Skip to main content

Getting More From Every Credit

Use current Auto controls, focused context and observed usage to manage AI spend.

Get more value from your AI budget by choosing the right tools, keeping tasks focused and saving repeatable work. These habits help you stay in control of spend while getting more done.

TL;DR — Credit Optimization Best Practices

  • Default to Auto. Auto chooses a model for the task and is more cost-efficient across a range of tasks when you consider successful completion and cost per completed task. Choose an available named model when you have a specific need.

  • Check The Multiplier. Hover over an LLM in the selector, or use the multipliers tab to see how much more or less credit-efficient a model is in relation to a ‘flagship model’ at 1X.

  • Let helpers handle specialized work. Helpers are optimized for their tasks, use fewer credits and do a better job on those tasks.

  • Start new chats for new topics. Relevant history and summaries help maintain continuity. A fresh chat for unrelated work keeps the context focused.

  • Turn recurring prompts into Skills and automations. Save a reusable procedure as a Skill, and ask Chat to create an automation for work that should run again.

  • Keep context tight. Upload only the relevant pages, paste only the relevant excerpt. Input size is part of the cost.

  • Think ROI, not just spend. High-value usage at higher cost beats low-value usage at any cost. Ask whether the credits are earning their keep.

  • Admins: use the controls available to you. Model access, per-user role limits, Extra Usage budgets and usage dashboards help you manage spend at scale.

Why Credits Vary More Than You Might Expect

Credits measure AI consumption. Usage depends on the model, input and output, context, files, tools and the work required to complete the task. Your package provides a monthly allowance shared across the workspace.

Credits are not a flat per-message fee. Models, context, files, tools, helpers and the number of steps all contribute to the cost of completing a task.

It gives you a set of controls for how you use AI. Understanding these dials is the first step to getting real value out of every credit you spend.

Start with Auto

Start with Auto for everyday drafting, summaries, questions and analysis. It gives Hatz the flexibility to choose a model suited to the work.

Auto is more cost-efficient across a range of tasks and uses fewer credits than comparable models when successful completion and cost per completed task are considered.

Choose an available named model when you need a particular capability or want to compare results, then return to Auto for everyday work.

Auto handles model selection for the request. The model picker shows your selected mode; it does not expose every model used behind an Auto response.

The habit to build: start with Auto, judge the completed result, and choose a named model intentionally.

Compare the credits used for successfully completed work, not just the price of a single model call.

Why Auto can still use more credits than you expect

A request can include research, file reading, tool calls and helper work before it produces the answer. Review those steps when a task uses more credits than expected.

If usage looks higher than the task warrants, check the files, context, tools and completed steps. The troubleshooting checks below help identify what drove it.

When You Want To Choose A Specific LLM: Understanding Model Multipliers

Every model in Hatz carries a multiplier — a number that tells you how much more or less credit-intensive that model is compared to the industry's flagship model, which sits at 1x.

  • A model at 0.5x has a lower displayed credit-rate comparison than the 1x reference. Compare completed task usage alongside the rate.

  • A model at 1x matches the flagship

  • A model at 2x has a higher displayed rate comparison. The amount of input, output and additional work still determines actual usage.

Use the multiplier for a quick comparison between models. Actual usage also depends on how many tokens are sent and received and the work needed to finish the task.

Where to find multipliers:

  • Hover over a model in the picker to see its displayed multiplier.

  • Open "View all models" and switch to the Multipliers tab for a sortable, filterable table of every available model

Keep an eye out for the Promo badge — Hatz occasionally runs promotional pricing on specific models, temporarily lowering their multiplier. The displayed number always reflects the current rate.

Before choosing a named model, check its multiplier. Use the Multipliers tab to compare available models, then judge value by the result and observed usage.

Use helpers for specialized work

Helpers, also called subagents, handle focused parts of a task such as research, file reading, calculations or document production. Hatz can combine their results into the conversation.

The practical result: tasks that would otherwise require the primary model to stretch outside its core strengths get routed to a tool designed for that job. That's better output and more efficient credit use.

Describe the result you want; Hatz can delegate suitable parts to available helpers. Your workspace access and the task’s permitted tools still apply.

Start Fresh Chats For New Topics

This is one of the most underappreciated credit drivers, and one of the easiest habits to change.

A conversation can carry relevant earlier messages and summaries into later work. Long, unrelated discussions can add context without helping the next task. Start a fresh chat when the topic changes.

The rule of thumb: when you're starting a genuinely new topic, start a new chat. The model has a cleaner signal to work from, and usage is easier to reason about because the chat is no longer carrying unrelated history.

Keep a thread when continuity helps—for example, editing a document or building on a result. Use a fresh chat for a separate topic.

Turn Repetitive Chats Into Automations

A weekly status summary, recurring report or standard client email is a good candidate for a Skill or automation. Save the instructions once and give each run the inputs it needs.

Manual, prompt-from-scratch conversations are the least efficient way to use AI for work that repeats. Every time someone reconstructs context, re-explains the format, or re-types the same instructions, that's both their time and your credits being spent on setup rather than output.

Skills keep a reusable procedure consistent. Automations run repeatable work. Keep shared project context in a Space.

Measure repeatable work by the time saved and consistency of its results. For example, a well-designed automation that produces the same report every week can save the team from rebuilding instructions and context each time.

Not sure where to start? Describe the result in Chat.

Ask Chat to save a reusable procedure as a Skill or create an automation. Hatz MCP also helps you find, inspect, create, update and run supported classic Workshop items.

  • “Turn the procedure we agreed in this chat into a Skill.”

  • “Create an automation that takes a weekly report and writes an executive summary.”

  • “Help me save a reusable procedure for drafting client follow-up emails from meeting notes.”

  • "Add a step to my existing workflow that translates the output into Spanish."

For classic items, ask Hatz MCP to inspect the inputs and steps before changing or running the item. For a new automation, use Chat’s automation tools and review the proposed setup.

Be Intentional About Files And Context

Choose files and excerpts that support the task. For a question about two pages of a 40-page document, identify those pages so Hatz can focus on the relevant material.

A few habits that make a difference:

  • Trim file uploads to the sections that are actually relevant to your question

  • Paste targeted excerpts rather than full documents when you're asking about something specific

  • Reuse appropriate context. Keep recurring project material in a Space, and identify the specific files needed for each task.

  • Keep questions focused. State the outcome, relevant inputs and constraints so Hatz can work without unnecessary back-and-forth.

Think in ROI, Not Just Spend

Judge spend alongside the value produced. For example, a 50-credit conversation that saves two hours of research may be an excellent trade for your team. Use that kind of comparison when deciding what work to expand.

Instead of asking, "how do I use fewer credits?" Ask: "am I getting real value from the credits I'm using?"

High usage driven by AI doing meaningful work is the system working as intended.

High usage from unrelated context, an unnecessarily expensive model or repeatedly rebuilding the same instructions is worth reviewing. Focus on the value produced per credit.

The tips in this article aren't about cutting AI use. They're about making sure that when you spend, it counts.

Troubleshooting Unexpected Or Runaway Credit Burn

Most high usage comes from the dials above — model choice, long threads, large files, and tool or subagent activity. Occasionally a single chat or run consumes far more than the work seems to justify. Use these steps to figure out what happened.

A chat spins with no response but usage still climbs

A request keeps processing in the background even if the answer is slow to appear, so credits can be consumed before you see any output. If a chat appears stuck:

  • Use the available stop control and inspect the task status. Closing a browser tab is not confirmation that server work stopped. Avoid resending the same request while its outcome is unknown.

  • Check completed results and external changes first. Once you know what remains, continue with a focused request and only the inputs needed for that work.

  • If the request was long-running, the assistant may stop the work on its own once it reaches a usage boundary and return a partial or wrapped-up result.

Repetitive or “junk” output loops

Sometimes the assistant gets stuck repeating itself or calling the same tool over and over without making progress, which burns credits while producing little of value. The system is designed to detect this kind of loop and wrap the run up gracefully rather than let it run indefinitely, but a loop can still consume credits before it stops. If you see repetitive or low-value output:

  • Stop the run, inspect completed work, then scope the remaining task. If a send or change has an uncertain result, establish its outcome before repeating it.

  • Narrow the request — one clear task, only the files and context that matter.

  • Note the time and the chat so support can review what the run was doing.

Auditing unexpectedly high usage for a specific model

If usage for one model looks higher than expected, work through it in order:

  • Confirm whether you selected Auto or a named model. Auto does not expose every underlying routing choice. Compare the usage report for the same task, workspace and period.

  • Check the displayed multiplier when comparing named models. A higher multiplier and more input or output can increase usage; task completion is the useful comparison.

  • Look at the relevant conversation context, large files and repeated file processing. Check whether the task could use a smaller, more focused input.

  • Check whether tools, subagents, or many steps were involved for that model, since each adds credits beyond the base message.

  • Review who and where. Admins can use the per-user usage views and the Usage Dashboard to see which users and features are driving usage for that model.

If usage still looks wrong after these checks, contact support with the tenant, the affected user, the date, time, and timezone, the model and Auto or named-model selection, and a link or ID for the chat, workflow run, or request so it can be investigated.

For Admins: Controls That Help At Scale

If you manage Hatz for a team or organization, there are additional levers available at the admin level that compound the habits above.

  • Set role-based credit limits to establish a ceiling for each assigned user. The role’s limit is not a shared group balance.

  • Review model access for roles and enable the models their work needs.

  • Use Auto as the everyday starting point and review the choices visible to an affected user.

  • Monitor usage patterns in the Usage Dashboard to identify where credits are going, who the heaviest users are, and whether usage patterns are producing the outcomes you expect

Visibility and policy together are how you stay in control at scale.

Did this answer your question?