The 2027 marketing budget: where should marketers focus their spend as AI changes discovery?

As we enter 2027, marketing budgeting is becoming harder than ever. Many of us remember when we could allocate a set amount to offline channels at the start of the year, have a reasonable idea of the return it would generate and repeat broadly the same process year after year.

Digital marketing changed that. Budgeting became more dynamic as performance became visible and spend could be moved between channels based on what was working. 

AI is accelerating that change again, as customers increasingly use tools such as ChatGPT and Google AI Overviews to discover products, research categories and compare providers before they ever reach a company’s website.

For marketing leaders, this creates a challenge. The customer journey is changing while visibility over that journey is becoming weaker. A customer may arrive on your website having already compared options, read reviews and built a shortlist elsewhere. What appears in your analytics as a branded search or direct visit may actually be the end of a much longer research journey.

For CMOs building their 2027 budgets, our recommendation is to avoid making a big upfront bet on exactly how AI will change their channel mix. Instead, the budget should be built around four areas: Understand, Protect, Prepare and Experiment.

1. Understand how your most valuable customers are behaving and where their purchase journeys are changing.

2. Protect the activity that continues to generate profitable growth, while being careful not to spend more simply to replace traffic or behaviours that have become less valuable.

3. Prepare for changes in discovery by ensuring your business can be found, understood and recommended as more research takes place through AI.

4. Experiment with emerging opportunities, with genuine budget available to test new routes to market before making larger commitments.

The balance between these areas will vary by business and should move during the year as the evidence changes. In this article, we explore exactly what that means for a CMO planning for the year ahead. 

1. Understand how your customers are changing

Before deciding how much budget should sit in search, social, content or emerging AI channels, marketers need to understand who their most valuable customers are and how they actually buy. That starts with a clear view of your ICPs, an understanding of the customers that generate the greatest long-term value and a proper understanding of what part of your marketing activity drives those sales.

CMOs should be asking the following:

• Which customers generate the greatest long-term value?

• How do those customers discover and research the category?

• What has their purchase journey historically looked like?

• Which parts of the website have traditionally played the biggest role before conversion?

• How has that that behaviour begun to change?

This matters because adoption of AI will not happen uniformly. A customer buying a low-risk, repeat purchase may already be comfortable asking an AI assistant to recommend a product or even go straight to a basket, while someone making a high-value or more emotional decision may use the same technology extensively for research but still want much more control over the final choice.

That same variation can exist within individual businesses. A retailer selling both large-ticket items and low-value consumables may see very different behaviour between the two. A B2B organisation may find procurement teams increasingly using LLMs to build an initial shortlist while senior decision-makers continue to rely heavily on referrals and direct relationships.

The growth in the use of ChatGPT for consumers has exploded over the past few years, the exposure to AI is no longer a ‘younger, more technologically advanced audience’ but almost everyone using the internet today. 

With this in mind, once you understand the customers you care most about, look at where their behaviour may already be changing. Many marketers are seeing changes in organic traffic and immediately wondering whether AI is responsible, but the headline number rarely tells you enough.

Marketing leaders should therefore ask:

• Which landing pages are gaining or losing traffic?

• Are changes concentrated in discovery content or high-intent pages?

• Are customers entering the website further through the funnel?

• Is branded search or direct traffic growing while generic discovery traffic falls?

• Are conversion rates changing by landing page?

  
  

For a retailer, a decline in traffic to blogs and buying guides is very different from a decline in product-detail pages. For a lead-generation business, a reduction in educational visits means something very different from fewer customers entering the sales funnel itself.

Increasingly, customers may be carrying out much of their early research away from the website. They can compare providers, read reviews and decide broadly what they are looking for before the business sees any identifiable activity. This creates a dark funnel, where part of the decision-making journey happens outside the company’s own analytics.

Even considering how customers use Google, the increase in searches now showing AI overviews is stark. Your website traffic will look different, but understanding what parts of it looks different is key.

This also has implications for attribution. A customer may use ChatGPT to understand a category, compare providers and narrow the market before searching for one of those companies by name and converting. Depending on the model being used, branded search may then receive most or all of the credit. Marketers know this isn’t true.

CMOs should therefore review whether their current attribution approach still reflects the way customers are buying. That means looking at traditional attribution alongside branded search, direct traffic, customer surveys, changes in landing-page behaviour and AI visibility.

The objective is to build a better factbase before changing the budget.

2. Protect what is still working

AI changing discovery does not mean established marketing channels suddenly stop working. For many businesses, paid search, SEO, affiliates, social and email will continue to generate the majority of profitable growth during 2027.

Those channels should continue to earn investment where the economics remain attractive.

The difficulty comes when the numbers start to move. If organic traffic falls, there can be a natural reaction to spend more on paid search to compensate. In some cases that will be the right decision. In others, the business may simply be buying back traffic that has become less valuable.

If generic organic traffic falls while high-intent traffic, conversion and sales remain relatively stable, some of the missing traffic may previously have represented customers carrying out research that is now happening elsewhere.

Marketing leaders should therefore understand which traffic has actually been lost, what those visits historically contributed to conversion and whether higher paid spend is generating incremental customers or simply replacing a vanity traffic metric that doesn’t move the bottom line.

The same challenge applies to conversion rate optimisation. Many businesses have spent years optimising the website on the assumption that customers arrive early, browse several pages, compare options and gradually move towards conversion.

If more discovery and comparison takes place inside an LLM, that behaviour may change. Customers may arrive already knowing what they want, which means the part of the website that received the most CRO attention may no longer have the same influence.

That does not make CRO less important. It changes where the investment should go… consider if your investment in your website should be focussed on how you move people through the purchase funnel, or how you retain them post-purchase so they repeat and come back to you again and again.

Marketing teams need to be looking at where customers now enter, what they already know when they arrive and which interactions still materially influence conversion. Resource that has historically been focused on helping customers navigate or discover may create more value further down the journey.

The principle is simple: protect spend that continues to work, but do not assume every lost click needs to be replaced or every part of the historical website journey needs the same level of investment.

3. Prepare for where discovery is moving

The third area of investment is preparation. Even where the commercial impact of AI is still developing, there are foundations businesses can put in place now to improve how they appear as discovery changes.

This starts with understanding how your brand is represented beyond your own website. LLMs draw from review sites, publishers, forums, social platforms and other third-party sources, which means reputation, PR and wider online presence increasingly influence how a business is described and recommended.

Marketing teams should therefore understand which sources shape the way their category is represented and whether the business appears strongly within them.

There is also a technical element. For businesses selling products, structured and machine-readable information is becoming increasingly important. If systems cannot clearly understand what a product is, its attributes, availability or price, the likelihood of it appearing in a relevant recommendation falls. We explore this very topic in our recent webinar: How to focus your marketing as AI re-shapes the customer journey.

The same principle applies in B2B. Services need to be clearly described, expertise needs to be visible and the information the business wants associated with its brand needs to exist somewhere AI systems can find and understand it.

This has implications for content strategy too. There will still be a role for gated content where the exchange gives the customer something genuinely valuable, but marketers should be more selective about what they hide behind forms. If you want your business to be associated with a particular topic, service or area of expertise, some of that thinking needs to be publicly available.

It also means reconsidering the role of the website itself. If an LLM increasingly handles parts of discovery, comparison and shortlisting, the website may have less work to do in moving every customer through those stages.

For a B2C business, that may mean putting more emphasis on product-detail pages, checkout, account creation, post-purchase journeys and loyalty. For a B2B business, it may mean concentrating more on high-intent service pages, proof points, lead conversion, onboarding and the ongoing customer relationship.

Marketing leaders should therefore consider:

• Can AI systems clearly understand what we sell and who it is for?

• What sources influence how our category is represented?

• Are we hiding valuable information that would be more useful if publicly accessible?

• Where in the customer journey does our website still have the greatest influence?

• Does our current CRO roadmap reflect how customers are now arriving and behaving?

  
  

This is where some budgets may start to shift during 2027: towards structured data, technical improvements, content, PR, reputation activity and the points in the website journey that still create the most commercial value.

4. Experiment with dedicated budget set aside

There is still considerable uncertainty around how quickly AI-driven behaviour will develop across different categories, which makes experimentation an important part of the 2027 budget.

For some businesses that may involve GEO activity or new approaches to content. Others may want to test new paid placements, different forms of third-party visibility, changes to landing experiences or new ways of measuring AI-influenced demand.

The amount set aside will vary, but the money needs to be genuinely available. Too often, organisations create an experimentation budget on paper while every meaningful pound remains committed to existing channels.

Some of this spend may also produce learning rather than immediate revenue, so alignment between the CMO and CFO is important. Both need to agree on the level of risk the business is prepared to accept and what evidence would justify further investment.

Each experiment should have a clear objective. Marketing leaders should agree what they are trying to learn, what would justify increasing investment and what would cause them to stop.

This gives marketers room to test new ideas while keeping commercial discipline around how money is deployed.

The operating model needs to support all four 

Understand, Protect, Prepare and Experiment should not be treated as four fixed budget allocations agreed once a year. The proportion going into each should change as the evidence develops.

That requires an operating model that allows money to move. Someone needs to be accountable for pulling together the evidence on how customer behaviour is changing, while Marketing and Finance need an agreed rhythm for reviewing performance and deciding when investment should change.

The exact allocation will vary by business: the point is to distinguish between spend that helps you understand change, protects proven returns, prepares for likely changes in discovery and tests areas where the commercial case is still developing.

Marketing leaders should consider:

• Who owns the overall view of how customer behaviour is changing?

• How often should the four areas of investment be reviewed?

• Which parts of the budget can be moved during the year?

• What level of evidence is required before reallocating spend?

  
  

This does not mean constantly changing direction. A large proportion of the budget may remain in the same channels throughout the year because they continue to work. The difference is that when the evidence changes, the business is able to act.

A marketing budget built in October 2026 and left largely untouched until the following year is unlikely to be sufficient if customer behaviour continues to move at its current pace.

The main question for 2027

There will be no shortage of predictions about how quickly AI will change marketing over the next twelve months. Some will prove broadly right and others will run well ahead of actual customer behaviour.

For CMOs, the immediate priority is to understand which changes are already happening within their own customer base and make sure their budget can respond. That means understanding the customer, protecting what continues to work, preparing for the changes that are becoming visible and creating enough room to experiment where the outcome is less certain.

As marketing teams finalise their 2027 plans, CMOs should be asking themselves one final question:

How much of our budget reflects the way our customers actually buy today, and how quickly can we respond to this changing?

If you'd like to understand how visible your business is inside AI tools today, or help preparing for a more agentic commerce driven future, get in touch here.

All views expressed in this post are the author’s own and should not be relied upon for any reason.