Artificial intelligence has moved very quickly from something agencies experimented with to something that is becoming part of everyday digital work.

At Dancing Badger, that has prompted a slightly different question from simply asking whether we should be using AI.

Where does it genuinely make the work better?

And, just as importantly, where does experience, judgement and human input still matter more?

Those questions are shaping how we use AI across research, analysis, content, technical work and our wider approach to digital growth strategy.

That distinction matters.

Digital agencies have always used technology to make their work more efficient.

Analytics platforms collect millions of data points that would be impossible to process manually. Advertising platforms automate bidding and audience selection. Ecommerce platforms manage transactions, stock and customer journeys at scale.

AI is another significant step in that progression.

But it also introduces something different.

It can summarise, interpret, draft, compare, organise and suggest.

That makes it extremely useful.

It also makes knowing when not to accept the first answer increasingly important.

We started with the work that takes time

Some agency tasks are valuable because of the thinking involved.

Others are simply necessary stages you have to complete before you can reach the interesting part.

Gathering information from several sources.

Organising a large volume of notes.

Comparing different datasets.

Producing an initial structure for a piece of content.

Working through repetitive technical problems.

Creating the first version of something that will ultimately need refining.

These are areas where AI can be particularly useful.

Not because the work suddenly becomes unimportant, but because reducing the manual effort involved gives specialists more time to concentrate on what happens next.

Research can become faster without becoming automatic

Research is one of the clearest examples.

Whether we are developing a search strategy, reviewing a market, planning a website or exploring opportunities for a client, there is normally a significant amount of information to work through.

AI can help organise that process.

It can help summarise material, group related themes, identify questions that need answering and explore alternative interpretations.

That can make the initial stage substantially more efficient.

But research isn’t finished when information has been collected.

Someone still needs to decide which sources deserve confidence, which findings actually matter and what is relevant to the problem being solved.

A technically correct observation may have almost no commercial importance.

Another apparently small detail may completely change the recommendation.

Context is what separates collecting information from understanding it.

AI can help us interrogate data more quickly

The same principle applies to data analysis.

Modern digital businesses generate an enormous amount of information.

Website behaviour. Search performance. Advertising data. Ecommerce transactions. Customer activity. Conversion rates. Geographic patterns. Changes over time.

AI gives us another way to question and interrogate that information.

It can help identify unusual changes, compare periods, group information and bring patterns to the surface more quickly.

That complements the work we already do through DB Analytics and our wider analytics work.

But there is an important distinction between identifying something and understanding why it matters.

If a conversion rate changes, AI can help highlight it.

Understanding whether that change relates to audience quality, seasonality, website behaviour, product mix, campaign activity or something happening elsewhere in the business requires a wider view.

Content is one of the easiest places to use AI — and one of the easiest places to get it wrong

For many businesses, content generation was their first meaningful encounter with generative AI.

The appeal is obvious.

A tool can produce a blog structure, social post, email draft, product description or page outline in seconds.

Used properly, that can be genuinely valuable.

AI can help develop ideas, explore angles, organise research, challenge a first draft and adapt information into different formats.

But the ease with which content can be generated creates its own problem.

More content does not automatically mean better content.

Without sufficient input and editorial oversight, AI-generated material can become generic, repetitive, inaccurate or completely disconnected from what makes the organisation distinctive.

That matters whether the content is being created to support SEO and search visibility, an email campaign or the wider website experience.

The best input is often the material AI cannot invent.

Genuine client expertise. Real experience. First-hand knowledge. Customer conversations. Original data. A distinctive opinion. The reason something was done in a particular way.

AI can help us work with those ingredients.

It shouldn’t replace them.

Development is becoming AI-assisted too

AI is also increasingly useful in technical work.

It can help developers explore implementation approaches, examine code, troubleshoot problems, create documentation and accelerate repetitive tasks.

For an agency working across websites and ecommerce, that creates some obvious efficiencies.

But producing code and delivering a reliable digital product are very different things.

A developer still needs to understand how that code fits into the wider system.

They need to consider performance, maintainability, accessibility, security, integrations and what happens when something changes six months later.

This is particularly important when working with platforms and technologies supported by partners such as Automattic and Shopify.

AI can assist the developer.

The developer remains responsible for the result.

Strategy is where human judgement becomes more important

AI becomes particularly interesting when it moves from producing things to recommending things.

Give a capable system enough information and it can suggest dozens of possible actions.

Increase advertising.

Create new landing pages.

Improve organic visibility.

Change the customer journey.

Develop more content.

Introduce new automation.

Many of those recommendations may be entirely reasonable.

The challenge is deciding which one deserves priority.

That is where understanding the client becomes essential.

What is the business trying to achieve?

Which customers actually matter most?

What can the business afford?

What can its team realistically deliver?

Which opportunity could create the greatest commercial value?

And which apparently attractive idea should be left alone because there is something more important to do first?

Those are the conversations behind our Growth Strategy approach.

We don’t assume that an AI answer is the right answer

One of the risks created by generative AI is how convincing an incorrect answer can sound.

AI systems can misunderstand context, make unsupported assumptions, misinterpret incomplete information or confidently present something that simply isn’t true.

That means speed has to be accompanied by verification.

If something is factual, the source matters.

If something is based on data, the data needs checking.

If something represents a client’s brand, it needs to sound like the client.

And if something is going to influence a commercial decision, somebody needs to understand how that conclusion was reached.

AI is changing what valuable agency time looks like

Perhaps the most significant change is not what AI can produce.

It is what happens to the time that no longer needs to be spent producing it manually.

If gathering and organising information takes less time, more time can be spent interpreting it.

If a first draft appears more quickly, more time can be spent improving the thinking and making it distinctive.

If analysis can be explored faster, more questions can be asked of the data.

If repetitive technical tasks become easier, developers can focus more attention on solving the harder parts of a project.

This is the opportunity we find most interesting.

Not simply doing exactly the same work faster.

Using that efficiency to spend more time on work that creates greater value.

AI is becoming part of the workflow, rather than a separate activity

The longer-term change is likely to be less visible.

Instead of somebody deciding to “use AI” for a particular task, AI-assisted capabilities increasingly become part of the tools and processes already being used.

Research becomes easier to interrogate.

Reporting becomes easier to question.

Information can move between different stages of a project more efficiently.

Repetitive processes can become easier to manage.

This direction also connects with the thinking behind Growth Strategy and the way we are developing Growth Intelligence.

The objective isn’t to ask technology to make every decision.

It is to make the right information easier for people to work with when they make those decisions.

So what does that mean for our clients?

Ideally, very little of this should feel like technology for technology’s sake.

The benefit should appear in the work itself.

Research can move more quickly.

More information can be considered.

Data can be explored from more angles.

Repetitive tasks can consume less specialist time.

And our team can spend more of that time interpreting, questioning, improving and solving.

Clients still work with the people behind Dancing Badger.

Our specialists still make recommendations, challenge assumptions, develop ideas and take responsibility for the work.

They simply have access to better tools.

AI will undoubtedly continue to change the way digital agencies work.

The tools we use in twelve months may look very different from the tools available today.

What is less likely to change is what clients ultimately need from us.

Understand the problem.

Understand the opportunity.

Bring the right expertise to it.

Make a clear recommendation.

And be accountable for the result.

AI can help us do all of those things more effectively.

But it is the people behind the technology who still have to decide what good looks like.