For the past few years, businesses have been encouraged to use AI everywhere.
Write more content. Analyse more data. Automate more conversations. Move faster with fewer resources.
LinkedIn was part of that push, introducing tools designed to help users generate and improve posts. Now, it is changing direction.
The platform is reducing the reach of generic, repetitive AI content, blocking automated engagement and allowing users to flag posts that “seem like AI slop”. It is also replacing its AI-powered post-enhancement tool with proofreading that preserves the writer’s original voice.
LinkedIn is not rejecting AI completely. It is drawing a line between content that has been assisted by AI and content where AI has replaced the thinking.
AI can identify a problem. Humans still need to understand it.
In CRO, AI can help businesses process huge amounts of quantitative data.
It can highlight:
- Falling conversion rates
- Common drop-off points
- Changes in traffic quality
- Underperforming landing pages
- Unusual behaviour across devices
- Gaps between acquisition and revenue
That makes it a valuable diagnostic tool. It can tell us where something appears to be going wrong much faster than a person manually working through every report.
But data doesn't always explain why it is happening.
A high exit rate on a product page might indicate unclear information, poor photography, confusing delivery options, a lack of trust signals or simply that the traffic arriving there was never relevant.
AI can surface the pattern. Human research, judgement and empathy help us understand the behaviour behind it.
This is the human side of CRO: building trust, reducing anxiety and creating an emotional safety net that makes someone feel comfortable enough to buy.
It means understanding the questions people ask before making a decision, the reassurance they need and the small details that make a brand feel credible.
Not every valuable interaction can be reduced to a conversion rate. Sometimes the value is familiarity, confidence or a positive experience that brings someone back later.
Optimising for people, not just metrics
If we focus only on what is easiest to measure, we risk creating technically optimised experiences that do not feel particularly good to use.
A shorter checkout is not automatically a better checkout if important delivery information disappears.
A more prominent call to action will not necessarily increase sales if the customer still doesn't trust the business behind it.
Removing content might make a page look cleaner, but it could also remove the information someone needs to make a confident decision.
Effective CRO combines quantitative insight with qualitative evidence:
- What are customers asking your service team?
- Which objections repeatedly appear during sales conversations?
- What language do customers use in reviews?
- Where do people hesitate during user testing?
- What makes someone choose your business over a competitor?
- What information would help them feel safer purchasing?
These are the human questions. The answers often come from talking, listening and observing rather than simply generating another dashboard.
Chris Garbutt at Salesfire, our CRO partner, says
AI is changing what's possible when it comes to CRO. The specialised AI-driven solutions now available are more informed, more intuitive and better equipped to respond to customer needs than ever before.
At Salesfire, we use AI within our solutions to analyse behaviour, build customer profiles and understand intent across sessions and devices. These insights help us deliver hyper-personalisation across key touchpoints and ultimately drive on-site conversions.
The same shift is happening across marketing
LinkedIn’s changes reflect a broader problem online: when everyone has access to the same tools, producing more content is no longer much of an advantage.
The advantage comes from having something worth saying.
SEO needs original experience
AI can speed up keyword research, identify content gaps and help organise information. But publishing ten articles that repeat what already exists is unlikely to create meaningful value.
Strong SEO content needs genuine expertise, first-hand experience and a clear understanding of what the person searching actually needs.
That might mean including original data, answering questions competitors avoid, showing how a process works in practice or adding insight from the people delivering the service.
AI can help structure that expertise. It cannot manufacture the experience behind it.
PPC needs an understanding of motivation
AI can analyse campaign performance, spot trends and support faster testing. But it cannot fully understand why a message resonates with a particular audience.
The strongest ads are not always the most polished. They are the ones that recognise the problem someone is experiencing and offer a credible next step.
That requires an understanding of the customer, the market and the emotional context around the purchase - not simply another variation of the same headline.
Maria Johnston, Aware's PPC Lead, offers her opinion on the use of AI
AI is especially valuable to me as a Paid Media Lead with ADHD. It helps me turn an overwhelming one line task into clear, manageable actions, condense my (notoriously) long notes, and handle repetitive work such as comparing order numbers and revenue across ad platforms, backend order exports and GA4. That gives me more time and mental space for strategy, creativity and consumer psychology which I gladly dont think AI has a good grasp on as of yet.I also use it as a learning and research assistant.
For example, I receive a weekly briefing on important developments across Google, Meta and Microsoft Ads, which helps me identify the changes that could genuinely affect our clients. It is also a useful sounding board for testing ideas, challenging my reasoning and identifying gaps in my thinking - a lot of people may ask AI for ideas and run with it, but I think asking for ideas and receiving non out of the box ideas is the best way to rule out 5 things you DON’T want to do (which is usually what is recommended) and gives you the chance create better strategy based on what you actually believe and think outside of the box.
However, I think advertising is currently suffering from the overuse of unedited AI copy and creative. AI can imitate a tone of voice and reproduce an approved format efficiently, but it cannot fully understand how someone’s identity, experiences or emotions will affect the way they perceive an advert. That human insight is still essential.For me, AI works best as a collaborator rather than a replacement. I use it to reduce admin, repetition and cognitive overload, so I can spend more time doing the parts of my job that require original thought, empathy and judgement.
Email needs a reason to be opened
Automation makes it possible to send more personalised emails at exactly the right time.
But timing and segmentation cannot rescue a message that feels empty.
Useful emails answer questions, offer relevant ideas, reflect the customer’s relationship with the brand and give them a genuine reason to return. The technology supports the relationship; it doesn't create one by itself.
Social needs a point of view
LinkedIn has said that posts using AI to refine language are still welcome. What it is reducing is the distribution of content that feels generic, repetitive or lacking in clear perspective. Its early systems reportedly identified this type of content correctly in 94% of tests.
That means the goal is not to make every post sound perfectly human. It is to make sure there is a human idea behind it.
A useful opinion, a real lesson, a specific example or an honest disagreement will usually contribute more than another polished post built around a familiar formula.
Holly Payne, Aware's Marketing Exec, brings another perspective:
Weirdly, LinkedIn has gone from pushing AI-generated posts, with its own built-in writing tool - to adding a “report AI slop” button. Bit ironic, isn't it really.
Social media hasn’t fundamentally changed, but the environment and creative process have. Audiences increasingly value content that feels genuinely human, which helped fuel TikTok’s growth from Musical.ly.
AI learns from existing communication, so it can only reflect, remix and repackage what already exists. It can present ideas differently, but it can’t create a genuinely original point of view shaped by lived experience. That still has to come from a person.
As more people use AI to create more content, simply posting regularly is no longer enough. When everyone can produce something polished in seconds, polish loses its value. What stands out is experience, personality and original thought.
The issue isn’t whether AI was involved. It’s whether a human had anything worth saying in the first place. AI can help package an idea, but it can’t replace the lived experience and judgement that make people trust it.
Use AI to support the thinking - not replace it
AI is most useful when it creates more space for human judgement.
Use it to:
- Process large amounts of performance data
- Find patterns that deserve closer investigation
- Summarise research and customer feedback
- Produce an initial structure
- Generate testing ideas
- Handle repetitive production tasks
Then bring people back into the process.
Question the output. Add context. Speak to customers. Include first-hand knowledge. Decide whether the recommendation actually makes sense for the business and the people using it.
Three actions to take this week
Review one customer journey as a person, not a dashboard.
Go through it from the first search or advert to the final purchase. Note every moment that creates confusion, uncertainty or mistrust.
Find the human source behind your next piece of content.
Before opening an AI tool, speak to someone in your business who has direct experience of the subject. Start with what they know, not what the tool can generate.
Look for the “why” behind one performance issue.
Choose a page, campaign or email that is underperforming. Use the data to identify the problem, then use customer feedback, recordings, reviews or conversations to understand the cause.
Aware's marketing director, Chris Munrow says:
AI helps me debug code and build things without relying on a developer, saving clients time and getting results quicker. I’ve used it to create calculators, Shopify landing pages, schema, product feeds and client reports, as well as automate meeting notes and repetitive admin.
The goal is to let AI handle more quantitative work so we can focus on qualitative work that needs human judgement.
But AI is only as good as the context it’s given. We’ve seen it mistake fixed double tracking for falling performance and a Cloudflare block for hundreds of site errors. It’s useful, but its output still needs to be questioned - not taken as gospel.
BUT...
That doesn’t mean AI isn’t valuable. Used behind the scenes, it can save huge amounts of time on repetitive, low-value work like organising data, summarising notes, formatting reports or creating first drafts. If it isn’t customer-facing and doesn’t need a unique point of view, who cares whether AI helped produce it?
The point is to spend less time on admin and more time on the work that actually needs human judgement, creativity and connection.
Paul Gray, Aware's SEO lead says:
It can speed up work, mapping keywords to pages, mapping keywords/prompts to blogs, on-page content, FAQs, generating JSON schema etc. It can just increase efficiency and productivityBut recently we have developed a system primarily through ChatGPT, which connects to our data such as GA4, Google Search Console, SEMRush, and Screaming Frog automated crawls and compiles monthly reports, weekly email recaps and daily anomaly emails.
Apart from occasional cross-referencing of data to ensure accuracy this is fully automated. It also takes the monthly report and puts it into a branded PDF doc that can be used to report to stakeholders or clients.
We talked to one expert who spends 4 hours per month per client doing reporting for clients and has 10 clients, this AI system reduces that time to a couple of hours across all clients just for checking accuracy, and would save the Reddit user around 38 hours a month (basically a full week of work)
Bringing it all together
LinkedIn’s change of direction doesn't mean AI has failed.
It means simply using AI is no longer enough.
When automated content becomes easier to produce, original thought becomes more valuable. When every brand can generate a polished message, trust, experience and genuine perspective become the differentiators.
The businesses that benefit most from AI will not be the ones that remove people from the process.
They will be the ones that use technology to understand more, move faster and create more time for the human work that drives meaningful growth.
If you need support & a team that does the work, get in touch.
