AI Agents · Fact sheet
Complexica Automations fact sheet
The facts about Complexica’s AI agents in one place: what they do, where they run, how they are controlled and how a project starts. Each answer links to the page with the full detail.
Last updated
Key facts
| Offering | Complexica Automations, the agentic AI business unit of Complexica Pty Ltd |
|---|---|
| Launched | 30 October 2026 |
| What it does | Designs, builds and operates AI agents that carry out multi-step business processes inside customers’ systems, with their people approving the steps that matter |
| Processes | 27 process patterns in six groups, from order intake to legacy application modernisation |
| Industries | Eight sectors, from CPG manufacturing and retail to ports and logistics, pharmaceuticals and financial services |
| AI models | Any frontier model; model-agnostic, so customers are not tied to one AI provider |
| Where agents run | Inside the customer’s IT or hosted by Complexica; either way, under the customer’s permissions |
| Human approval | Approval points agreed for each process; high-risk or commercial steps wait for an authorised person |
| Audit trail | Every run logged for audit: what the agent did and why, and the work it produced |
| Certification | ISO 27001:2022, issued by TQCSI International |
| Engagement | Four stages: Discovery Workshop, scope and design, build and validate, operate and improve |
| First step | A Discovery Workshop |
| Research base | 25+ years of applied AI research; 300+ peer-reviewed articles and 25 books |
| Head office | West Lakes, South Australia; +61 8 8355 5050 |
On this page: What it is · What agents can do · Technology · Security and control · Working with Complexica · Proof and company
What it is
What is Complexica Automations?
Complexica Automations is the agentic AI business unit of Complexica Pty Ltd, launched on 30 October 2026. It designs, builds and operates AI agents that carry out multi-step business processes inside customers’ systems, with their people approving the steps that matter.
The agents take on high-volume work, such as orders, invoices, checks and requests, and run it faster, cheaper and more consistently.
More: Overview
How is an AI agent different from a chat assistant or RPA?
- A chat assistant answers. You decide every next step, and you still do the work: copy, paste, check, update the system.
- Rule-based tools such as robotic process automation (RPA) repeat the same fixed steps until the input or the process changes.
- An agent is given a goal. It plans the work, acts in your systems, checks the result and keeps going until the job is done, within limits you set. It asks a person before anything that matters.
More: Technology & governance
How is it different from chat tools, low-code tools and consultancies?
Chat assistants, low-code tools and generalist consultancies are quick to start on simple tasks. Complexica Automations works on complex agentic systems: governed, multi-step processes that combine code and control.
On Complexica’s market map this is the Gold Zone. Agentic design tools such as Claude, OpenAI Frontier, Gemini Enterprise, n8n, crewAI and Azure AI Foundry, and the large consultancies, sit in the low-complexity zone. Large-scale systems such as Complexica Decision Cloud and Palantir Foundry sit across the top.
More: Overview
What agents can do
Which processes can AI agents run?
27 process patterns, in six groups. Work can arrive as email, PDF, photo, spreadsheet or ticket; the agent does the steps in between, and a person approves before anything is posted.
Orders, quotes and invoices
- Order intake: reads orders that arrive by email, PDF, photo or handwritten note, matches each line to your products, checks stock and price, and drafts the order. A person confirms before it posts.
- Quotes from customer requests: reads requests with loose descriptions or competitor part numbers, applies the customer’s pricing, suggests substitutes and drafts the quote for approval, then follows up quotes not yet answered.
- Supplier invoice matching: matches invoice lines to the purchase order and goods receipt, posts clean matches and sends price or quantity variances to accounts payable.
- Material lists to draft orders: reads emailed lists, spreadsheets and job-site photos, matches each line to your catalogue and builds a draft order for staff to approve.
Quality, compliance and documents
- Supplier certificate and specification checks: reads, classifies and extracts certificates of analysis and product information forms, validates them against approved specifications and routes exceptions to quality assurance.
- Regulation and labelling checks: monitors regulation by market and checks ingredients, labels and claims against it, with every finding linked to its source.
- Regulated and promotional document checks: checks draft materials against approved claims and your code of conduct, or assembles approved text into draft documents for reviewers to edit and approve.
- Complaints and non-conformance: classifies complaints, links them to products, batches and specs, and drafts responses and corrective actions for review.
- Contractor compliance: checks contractors’ safety papers and certificates before anyone starts work, and flags what is missing or expired.
Customer service and support
- Ticket diagnosis and resolution: investigates each ticket across code, data and logs, finds the root cause and drafts the reply. If a fix is needed, it is built and checked before release.
- Customer self-service: answers everyday questions and handles bookings and requests at any time, and hands sensitive cases to staff with a summary.
- Track-and-trace enquiries: answers “where is my freight?” from live tracking data, and opens a case for a person when it can’t resolve the query.
- Warranty, returns and parts: answers product questions, tracks orders, registers warranties, starts returns, finds compatible parts and books service.
Finance and commercial
- Retailer deductions recovery: gathers deductions from customer portals, proof-of-delivery records and the ERP, matches each claim to the trade agreement and drafts disputes for an analyst to approve.
- Contract and obligations register: builds one searchable repository of agreements with an obligations and renewals register, alerts, and answers that cite the source document.
- Approval drafting: takes a quote through to a checked approval and purchase order, ready for sign-off.
- Reporting and commentary: gathers the numbers, explains the drivers and drafts the commentary for review.
- Claims and loan document checks: classifies documents, extracts the figures and checks them against policy before a person makes the decision.
Operations and supply chain
- Operations monitoring: tracks throughput, downtime, quality and waste from machine data, and flags what needs attention.
- Market and disruption watch: watches commodities, currency, weather and freight, and sends early alerts when something changes.
- Bookings and appointments: books and reschedules pick-up, delivery and dock times, follows up drivers and sends exceptions to operations.
- Process mining: shows how work really moves through your systems, and where it stalls.
Knowledge, systems and IT
- Ask the business: gives plain-English answers from approved data, with the sources attached.
- Living SOPs: captures procedures and know-how, keeps them searchable and keeps them current.
- Spreadsheets to governed apps: rebuilds high-risk spreadsheets as web apps with approvals, access control and an audit trail.
- Legacy application modernisation: maintains and progressively modernises an in-house system, with an engineer approving every release.
- Website and content changes: makes changes to a website from a single message, and checks every affected page before and after it goes live.
More: Use cases
Which industries does Complexica Automations work in?
Eight sectors. These are the agent use cases seen most in each.
| Sector | Most common agent use cases |
|---|---|
| CPG manufacturers | Deductions recovery; order intake; field sales preparation |
| CPG retailers | Basket-building assistant; supplier negotiation; supplier and retail media support |
| Wholesale and distribution | Order intake; quotes from customer requests; supplier invoice matching |
| Ports and logistics | Quote and order intake; bookings and appointments; track-and-trace enquiries |
| Construction and industrial | Material lists to draft orders; quote drafting and follow-up; ordering by message |
| Pharmaceuticals | Pre-call planning; regulated document drafting; promotional material pre-checks |
| Financial services | Claims triage; loan document checks; service triage |
| Consumer goods | Customer service, warranty and parts; retailer deductions; product listing checks |
More: Use cases
What makes a process a good fit for an AI agent?
The best first agents take on processes that simple tools can’t. A process fits when it has these four things:
- Many steps across systems: the work moves between your ERP, CRM, inbox and shared files.
- Rules to check against: specs, contracts, prices or policy decide what is right.
- Exceptions that need judgement: and a full audit trail of who decided what.
- A need for controlled access: under your permissions, with nothing it can reach that you haven’t granted.
More: Use cases
Which AI agents has Complexica built?
Every agent runs on Complexica’s own work before it is built for anyone else.
| Agent | What it does |
|---|---|
| Support Agent | Diagnoses tickets and drafts replies for an engineer, and hands approved fixes to the Dev Agent. An engineer approves each step. |
| Dev Agent | Turns approved fixes into code. Never publishes itself. |
| QA Agent | Checks every fix before anything goes live, then asks an engineer to review before release. |
| Customer Support Agent | Diagnoses and resolves customer issues. |
| Cyber Agent | Finds where you’re exposed and proves which vulnerabilities are real. An engineer approves each fix. |
| Legacy Management Agent | Updates and maintains legacy codebases, and turns old desktop systems into modern web apps. |
| Legal Agent | Drafts and reviews legal contracts, with every finding linked to its source. |
| Commercial Agent | Drafts and reviews scope proposals. |
| Website Agent | Makes changes to complexica.com from a single Slack message, and checks each affected page before and after it goes live. |
| Demo Design Agent | Makes and checks changes to demos. |
More: Overview
Technology
Which AI models does Complexica Automations use?
Any frontier model. Complexica stays model-agnostic and matches the model to each task, balancing accuracy against running cost. Customers are not tied to one AI provider, and an agent can move to a better model without rebuilding its controls.
More: Technology & governance
What is an AI agent made of?
Agent = Model + Harness + Enterprise.
- Model, the reasoning: reads context, plans and chooses the next step.
- Harness, the system around the model: connects to your systems, remembers state and runs multi-step work. It tests its work, rolls back and stops.
- Enterprise, your standards: permissions, audit, approvals and change control, wherever the agent runs.
More: Technology & governance
Which systems does an agent connect to?
The systems you already run: ERP, CRM, inbox and shared files. Agents also fit the tools your teams use, such as Slack and ticketing, with no need to replace the systems you rely on.
More: Technology & governance
Can agents work with legacy systems and spreadsheets?
Yes. Agents maintain and progressively modernise in-house systems, with an engineer approving every release, and rebuild high-risk spreadsheets as web apps with approvals, access control and an audit trail. Complexica’s Legacy Management Agent turned a car-distribution optimisation system from a desktop application into a modern web app.
More: Use cases
Do agents improve over time?
Yes. When someone says what was wrong, and why, a scheduled review turns it into memory. The next similar job starts from it.
More: How it works
Security and control
Where does an agent run, and what can it access?
Either inside your IT or hosted by Complexica. Wherever it runs, it can only reach the systems and data you grant, and every step is logged, so you can audit what it did and why. Each agent has a defined identity and controlled access to secrets, and permissions are agreed for each process under your policies.
More: Technology & governance
Who approves what an agent does?
You decide. Approval points are agreed for each process during design, and can change as trust grows. Routine steps run on their own; high-risk or commercial steps wait for an authorised person. People handle exceptions and decisions; agents handle the repetitive work in between.
More: Technology & governance
Is there an audit trail?
Yes. Wherever an agent runs, you can audit every run: the logs show what the agent did and why, and the work it produced. Each run is recorded at three levels:
- Summary: what was found, what was done and who approved it.
- Every step: each action in order: what it looked up, checked and decided, and why.
- Evidence: the records, queries and logs behind each conclusion.
Conclusions link back to the document, record or data they came from, and every change records who approved it. Every release is versioned and evaluated before it goes live.
More: Technology & governance
Is Complexica certified?
Yes. Complexica is certified to ISO 27001:2022; the certification was issued by TQCSI International. Every agent deployment is governed by the same information security management system that governs Complexica’s Decision Cloud®.
Working with Complexica
How does an AI agent project run?
Four stages with one team. At every gate, you decide whether to go further.
| Stage | What happens | What you get | Your decision |
|---|---|---|---|
| 1. Discovery Workshop | A working session with the people who run your processes | A shortlist of processes, and what each could be worth | Which processes to take further |
| 2. Scope and design | We map the process and design the agent around it | The agent design, its controls, and build and running costs | Whether to build, and in what order |
| 3. Build and validate | Built on our control layer and connected to your systems | A tested agent, versioned and evaluated before it goes live | When it goes live |
| 4. Operate and improve | We run it and review its work | Traces, usage and alerts, and an agent that keeps improving | Which process comes next |
More: How it works
What is the first step?
A Discovery Workshop: a working session with the people who run your processes. You leave knowing where agents fit, what they are worth and the next steps.
- Who takes part: the people who own and run the processes you want to improve, and someone who knows the systems involved. From Complexica, the team that builds and runs its agents.
- What to bring: the processes that take the most time, cause the most rework or hold up the business. Examples of the documents, emails or spreadsheets involved help get specific.
- What comes next: a clear view of where agents could help and what they could be worth. If it makes sense to go further, the next step is scoping and design for the processes you choose.
What does an AI agent cost?
Each agent is designed with its value and cost set out up front. At the scope and design stage you receive the agent design, its controls, and its build and running costs, before you decide whether to build.
More: How it works
What does Complexica need from us?
- A project lead: one person who keeps decisions moving and brings in the right people.
- Process experts: the people who do the work today, for short, regular sessions on the rules and the exceptions.
- An IT contact: someone who can arrange access to the systems and data the agent needs, under your policies.
More: How it works
Who runs the agent after go-live?
Complexica does, with you. It operates each agent, reports on its work, and improves it from every reviewed job. You get traces, usage and alerts on every run, and exceptions are routed to your people.
More: How it works
Proof and company
Who uses Complexica’s AI agents today?
Complexica itself: every agent it offers has run on its own work first. Three examples:
- Support: a chain of agents. The Support Agent investigates a ticket read-only and posts its findings. When an engineer says yes to a code change, the Dev Agent builds it and the QA Agent checks it, then asks for review.
- Legal: the Legal Agent checks signed agreements, redlines and emails, and reviews a draft against Complexica’s standard positions to return a memo and a redline. Every finding links to its source, and the authorised reviewer decides.
- Website: a change to complexica.com is one Slack message. The agent edits the site, checks every affected page before and after it goes live, and replies in the thread.
At launch on 30 October 2026, Complexica Automations was in discovery with customers on knowledge and answers, operations monitoring, market and disruption watch, process mining and contractor compliance.
More: Use cases · Launch announcement
Who are Complexica’s software customers?
Complexica’s software customers include Lion, Asahi Beverages, Metcash, Bunzl, Pfizer, PFD Food Services, Boehringer Ingelheim, Costa Group, Aurizon, Toll Group, Linfox and Viterra.
More: Overview
Who leads Complexica Automations?
The team that designed, built and runs the agents inside Complexica.
- Spiros Kaye, Head of AI Automation, leads the technical side, taking frontier AI research through to working systems.
- Adam Michalewicz, Manager, AI Automation, leads the commercial side, helping customers identify where automation can deliver the greatest value and scoping those engagements.
More: How it works
What research is Complexica Automations built on?
More than 25 years of applied AI research at Complexica. Complexica’s staff have published more than 300 peer-reviewed articles and 25 books, and built Decision Cloud®. International thought leaders in artificial intelligence advise Complexica through its Scientific Advisory Board, and its university research partners are the University of Adelaide, RMIT University and the Polish-Japanese Academy of Information Technology.
More: About Complexica
Where is Complexica based?
Complexica Pty Ltd has its head office in West Lakes, South Australia. Phone +61 8 8355 5050.
More: Contact
Discovery Workshop
Find the processes worth automating
A working session with the people who run your processes. You leave knowing where agents fit, what they are worth and the next steps.
Some of our Recent Awards
Copyright © 2026 Complexica Pty Ltd
