, and why." "My clients don't want to pick modes, drag nodes or learn prompting, and my power users want more." "Enrichment tenant by tenant is a chore, every client is at a different point with AI, and a few power users block standardization."

Hatz 2.0 is our answer to all of it. Same mission, with more people on board: MSPs, power users, and first-time users.

What that means

At Hatz, we're invested in making AI do more work for you, without the cost. That's not a slogan; it's a standing research effort. We spun up a dedicated AI research team with one mission: figure out how to measure what "better" actually means for small businesses, then engineer toward it. The team built a custom benchmark from how real MSPs and SMBs use AI, and scored everything on it: how Hatz performed, how individual models performed, cost-to-complexity, and how our model routing was doing. Then we rebuilt the engine of the whole platform against that benchmark.

The clearest proof is Auto mode. With Hatz 1.0, users chose between Lite, Performance, and Turbo, balancing cost, output, and speed themselves. After looking into several months of real Auto Mode data, we found a way to make it even easier to maximize all three at once. Hatz 2.0 replaces the three tiers with one Auto mode that picks the right model for each task. Under the hood it even mixes models within a single task, a heavyweight model for the complex analysis and a lighter one for reading the file or generating the PDF, with a cost-complexity algorithm that factors in cached input. On our SMB-1 benchmark, the new Auto mode is:

  • 24% more accurate than Turbo
  • 12% cheaper than Performance
  • 15% faster than Lite

And against the leading frontier models running on our platform, the gap is even wider:

  • vs. Sonnet 5.5: +12 percentage points in critical pass rate, 68% fewer credits, and 68% faster
  • vs. Terra 6.1: +6 percentage points in critical pass rate, 76% fewer credits, and 76% faster

Your users no longer have to pick their tradeoff. Auto mode routes each task to the model that wins on cost, output, and speed together. That is smart cost routing doing exactly what it's for: control on spend, with no model lock-in. And because every change we make now runs against the benchmark, we iterate quickly and reliably, always moving forward. We'll publish the full benchmark results on our blog and keep them updated.

The new engine also stretches what long-running work can do. Hatz 2.0 can run for hours, with dynamic decision models that stay aware of your credits. You can spin up multiple prompts at once, watch them run in the background, and switch between them.

What's in it, and why

Hatz AI gives MSPs and SMBs full control over their security, their spend, and their scope. Everything in Hatz 2.0 falls into three buckets: easier to manage, easier to use, and easier to sell.

Easier to manage: control over spend, for the MSP's own business

Itemized Billing. Every line on the invoice now has a reason behind it. The Subscription Overview itemizes every charge by client, month by month, meaning cleaner invoices and easier client conversations.

Per-User Reporting. Cost now follows the person actually using it. See usage by person, for one client or all of them, and export it. It's available through the API too, so it flows into your own reporting. The result: predictable costs and accounts that are easier to defend and expand.

Easier to use: for average users, power users, and teams

For the average user, Hatz 2.0 and Automations in Chat. No more picking modes; Auto mode just works. Building an automation no longer means learning a workflow builder: you ask, schedule, and edit by typing in chat. Value on day one, with no training required. Frictionless from day one, for every team.

For the power user, AI Gateway. A few power users want direct model access, and that blocks standardization. The AI Gateway puts raw models on Hatz billing at the provider's public API price plus a simple 10% fee, with a budget for each client. Power users stay aboard, spend stays centralized, and deals move forward. And for tools without an API at all, Hatz can combine the browser with vaulted credentials and sign into the site itself; compatible credentials are proxied so they never touch the code sandbox.

For the team, Spaces, Skills, and Notification Center. A space is a shared home where chats, files, automations, and artifacts use the same instructions and context, shareable with specific people, a group, the whole organization, or a managed customer. Skills let you save a task once and share it with the whole team, which partners on our roundtable called a major win for migrating power users off other platforms and for lowering credit usage in large workflows. The Notification Center gives admins one inbox for what finished, what needs attention, and what's new, with controls over what, if anything, gets sent to your customers. That's scope control: you decide where AI starts and ends.

Easier to sell: in front of a client

Activate updates. A new UI with bulk enrichment, plus refreshed Activate and Learn videos, meaning easier demos and faster onboarding.

Hatz 2.0 for every client. Your clients fall into two groups: those barely using AI and those really using AI. The new experience works for both. Meet your clients where they are, crawl, walk, run, from first prompt to daily habit.

AI Gateway. It answers the power user objection in the room: unblocked deals and net-new pipeline.

The line to bring to clients: "However you use AI today, Hatz meets you where you are and wraps controls around it, so you get the most productivity without risking your data security, spend overages or scope creep."

What's next

Hatz 2.0 is just the beginning of where Hatz is going to take its users and its partners. We will continue to make AI easier, and feedback from our partners will drive us along the way. If you have a use case that isn't stacking up as well as it should, send it to us: we can tune Auto mode against it, and every change runs against the benchmark so we know we're making progress. Follow along on our roadmap at hatz-ai.canny.io/changelog.

For your customers, the outcome is confidence: they trust your controls around their data and guardrails, and they see the ROI on their decision. For your MSP, the outcome is revenue: you decide how AI shows up across every client, and you capture recurring margin that compounds as they grow. AI compounds, software doesn't.

Our advice to partners getting started with Hatz 2.0? "Just get started. It's easier than you think."

Control for every business. AI for every team. Check out the latest Partner Round table for a deep dive into Hatz 2.0: watch on YouTube.