AI-active
Current public capability, product, operating claim or delivery case.
A source-backed review of what 100 Australian agency websites publicly show about AI activity and agency-level integration.
AI activity can be visible without a public operating model behind it. The integration measure is intentionally narrower than the AI-active classification.
All figures come from a locked 100-agency register. The categories are evidence classifications—not maturity scores.
Current public capability, product, operating claim or delivery case.
Meaningful commentary or experimentation, but no current capability or operating claim.
No qualifying public AI evidence found after the required source coverage.
To qualify, one retained first-party source had to establish a repeatable workflow, operating approach, leadership owner, team enablement, safeguard or integrated proprietary system. An AI offer, article, product or case did not qualify on its own.
These results are not combined with the website audit. They are shown as context with their own samples and methods.
of 92 surveyed small marketing businesses reported currently using AI.
Separate April 2025 Australian small-business survey.Beginner / Emerging / Advanced among 60 senior Australian marketing leaders.
Separate readiness study spanning strategy, leadership, skills, governance, data and deployment.The audit used a purposive, quota-based frame of 100 distinct Australian agency entities. Agency type was the primary balancing dimension; geography was secondary. This provides breadth of observable public evidence, not statistical representation of every agency.
Seven categories prevent the result being driven by one form of agency.
| Category | n |
|---|---|
| PR / communications | 28 |
| Creative / advertising | 20 |
| Digital / performance / SEO | 12 |
| Corporate / reputation / public affairs | 10 |
| Brand / design | 10 |
| Social / content / influencer | 10 |
| Integrated / full-service | 10 |
The frame includes major centres and national/multi-city organisations; this is not a population-weighted geography.
| Location | n |
|---|---|
| Sydney / NSW | 42 |
| Melbourne / VIC | 36 |
| Brisbane / QLD | 8 |
| National / multi-city | 6 |
| Perth / WA | 4 |
| Adelaide / SA | 4 |
The audit was deliberately bounded at 100. Candidates needed to be active Australian agency entities with a distinct first-party website and sufficient accessible material for the coverage rules. Eligibility, accessibility and entity checks were completed before AI content was assessed. Non-inclusion does not mean an agency lacks AI capability or internal strategy; agencies outside the frame were not scored or ranked.
What this frame supports: exact findings about the 100 entities audited. What it does not support: a claim that 9% of all Australian agencies have an AI strategy or that the frame represents the whole market.
These are not rankings. Each row links to a retained first-party source and explains the narrow dimension that qualified.
| Agency | Qualifying dimensions | First-party source |
|---|---|---|
| DevotionCreative / advertising | Agency Operating Approach | Open first-party source ↗Evidence and decision“results. We harness the power of AI for the heavy lifting, so we have time to focus on everything that makes us human; strategy, creativity, and storytelling. Head of Growth and Digital Marketing Marketing Q&A with Connor Chmielweski What do you mean by AI-first? We believe that artificial intelligence is the future of marketing so we strive to intelligently automate as many tasks and processes as possible. We outsource repetitive data entry, campaign optimisations, data analysis and much more so our team can focus on the big picture. Do you work with start-ups? We love workin” Devotion describes an AI-first operating approach and using AI for repeatable heavy lifting so staff focus on human strategy and creativity. |
| First Page AustraliaDigital / performance / SEO | Repeatable Coordinated Workflow · Proprietary Operating System | Open first-party source ↗Evidence and decision“The First Page platform is powered by our proprietary technology - SENTR™. Built with you in mind, SENTR™ is the system that distinguishes First Page from other agencies. AI decision-making. SENTR is a proprietary tool that brings campaign management, reporting and approvals together in one platform. So we spend less time on admin and more time executing and optimising your campaigns.” The proprietary SENTR platform publicly combines campaign management, reporting and approvals and identifies AI decision-making as part of the system. |
| HypetapSocial / content / influencer | Repeatable Proprietary Workflow · Risk Safeguards | Open first-party source ↗Evidence and decision“ce Working with Hypetap. Our co-founders saw a better way to work with influencers. They called on their experience to design and build the Hypetap platform. Now, the benefits are available to your team. Designed for complete end-to-end campaign management and reporting, we use: AI Matching Exclusive AI technology to match your brand with aligned influencers. Profanity monitoring Alerts to flag controversial language and hate speech in posts and commentary. Conflict assessment Inbuilt conflict assessment tool to minimise campaign risk. Campaign dashboard R” Hypetap describes an end-to-end campaign platform using AI matching alongside brand-safety and conflict-assessment controls. |
| Paper MooseCreative / advertising | Repeatable Coordinated Workflow · Proprietary In-House System | Open first-party source ↗Evidence and decision“Moose Review uses marketing science and AI to assess and test creative. Moose Review combines emotions and marketing science to create a score using our proprietary methodology calibrated against other successful campaigns to help the team understand the likelihood of driving ROI. Moose Review runs reviews with digital twins. Synthetic humans or ‘synths’ as we call them internally have been proven to provide responses up to 87% similar to the real humans they are modelled on, speeding up creative review times by 100x.” Paper Moose describes a proprietary in-house AI creative-testing system with synthetic personas, a calibrated methodology and repeatable review workflow. |
| SenateSHJCorporate / reputation / public affairs | Repeatable Coordinated Workflow · Governance | Open first-party source ↗Evidence and decision“munity engagement – when communities are affected by projects, reforms or decisions. Stakeholder engagement – when influence, relationships and sector confidence matter. Deliberative engagement – when informed public input can strengthen the quality and legitimacy of a decision. AI-assisted engagement – when technology can help identify insight, test messages and analyse complex feedback with strong governance. Employee engagement – when internal trust, connection and alignment are critical to performance and change. If your project, reform or decision cou” SenateSHJ describes AI-assisted engagement and research processes used to identify insights and analyse feedback with strong governance. |
| TBWA AustraliaCreative / advertising | Leadership Ownership | Open first-party source ↗Evidence and decision“BWA\Melbourne Elly Liley Chief Strategy Officer, TBWA\Melbourne Blake Crosbie Group Managing Director, TBWA\Sydney & +61 Managing Director Paul Reardon Chief Creative Officer, TBWA\Melbourne Kiefer Casamore Executive General Manager, TBWA\Melbourne & Eleven PR Lucio Ribero Chief AI and Innovation Officer, TBWA\Australia Who we are TBWA is The Disruption Company®. We use creativity to help businesses acquire a greater share of the future. We are a local agency across Sydney and Melbourne with a strong global backing. That means we have the support and exper” TBWA Australia publicly identifies a Chief AI and Innovation Officer, satisfying the leadership-ownership criterion. |
| The Content DivisionSocial / content / influencer | Coordinated Workflow · Team Enablement · Governance/Client Safeguards · Client Disclosure | Open first-party source ↗Evidence and decision“The Content Division has developed an Artificial Intelligence (AI) policy which outlines how we use AI tools to assist in our service delivery, the tools we have approved to use internally, and the extreme importance of human oversight. This policy applies to all employees of The Content Division who use AI tools in their work. Inform clients when AI tools are used in their projects. All AI-generated content must be reviewed by a human before delivery to clients. Participate in ongoing training sessions on AI tools and best practices. Use AI tools that are integrated into The Content Division's workflows to maximise efficiency.” The published policy applies to all employees, defines approved tools, integrates AI into workflows, requires training and human review, and mandates client transparency. |
| ThinkerbellCreative / advertising | Agency-Wide Operating Model · Dedicated Ownership · Workflow Redesign | Open first-party source ↗Evidence and decision“pletely new way of working. 3. What’s your unfair advantage over the networks today? Our unfair advantage isn’t just a nicer mindset, it’s a completely restructured model that’s brand led and future fit. 4. How are you using AI to compete above your weight? We’ve operationalised AI across the business, from workflows to execution. We created a dedicated engineering and systems team to map out every department, from finance, earned, studio to creative. Their brief was to create AI tools from the ground up, to our model, to enhance what we do. We use it to s” Thinkerbell states it operationalised AI across the business and created a dedicated engineering and systems team to map every department and build internal tools. |
| Zib DigitalSocial / content / influencer | Repeatable Coordinated Workflow · Human Oversight | Open first-party source ↗Evidence and decision“wns your account, a content + ads team behind them, transparent monthly reporting in commercial language, and month-to-month terms. No juniors on rotation, no offshore teams you've never met. The human you talk to in week one is the human you talk to in year three. Are you using AI in social media work? + Yes, agents do the repetitive work (creative variant generation, audience research, copy drafts, reporting briefs). A senior human still does the judgement: what to test, what to kill, what your number tells us next. We've written about how this works on” Zib specifies repeatable tasks handled by agents and the senior-human decisions retained in its service workflow. |
The study supports exact counts in its final inclusion frame. It is not a probability sample of every Australian agency.
100 entity-unique agency websites. The final inclusion frame was built to provide breadth across seven agency segments.
Selection before topical inspection. Eligibility and accessibility screening were completed before AI content was assessed.
First-party evidence only. Every positive classification has a retained URL and excerpt. Client leakage, page chrome and generic innovation language were excluded.
Classification-aware coverage. No-signal classifications required homepage, capability/work surface and targeted AI-source coverage. Positives required direct retained evidence.
What it cannot tell us. The report does not establish private use, maturity, motives for silence or which agency is better at AI.
The practical work is deciding where AI belongs, who owns the standards, how people and models collaborate, where review sits, how client trust is protected and what value the agency can honestly claim.
Using AI occasionally is easy. Integrating it across workflows, people, safeguards and client value is the real strategic work.
That is why the useful response to this research is not simply to publish an AI service page. It is to build a credible operating approach, then communicate the proof that follows from it.
An AI Integration Strategy Sprint maps current use, identifies priority workflows, establishes ownership and safeguards, and creates a practical 90-day roadmap before public positioning gets ahead of the substance.
Read the methodology before acting →