Using AI Agents inside your Tekmatix workflows just became significantly more affordable.

A new round of performance and cost optimisations has been released across the AI Agent execution process, reducing the cost of running AI Agent Actions without requiring you to change your workflows, prompts or setup.

In internal testing, AI Agent executions are now approximately 40% to 50% cheaper, depending on the complexity of the agent.

And importantly, these savings have been achieved without reducing agent behaviour, reasoning quality or output quality.

What Has Changed With AI Agent Actions in Tekmatix?

Several behind-the-scenes improvements have been introduced to reduce the number of tokens and resources required each time an AI Agent runs.

Up to 50% Lower Cost Per AI Agent Execution

The biggest improvement is simple: AI Agent Actions can now cost considerably less to run.

Depending on the type and complexity of your agent, execution costs may be up to 50% lower than before.

This becomes particularly valuable when you have AI Agents running repeatedly inside high-volume automations.

Prompt Caching Reduces Repeated Token Usage

AI Agents often need to reference the same information across multiple stages of an execution.

Previously, some of that repeated context needed to be sent again during subsequent AI calls.

Prompt caching now allows repeated context to be reused instead.

This reduces unnecessary token consumption and makes multi-step AI Agent executions much more efficient.

Leaner System and Tool Prompts

The underlying instructions used by AI Agents have also been refined.

System prompts and tool instructions have been rewritten to be shorter, clearer and more precise.

This means fewer tokens need to be processed during every AI Agent execution while still giving the agent the instructions it needs to complete its task effectively.

Small reductions like this can create substantial savings when multiplied across hundreds or thousands of workflow executions.

Supporting Tasks Now Use Right-Sized AI Models

Not every part of an AI Agent execution requires the same level of AI reasoning.

Supporting operations such as:

• Summarisation

• Structured data generation

• Memory creation

• Internal processing tasks

can now use lighter, more efficient AI models where appropriate.

More advanced reasoning continues to use the models required for the core AI Agent task.

This helps reduce unnecessary AI processing costs without compromising the quality of the agent's primary work.

Fewer Tokens Are Used Overall

Together, these improvements mean AI Agent Actions now consume fewer tokens during execution.

You are therefore benefiting from savings in two ways:

1. More efficient processing during individual AI calls

2. Fewer tokens required across the complete AI Agent execution

For businesses running AI automation at scale, those savings can add up quickly.

Why Cheaper AI Agent Actions Matter

AI Agent Actions allow you to build intelligent workflows that can interpret information, make decisions, work with tools and complete more complex tasks than traditional step-by-step automation.

Reducing the cost of running those agents means you can now consider using AI Agents in more parts of your business.

Instead of asking:

"Can I afford to use an AI Agent here?"

you can focus more on:

"Is an AI Agent the best solution for this workflow?"

That is an important shift.

Complex AI Agents Can See the Biggest Savings

The most noticeable reductions are likely to occur with larger and more complex agents.

For example, AI Agents that:

• Work through long conversations

• Use multiple tools

• Analyse substantial amounts of information

• Perform several reasoning steps

• Run across complex workflows

typically consume more tokens.

Because the new optimisations reduce repeated and unnecessary processing, these heavier AI Agent executions may see some of the largest savings.

High-Volume AI Workflows Become More Practical

Cost optimisation also makes AI Agents more practical for workflows that run frequently.

You might use AI Agent Actions to help with:

• Lead qualification

• Customer enquiry processing

• Contact research

• Conversation analysis

• Data extraction

• Internal task routing

• Follow-up decisions

• Summarising customer information

• Generating personalised responses

• Updating CRM information

• Processing form or survey submissions

• Supporting complex automated workflows

When an automation runs hundreds or thousands of times, reducing the cost of every individual AI execution can make a substantial difference.

Easier Entry Point for Businesses New to AI Agents

If you have been curious about experimenting with AI Agents but concerned about usage costs, these improvements also reduce the barrier to getting started.

You can test more use cases, build more sophisticated agents and explore where AI Agents could remove repetitive work from your business without the same level of execution cost.

Do You Need to Change Your Existing AI Agent Workflows?

No.

There is nothing you need to update, migrate or reconfigure.

The cost optimisations are already live and automatically apply to both new and existing AI Agent Actions.

Your existing agents can continue running as they already do while benefiting from the new efficiency improvements behind the scenes.

How to Use AI Agent Actions in Tekmatix

To add an AI Agent Action to a workflow:

1. Open Tekmatix and go to Automations.

2. Open an existing workflow or create a new one.

3. Add a new workflow action.

4. Select the AI Agent Action.

5. Configure the instructions, tools and information your agent needs.

6. Test your agent to make sure it performs the task as expected.

7. Publish or activate your workflow.

Once active, the AI Agent can perform its assigned task whenever contacts reach that point in your automation.

Build Smarter Automations Without Building Longer Workflows

One of the biggest advantages of AI Agents is their ability to handle situations that would otherwise require large numbers of workflow branches, conditions and individual automation steps.

Rather than creating an enormous workflow to account for every possible scenario, an AI Agent can often analyse the situation and determine the appropriate next action dynamically.

With AI Agent executions now costing up to 50% less in internal testing, using intelligent automation at greater scale has become considerably more practical.

The improvements are already live inside Tekmatix, so if you already use AI Agent Actions, you can start benefiting from the reduced execution costs automatically.