AI and its effects on digital agencies and consultancies

With Google’s presentation of AI Mode, last week I talked about the effect this particular AI — far more than the others — is already having on digital businesses:: we are facing a revolution unlike anything we have seen before..

That was the first article on the subject.

This is the second.

Here I am going to focus not on digital businesses, but on the providers of these digital services: agencies, consultancies and freelancers.

Because the truth is that, for many—if not all—what is coming will mean a significant change in our day-to-day work.

I am going to break it down area by digital area, explaining where each one stands right now.

SEO agencies

This is the most obvious case, because it is the area most affected.

SEO is going to change completely. First with the gradual rollout of AI Overviews (AIO from now on) and, soon, AI Mode (AIM). Writing to rank your blue link in the top ten—first if possible—is over.

It is over because we have already seen that, even if you manage it, the organic traffic you receive will be a third, or less, of what you used to get.

Therefore, ranking metrics matter much less. It is time to look for alternatives.

Perhaps the most obvious metrics in this case will be appearances and mentions in AI, which will complement traditional rankings.

The next thing people will look at will, of course, be traffic. Specifically, the sum of SEO traffic + AI traffic, which, however much you add them together, will still be lower than what Google used to bring us.

And because it will be lower, attention will shift to the third and most important current metric. The real KPI: conversions or revenue, which is what this is all about.

If, as an agency, we can show that the SEO work we are doing helps sell more and is profitable for the client, there will be no problem.

But we have to prove it.

And this is where things get complicated, because direct conversions from AI traffic are easy to attribute, but what about indirect ones? What happens to them?

As professionals, we have to design an analytics system that proves beyond doubt that the content we are creating and positioning helps sell more.

And in many cases that will not be easy, in my view.

Will it be difficult because we are not actually providing any value to the client?

No.

It will be difficult because, in a world where tracking is becoming increasingly complicated, the only valid way to prove it would be to stop doing what we are doing and measure.

But because SEO effects last, we do not even have that option, whereas we do in paid media.

So unless we can develop an analytics system with indisputable results, in many cases continuing to hire these services will be an act of faith, vision and trust on the client’s part.

A bet, really.

And I think we all understand that small clients, where tactics tend to take precedence over strategy, prefer certainties to bets.

So if your clients are like this, you had better prepare a convincing case.

By contrast, for large clients with resources, things should not change as much. Even if the tangible and, above all, demonstrable financial results—that is the key to all this—are worse, they cannot afford to do nothing and let the competition overtake them.

In fact, we have previous experience with similar cases, because upheavals like this have already happened in other channels.

I specifically remember how 12 or 13 years ago Facebook’s algorithm changed completely. At TiendAnimal, from one day to the next, this source started bringing the website a third of the traffic it used to. And it kept declining progressively.

And that was despite having two profiles with more than 130,000 users in total, which was huge at the time and far above the competition.

Well, despite the algorithm change, the company obviously did not collapse.

Nor did it stop publishing on Facebook.

The strategy was simply rethought and resources were rebalanced.

With the arrival of Google’s AI, I think something similar will happen, at least for now. As I said, probably in large accounts. We will have to see what happens with small ones.

Content marketing agencies

Continuing along the SEO line, the next step is to talk about content agencies, which often go hand in hand with it.

As in the other areas, the arrival of AI is making the very good agencies sweat, and their revenue from these services will fall somewhat.

Average agencies will either improve and become good ones, or they will have a hard time.

And the bad ones… well, you know.

Because AI has almost completely removed the barriers to entry for creating all kinds of content: with a little know-how and a premium Freepik plan, any sharp person can produce amazing material.

That is on the audiovisual side. If we are talking about text, ChatGPT, Gemini and Claude generate pieces that are good enough for many clients, who see no need to pay for something that, in their eyes, is no better than what AI produces.

But that is the pure execution side. Fortunately, freelancers and agencies that sell content creation still have something solid to hold on to.

We are talking about the strategic side. And it is essential, because even if AI generates the piece, if you do not know which pieces to create, you are no better off.

It is here, in strategy rather than execution, that professionals will still be able to add value.

Make that clear.

And, by the way, AI does not only bring disadvantages: if your agency has already incorporated it into its workflow, you will have noticed the increase in the speed of producing deliverables. Now maintaining high quality generally entails lower operating costs and is a combination of judgement and strategy.

How do you see yourself in this scenario?

Paid media agencies

Here I include both social media and Google agencies, because both channels are heading in the same direction: both platforms want to automate as much as possible with AI.

In other words, what both companies want is to replace a large share of the services these agencies provide, making it easier for advertisers to create campaigns.

The approach is simple: they ask the advertiser three things:

1. What they want to advertise (website, products or services).
2. What objective they have (sales, leads, brand awareness).
3. How much budget they want to invest.

And the platform’s artificial intelligence will take care of everything else:

• Carry out the full campaign setup.
• Create the complete ads: text, images, video and audio.
• Show the campaigns to the users most likely to convert.

Put like that, it sounds very good for them and for the advertiser—not for the agency, of course. But we still need to see how effective it really is and compare it with the expert team’s performance.

And yes, for small advertisers with limited budgets, it may make perfect sense, because even if it converts slightly worse than a manually managed campaign, you save the setup and maintenance costs, so in terms of real profitability it may work for you.

But I very much doubt that, in the short term, major advertisers investing six or seven figures a month on these platforms will take the leap of faith involved in removing human intervention.

If there is no solid proof that AI works better, they certainly will not. And even when there is proof, we will see. The initial risk and uncertainty are obvious.

Could it happen in the future?

Yes, it could.

In fact, I would not be surprised if the process evolved as follows:

The advertiser hires an agency, and the agency in turn uses AI to execute the entire process.

In other words, AI ultimately performs the work, but with one important nuance: throughout this process there will always be human oversight and accountability behind it.

And once it is clear that AI can adjust itself properly, paid media agencies will indeed have little left to contribute.

But as with media agencies—and all the others, really— the strategic side may keep paid media agencies going for a while. Especially those covering paid media as a whole rather than a single channel: it is essential to understand every channel where money can be invested, build the right channel mix and know the best practices that work within each one.

Even so, I think that when the time comes they will go through what SEO agencies are suffering now: they will be constantly under the sword of Damocles, with clients wondering after every invoice whether to eliminate that maintenance cost and entrust everything to the platform’s AI.

Analytics consultancies

They are at a very interesting point and I believe they have little to fear. At least there is no major short-term AI threat in sight. More likely a medium- or long-term one.

But be careful: this is moving so quickly that it could change overnight.

The task now is to measure the effect of AIO and AIM.

I mean the real effect: the euros in the account at the end of the month, not traffic or rankings.

And, as I mentioned when discussing SEO, it is not an easy job, determining how much revenue we are losing because of the emergence of Google’s AI.

Here, consultancies that find the most applicable model—knowing in advance that the effect differs in every project—will hold a winning card.

One that will make it perfectly clear why analytics should sit at the centre of every business.

Google’s AI itself will have little or no effect on these consultancies, but AI in general will.

Because, when used well, it allows or helps business managers themselves to carry out complex analyses.

I am also convinced that AI is far more capable than any human of finding hidden patterns in large datasets.

That is why I do think the analyst’s job will change in the medium term. In fact, I do not think analysts will perform the bulk of the analysis themselves; I imagine a process something like this:

1. The analyst defines the KPIs and how they are measured.
2. The technical team implements them.
3. The analyst reviews the implementation and ensures data quality.
4. AI uses the collected data and performs a threefold analysis: descriptive, diagnostic and predictive.
5. The analyst gathers these insights and presents the prescriptive analysis. In other words, they build the complete narrative and explain which actions should be taken.

As I said, in step 4—the one with the heaviest analytical workload— AI will be able to outperform a standard analyst. So, if you consider yourself one, I would recommend strengthening the other three points:

#1. Understand the business in order to set KPIs.

#3. Know how to audit both the data itself and the way it is collected.

#5. Understand the business and propose measures that improve results.

In short: business strategy and data understanding.

Of course, consultancies will also need to know how to handle AI so that it performs its role, but I would say that is more of a technical skill than an analytical one. And as AI advances and becomes easier to use, that skill will become less valuable.

As I said at the start of this section, I think this is a good time to be an analyst—and prove it.

Digital business consultancies

Changes.

I am seeing, suffering and enjoying a great many changes here as AI develops.

Some work in favour of consultancies—and freelance consultants—and others work against them.

In their favour:

• Perplexity, Deep Search and similar tools make it possible to carry out in-depth analyses of a huge range of topics while using fewer people and less time. Market and competitor analyses that are sufficient to establish the context for almost any digital business. That said, these studies need to be reviewed and refined; it is not as simple as waiting for the machine to deliver 100% of the work perfectly on the first try. But the cost savings are astonishing.
• Internal consultancy processes such as documentation, transcription and meeting summaries make coordination between teams enormously easier. It is now possible to keep meeting attendance to a minimum while giving everyone else an immediate summary, transcript and the precise video section they need to watch for more detail. Meetings are now for listening and contributing, rather than taking notes.
• Creating GPTs as knowledge hubs: if they are given the right documentation, these GPTs will complement and streamline extremely important consultancy processes such as:
  • Employee onboarding.
  • El desarrollo de propuestas comerciales (incluyendo pricing).
  • El desarrollo de las tareas en sí.
• Decision-making.
• Finally, AI is a new line of business. In other words, implementing AI in companies, or training them to do so. There is obviously a great deal of business here in the short and medium term. Further ahead, given the extraordinary pace at which this field is evolving, we will see.

Against them:

• The barriers to entry for business consulting work have fallen. This means clients may think they can carry out these analyses or studies themselves without needing a consultant. Here again, experience and the strategic side are what consultants must hold on to.
• If AI evolves to the point where it can devise a marketing, financial or strategic plan for the trivial price of, say, €250 a month—roughly what the most premium ChatGPT and Google Gemini plans cost right now—it could mean that only the largest consultancies survive. As with agencies, these multimillion-euro projects will require human supervision and experience. Until proven otherwise.

In short, I think that if you are an online business consultancy, you will have to:

• Bring AI into your operations as much as possible—starting yesterday.
• Sell that know-how adapted to companies and position yourself as an expert in the field.

UX and CRO consultancies: will we design for AIs?

At first glance, they seem the best positioned, because in my experience AI’s attempts to become relevant in this digital field lag behind the others.

Yes, AI can generate interfaces relatively quickly, but without an understanding of how users’ minds work when converting, what it suggests may work… or it may not.

And that is inherent in the current artificial intelligence model: for now, AIs do not understand. They only return the most probable answer.

I can certainly see them as assistants: they speed up the entire digital product creation process, but always with an expert behind them.

That said, my question is whether, before too long, the web experience as a whole will change completely. Whether today’s link-based browsing model will become a more conversational one.

Whether, instead of using Amazon’s search engine, we will ask it directly and converse with it.

It does not sound crazy: to some extent, social media and infinite scrolling already redefined how we used the internet.

What I am certain of is that the code beneath digital products will change to adapt more closely to what AI wants. It is not clear whether this will fall to SEO, the web developer or, perhaps in a few years, the AI Experience Manager.

Jakob Nielsen’s position —the father of the concept of web usability—seems to point in this direction.

And I am not going to question the vision of someone who was working on and defining the relationship between humans and machines long before anyone else.

Software development consultancies

Here, as with analytics consultancies, I have my doubts about whether AI helps or harms them.

On the one hand, in the short term AI is increasing their profitability: they are producing code much faster and with fewer people than before.

That is a fact.

For now, digital businesses—not a little personal side project—still need developers who know what they are doing, understand security and performance, and can see the project as a whole rather than as isolated features.

Will AI eventually replace them completely?

Well, perhaps in the future it will, but I think that will take longer than replacing paid media agencies, for example.

Moreover, given their technological background, making the leap into corporate AI implementations —business consulting focused on AI—seems logical and even achievable. At least, that is how I see it.

The human team

That is the current state of affairs across the different types of digital agencies and consultancies.

Each one faces a different level of uncertainty and AI penetration in its services.

However, they all have something in common. Something related to their internal teams which, from what I have seen, should concern you if you are finishing your studies and plan to focus on any of these areas.

I say this because AI is causing fewer junior profiles to be hired. Why would I invest in training them when AI is already trained, works 24/7, never gets sick and costs next to nothing by comparison?

And that is precisely the question you need to ask yourself if you are under 30 and have fewer than three or four years’ experience in any of the areas mentioned above.

I think you are in a tough spot and will have to work your arse off. Or take advantage of your youth and prepare to become an AI expert.

Whether you are an AI expert or merely use it as a tool in your current job, I think the junior profile most in demand will have these characteristics:

• All-rounder: someone capable of getting an acceptable mark in several fields. AI will execute, but you will give the instructions, backed by a strategy and a quality threshold.
• Closely tied to the business, rather than to one speciality—not a “SEO specialist” or a “content marketing expert”).
• AI power users and, to a lesser extent, automation experts. Thanks to AI, the volume of work that can be produced multiplies. Creating pieces, creative assets and ads will take minutes rather than hours, leaving the rest of the time for strategy and analysis.

That is how I see the outlook for digital professionals who are just starting out. At least within companies. I am seeing more and more young entrepreneurs, where the skills required are quite different.

Although both groups rely on AI.

Conclusion and closing thoughts

I do not know what you thought of this review of every digital area, all of which AI is already shaking up considerably.

In truth, I did not write it for you. I wrote it for myself. AI is changing the work I have been doing for 15 years far more than any previous disruption.

And I need to organise my thoughts.

So I can decide which direction I want to take as a digital business consultant.

Because even if everything remains more or less the same in the short term, I am absolutely certain that the landscape will change completely in the medium term.

In 12 months? In 24?

I do not know, but in my mind the first date is labelled more likely than the second.

If you work client-side—that is, at a company that hires agencies and consultancies— the short-term scenario will not change as much, or as quickly.

But it will for agencies and consultancies.

So there are really three things to think about:

1. The services that will be in demand.
2. The clients who will demand them.
3. And how to acquire the knowledge needed to provide them.

Because nobody wants to become the Nokia or BlackBerry of the post-iPhone era.

And I think that unless we adapt, many of us are heading exactly there.