SaaS Startup marketing
November 25, 2026

All my customers are agents

Secret SaaS: the series in which I interview SaaS buyers, founders and marketers to learn new lessons in SaaS marketing.

Will AI agent customers forever change the way B2B companies do business?

There are tons of infographics and insights out there on how to use AI to improve your marketing and business processes. But what about the other side, when your customers are using agents and AI for their buying process? 

AI is shifting how we think, work and do business. It’s been on my mind a lot, but in particular, I’ve been looking into how it’ll affect marketing a B2B business. How will B2B businesses attract customers when those customers are sending out bots and agents to do most of their procurement work?

My particular interest is B2B because that’s where my career has been focused over the past decade plus. 

Individuals are starting to use AI agents to book holidays, buy clothes and electricals, entrusting them with credit card details and to make important financial decisions. There’s about to be an influx of retail agents from the likes of Google and Meta over the coming months.

In some cases, the human will never interact with the brand at all, apart from maybe a final approval to go ahead with the transaction. This means that the whole marketing funnel piece is being turned on its head. How are we meant to attract and nurture customers now? What rules should we be playing by?

There’s a lot that B2B can learn from B2C when it comes to agentic procurement and purchases, but there are still a good few differences between how the two might operate. 

Generally, B2B transaction value is higher, typically there are more stakeholders and decision makers and the buying cycle is longer, but AI agents can still be involved and businesses need to be prepared for them.

For years, I’ve been harping on that B2B marketing is actually like B2C because you’re still talking to humans, even if you’re marketing to a business, it’s not some faceless corporate entity, you’re always talking to real people.

But what happens when there are no humans making the purchasing decisions? Is in-bound dead? Can we still use emotional advertising and messaging or do we have to become bland and functional?

I’m just a curious person here, I’m on a quest to find the answers, but because everything is moving so quickly, these ‘answers’ are sure to keep evolving.  

But in this snapshot, I’m hoping to find out how B2B businesses can respond to agentic customers, maintain a brand and keep those leads and sales coming in. If you have thoughts, please send in your answers on a postcard!

‍

Let’s get dystopian… some thoughts on the dead internet theory

The dead internet theory is something I’ve long found fascinating. If you’re not familiar with it, it’s the concept that if all content and interactions are automated and AI generated, there will be a point at which humans on the internet become a minority. 

And if the AI algorithm is being taught on its own outputs, it’ll create more slop and the internet will just be a senseless mess. Time will tell on this one I guess!

When bots are ruling the majority of the internet, then what’s the point? Do we create a new humans-only internet with authentic connections… or is that what we call real life? I digress, this is philosophical talk for another day.

Let’s get back to the dead internet theory and how it might play out for B2B businesses and change the whole relationship between business and customer.

Take these examples:

The Claude website

You’re the founder of a B2B SaaS and run up a website in Claude in a few hours. It’s automatically optimised for schemas and SEO and you’re finding that most of the traffic is bots from LLMs or agents doing their research on your services. 

If no humans are visiting the site, then what’s the point in having a fancy website? It’s like the ghost kitchens built for Deliveroo, they don’t have actual premises (in this case, an attractive website) where people can visit, or indeed stumble across, instead they have a menu which people access though the Deliveroo app and a kitchen where they prepare the food.  

‍

The LinkedIn post

Or alternatively, you have a LinkedIn page for your B2B business. As a busy founder, you don’t have time for marketing and so outsource it to an agent to generate and publish your weekly posts. It looks like a potential customer has liked and commented on your post, but they’ve actually outsourced this to an AI agent. In this whole interaction, no humans have actually seen this content.

When AI bots and agents are generating the marketing content as well as consuming the marketing content, it creates an empty loop of machines talking to machines. 

More questions than answers here, these are extreme scenarios but it’s worth keeping them in mind as they could well become a reality, impacting the fundamental way that we do marketing and how ‘customers’ interact with brands.

My focus for the moment isn’t on how to use AI in your own marketing processes, there’s a lot of stuff about that already. 

I’m interested in how marketing will have to change when a potential customer starts outsourcing their B2B procurement decisions to an AI agent, when one day you wake up to discover that most of your inbound visits and enquiries come from bots and agents.

This change will involve a whole rethink on how companies market themselves and attract business and not enough people are talking about it yet.

Examples of procurement agents

Before we dive into how agentic commerce might affect B2B, it’s easier to understand how it’s likely to change B2C when individuals are using agents to help them with their everyday shopping decisions.

The state of B2C agentic commerce

The most concrete examples of agentic commerce are coming from B2C scenarios: people are starting to outsource some of their purchasing decisions. There are varying degrees of autonomy but agents are more and more involved in the comparison, shortlisting and sometimes purchasing.

B2C agentic scenarios might include the following:

Shopping for clothes

“I’m going to a wedding and need some navy blue leather court shoes to match this dress, they must be closed-toe, with a heel that’s no higher than 2 inches. Please bring up a list of options, from UK sellers, my budget is £80.”

Home improvements

“Here’s a photo of my broken Hoselock spray gun, please find and order the cheapest replacement with the fastest delivery.”

Booking a hotel room

“Please book me a 2 night stay in a hotel room for two people. The hotel should be within 1 hour’s drive of Birmingham. It must be in a market town which has a highly-rated tea room for a cream tea, countryside walks nearby, spa facilities and breakfast included. The budget is £500, please book the cheapest weekend in October”. 

While not everyone is entrusting their purchasing process to agents yet, we know that the way people are shopping is changing, 67% of shoppers now use LLMs for search and discovery when buying online, meaning the traditional Google Search and website browsing is becoming less prevalent.

There are still many questions on the B2C side about how commerce agents might affect the marketing, customer service and loyalty journey which you can read more about in this Substack article.  

As a business owner, your first question should be: how do I build my brand, stay discoverable and maintain sales when the audience is no longer human and instead being taken over by agents?

‍

What will agentic commerce look like for B2B?

Agentic commerce isn’t as simple as B2C for businesses buying from other businesses. On the whole, the B2B buying journey is longer, higher value and more complex. 

B2C commerce agents will mostly be optimised for convenience and price, while B2B agents are a little more complicated in that they’re focused on risk, contracts, reputation, usage and compliance. 

With 86% of business leaders claiming their companies now rely on AI agents for daily operations (Sifted), there is no doubt their function will spill into the procurement space. 

Again, Sifted have said that fully agent-run workflows remain uncommon at the moment. Typically, agents are handling one or two steps in a larger process, with humans overseeing and managing the rest (Sifted - Rise of the AI Agents) but even out sourcing one or two steps in the procurement process will have a profound effect on the way that marketing is done.

Outsourcing even a couple of phases in the procurement cycle will fundamentally alter marketing strategy. Below is a breakdown of the 5 key stages in the B2B buyer journey, along with real-world agent examples:

‍

Procurement stage

Market example

Key capabilities

1. Social Engagement

Buzzli

Automates LinkedIn content generation and profile interactions

2. Email Filtering

Fyxer

Prioritises high-value, relevant messages for review

3. Vendor Evaluation

Lyzr

Aggregates supplier details, benchmark metrics and pricing structures

4. Negotiation

Globality

Independently handles supplier pricing and contract negotiations

5. Contract Renewal

Tropic

Monitors internal software adoption, drafts technical requirements, and manages renewal terms

‍

Marketing to the AI-powered B2B buyer

1. Social engagement: traditional measurements aren’t as reliable

Customers are using agents and bots to manage their social presence, from liking and commenting on articles, to auto-generating posts. 

It’s becoming harder and harder to rely on social engagement to measure intent as there are fewer humans in the process interacting with your brand. 

Real people are no longer being influenced or getting brand awareness through the traditional social routes. 

‍

Proactive marketing:

‍

  • Lean into human activities such as going to events, in-person roundtables, coffee meet-ups, where real humans are all in the same room.
  • Show up in the closed communities: Reddit and Slack groups, podcasts, vetted spaces for real people.
  • Content-wise, be real and unique by producing case studies and real-world problems and solutions, use different multimedia formats like videos, and very opinionated pieces from your own point of view to get more cut-through. 
  • Give traditional ‘clicks’ and ‘views’ less weighting. Look more at high-intent actions such as attending a webinar or downloading a deep-dive report. (Personally, I’m waiting for some more insights in this book when it gets published, One Click Marketing).  

‍

2. Email filtering: The agent decides what content gets seen by the human

The AI agent is becoming the inbox gatekeeper, summarising what matters and forgetting the rest.

This could mean that the traditional email nurture sequence isn’t quite as effective as it once was as you’re no longer vying for the attention of a curious human. 

Proactive marketing:

‍

  • By all means, keep the clever copywriting and catchy subject lines in your mass emails but it’s also necessary to skill up on the technical and ‘write for the machine’.
  • Be clear in email copy about the value proposition, relevance and action required is, all in plain text. If it hits on a subject that the agent ‘thinks’ is relevant to the user, it’s more likely to get through.  
  • Try to weave in highly-specific data, it could be based on the industry or something that’s particular to the company so that the AI recognises the keywords and is more likely to present it to the user.
  • Find ways to bypass the digital inbox through direct mail, LinkedIn messages and referrals. Account-based marketing practices are still a strong contender here too.

‍

3. Vendor evaluation tools: aim to get on the list

Customer agents might already be scraping the web and consolidating data on your brand and comparing your services against competitors before they have even filled in an enquiry form.

The ‘auto RFP’ is a real possibility: that you’re pitching for work with data and content that’s publicly available, without even realising it.

It’s likely that you’ll lose deals without ever knowing you were considered for them in the first place, just because some information was missing from your website.

‍

Proactive marketing:

‍

‍

  • Make sure that your website is crawlable, use a JSON-LD schema for your product offering or catalogue which features pricing, features, specifications and reviews. A JSON-LD schema is added to your HTML website script. It features structured data which can be easily crawled by AI agents. For more information on how to structure JSON-LD schemas, visit schema.org.
  • Consider putting more info up front than you would have before, this includes pricing, technical specifications, detailed case studies, original research and thought leadership - the more unique, the better.
  • Traditional SEO principles (and subsequently AEO - answer engine optimisation) still apply here too in the form of 3rd party sources for credibility such as review sites, article mentions and backlinks.
  • Maybe it’s time to ditch the gated content, only keeping it for super-high intent actions such as downloading a very specialist report, signing up for a trial or accessing a demo. 

‍

4. Vendor negotiation: be prepared for the influx

There might be a time when your potential customer’s agent visits your website sending questions and queries automatically, then shortlisting autonomously based on specific criteria. 

Fielding these queries manually will no doubt be quite arduous and you might miss your chance if you take too long to respond. 

I wouldn’t always recommend fighting AI with AI but at this stage, if you’re starting to get an influx of agent queries, then maybe it’s time to consider it.

Proactive marketing:

‍

  • Start by working on a flawless and up to date knowledge base for AI agents to crawl, making sure that the data is structured and up to date from specification sheets, to case studies, certificates to pricing tiers. 
  • Consider having your own agent that sits on the website to answer the agent questions at scale. Ensure that there are robust guard rails in place as you don’t want your vendor agent to be promising the moon on a stick or giving your service secrets away for free!
  • All these points are very uniform and technical but there’s still room for creativity here. At SOME point, a human will be involved in a decision so you have to be clear here on your value proposition and what sets you apart from competitors. 
  • The place where you absolutely need the human touch is for an RFP response. Many companies are going the other way and automating it all. But by doing this, you’re just adding to the homogeny and making it harder for the bid manager to shortlist when every response sounds the same. Instead, try and go the other way. Once you’ve covered off the data and technical points, find ways to stand out and be memorable.

‍

5. Contract renewals: don’t wait until the year is up, be proactive 

This one’s specifically for SaaS providers: renewal agents are now able to track and monitor internal usage data for all types of software, specifying exactly how much the software is being used, by whom and what functionality is needed.

When it comes to renewal, the agent is able to drive a hard bargain if certain features or licenses aren’t actually being utilised. It’s also able to write an RFP ready to send out to the competition if the software isn’t fit for purpose. 

Having a few super-users on side and a brilliant account manager probably isn’t enough if the renewal agent data shows that your software is overpriced and hardly used.

Proactive marketing:

  • Don’t see the sign up or subscription as a ‘win’ then forget about it. Retention efforts must begin as soon as they become a customer. Drive usage through education, gamification and tutorials. Keep a pulse on user satisfaction and complaints too, dealing with any issues quickly and feeding that into product development if needed.
  • Be proactive on ROI reporting, create your own monthly report for the client that helps to feed the buyer’s own internal systems, showing the exact ROI, time saved and revenue generated so that the AI factors in your data too.
  • Share proven quantifiable benefits of the software such as speed to implementation or service cost reductions.

‍

What does B2B agentic commerce mean for the marketing funnel?

Simplistically, the traditional marketing funnel has 3 parts…

  1. Awareness: Getting seen by your target audience, people start to associate your brand with what you provide. This step is for building recognition of your brand so that people know what you do and how you can help them.
  1. Consideration: Once you’ve built up recognition, now it’s time to become the ‘favourite’. Here, you get people to engage with your brand and know why they should choose you over the competitors. 
  1. Usage: This is where the conversion to purchase happens: they trust you enough to buy from you. 

‍

At each stage of the funnel, you’re establishing a relationship with someone in your target market, from knowing about the existence of your brand, to building up trust then an eventual conversion. 

But if you add AI procurement agents into the mix, the whole marketing funnel and what it stands for quickly becomes irrelevant, because AI agents operate differently to humans. 

If an AI agent is asked to shortlist some providers and send out an auto-generated RFP, in a sense that means the middle of the funnel has been flattened. 

In the classic B2B funnel, buyers spend weeks moving through blog posts, gated white papers and email nurture sequences. When the agent is involved, the entire search, comparison and recommendation phase can happen in seconds. 

So humans are only really going to be actively involved at the top of the funnel and at the bottom. 

At the top of the funnel some B2B buyers will have their own shortlist in mind after attending an event, or ask trusted peers in groups what their recommended tools are.

And at the bottom, once an agent has shortlisted and recommended a possible vendor, the human is likely to step in to evaluate brand reputation, legal safety as well as the ‘gut feel / je ne sais quoi’.

‍

The two-track marketing system

Spending all your time trying to ‘crack the code’ by thinking solely about attracting the AI agent will only lead to dull and homogenous companies. 

To really market effectively when your B2B customers are sending out AI agents, you need to have a two-track system: one track built for machine-readable logic and another reserved for human trust.

  1. Be appealing to the agents and bots, get your foot through the door. 
  2. Build the brand for real people, aim to be ‘unscrapeable’.

Jamie Smith, expert in B2C personal agents stresses the importance of brands to create a connection with their customers. Agents risk making the transactional experience 'functionally excellent, but emotionally invisible', meaning you might provide an excellent service or appropriately-priced product but the customer won't remember who you are. In order to adapt, he says that brands will need to create emotional connections elsewhere, through demonstrating values, creating experiences and helping customers accomplish something meaningful.

What this looks like when you put it together:

Stop 

Start 

Marketing for the agent

Gating basic content and information 

Put more information up front such as white papers and contact information

Hiding pricing tiers

Make pricing and packages comparable and readable for the bots

Irregular product information 

Structure JSON-LD schemas about product and specifications

Confusing jargon and ambiguous claims 

Clear and proven ROI proof points and case studies about how the product has benefitted customers 

Vague product features and benefits

If you’re providing B2B software, publish detailed API documentation so the agent can verify technical fit

2-3 days response time to in-bound enquiries

Consider a bot which is available 24/7 to answer questions for the procurement agents (with strict guardrails in place) in order to compare specifications, negotiate initial contract terms.

Marketing for the person

Generic, keyword-stuffed articles

Unique and original studies and content 

Inoffensive, (boring) social updates


Expert and opinionated founder-led thought leadership. Consider different media such as video content and podcasts.

Relying solely on public online marketing channels

In-person events, industry Slack groups, invite-only spaces, podcasts

Jargon-heavy, AI-generated branding content

Clear brand messaging on what really makes your brand different 

Marketing just for the bots

Be creative and unique!

‍

How will buying agents affect the future of inbound B2B marketing?

It’s becoming clear that getting the attention of humans is becoming harder and harder. When you can create content at the click of a button, there’s no end in the supply but when it comes to giving it attention, we still have the same amount of hours in the day, with a million other things to look at.

But when you replace that attention with AI agents being sent out to engage with posts, carry out research on companies, create shortlists and negotiate deals on behalf of their human clients, then that changes the whole way we market and get sales in B2B. 

I don’t believe that the B2B agentic commerce wave will happen overnight, and when it does happen, companies won’t be entrusting their whole purchasing decisions to an agent. There will ALWAYS be humans in the process.

And as a business owner dealing with in-bound customer agents, be aware of their imperfections. As Michelle Johnson says, the more AI agents that are used, the lower the success rate is. If 10 agents each have a success rate of 95%, then the end-to-end success rate is just 60%. So if you get an absurd question or request, it’s probably the agent’s fault and you’ll really need to reach out to a real human for clarification.  

There will be a small number of companies entrusting everything to the AI agent, others who outsource a couple of purchasing stages, and another group who will be forever skeptical. 

The old way of winning in B2B is to build up a content list and following, the new way is to get onto the AI Agent’s ‘list’ then work on winning the human over.

I am convinced that one day soon, there will be a rebellion on AI homogeny and brands which set themselves apart from others will be those that can prove their ‘humanness’ will succeed. 

We’ve seen ‘analogue’ resurgences like this before… 

  • A preference for UK call centres in the face of overseas cost cutting.
  • The fact that paper books are still more popular than ereaders.
  • That music artists are still putting records out for a less digital listening experience.

‍

Be creative, be unique

Business owners and marketers will have to adapt to non-human AI agents but we must not turn this into a systematic and mindless checklist of an exercise. By focusing all our attention on how to win over the AI agent, don’t lose sight of the soul of your business, what makes it different and unique.

Here’s how we can adapt:

‍

  1. Structure marketing in 2 tracks: winning over the ‘machine’ (AI agent) and building up trust and credibility with the human.
  1. But be bold, be different, be creative, build a strong brand for that point when a messy, emotional, irrational human is involved in the mix.

The way to win and keep the sales coming when your customers start becoming agents isn’t all about being technical and ‘marketing to the bots’. Realise that the agent can get you onto the shortlist but the human should often have the final say. 

Got any input?

This is very much a work in progress, if you have any ideas about how B2B agent procurement will affect the future of marketing, please reach out. You'll also find some bitesize content on this on my LinkedIn page.

Need help adapting your marketing to AI procurement agents?

If you have a B2B company and need to figure out how to adapt your marketing for AI agents, get in touch to arrange a call with me.

" Are you ready to bring in a fractional marketing director for your B2B company? From ongoing support to one-off projects, strategy to delivery, harness marketing expertise at a fraction of the cost of an employee."
‍
- Georgina Willison, Fractional Marketing Director, Marwhal.