The Death of SaaS: Why I'm Building 7 Agentic Products to Replace the Tools You Pay For

Every era of software has had a defining lie. In the desktop era, the lie was productivity: tools were sold on the promise of doing work faster, when in reality they introduced new work to manage the tool. In the SaaS era, the lie was collaboration: dashboards multiplied, seats multiplied, and the marketing team that ran on five tools in 2010 ran on thirty by 2022. The work did not shrink. The work moved into the tool.

The lie of the next era is already visible. It is the lie that AI is a feature you bolt onto your existing software stack. It is not. AI is the thing that finally lets you delete the stack.

1. The unit of value flips from access to outcome

A SaaS product sells you access to a workspace. You log in, you configure, you operate, you measure, you adjust. The vendor's job ends at the login screen. Yours begins. This is why a £400-a-month Klaviyo licence still requires a £4,000-a-month operator to make any revenue out of it. The tool is sold. The outcome is your problem.

An agentic product sells you the outcome itself. You do not log in. You do not configure flows. You do not segment audiences. You give the agent the brand, the catalogue, the goal — and the agent produces the revenue. The vendor's job ends when the outcome is delivered. The seat licence, the operator, the dashboard, the reporting layer all collapse into a single line item on the invoice: revenue produced.

2. Why I am building seven, not one

A single agentic product is a feature. A portfolio of seven across different categories is a thesis. The thesis is the asset. Once the market has seen Autothink replace email, credit control, checkout, social, CI, address validation and references, the question is no longer does this work? The question becomes what's next? — and the founder who has shipped the first seven owns the answer.

The seven, in plain language:

• AutoEmails — replaces Klaviyo, Mailchimp, HubSpot Email. The agent owns the email and SMS revenue line. No flow builder, no segment editor, no template designer. (See the deeper piece: why we built AutoEmails.)
• AutoSettle — replaces the credit control function inside Xero, QuickBooks and the average finance team's chase spreadsheet. Collects the cash. Names the deal-breaker. Logs the call.
• AutoTill — replaces the static ecommerce checkout. Turns the most important page on the site into a closing conversation.
• AutoSocials — replaces Hootsuite, Buffer, and the average social-media manager's calendar. Owns the post, the reply, the DM, the brand voice.
• Competable — replaces Crayon, Klue, and the £40k-a-year competitive-intel analyst. Always-on. Always current.
• AutoPostcode — replaces Loqate / PostcodeAnywhere for UK address validation, autocomplete, and address-data quality. (Already live — see /tech-stack/autopostcode.)
• AutoRefs — replaces the references-chasing process inside lettings, recruitment and finance. Books, chases, verifies, reports.

None of these are AI features grafted onto existing categories. Each is a replacement for a category. The naming is deliberate. Auto- as prefix tells the buyer the agent does the work. The category noun tells the buyer which incumbent to cancel.

3. The CFO becomes the buyer

In SaaS, the buyer is the head of the function. The marketing director buys Klaviyo. The finance manager buys Xero. The sales director buys Salesforce. The CFO signs the cheque, but the buyer has already chosen.

In agentic software, the CFO becomes the buyer. Because the line item is no longer tool access — it is outcome produced. The CFO can read the invoice. £18,400 of email-attributed revenue, £6,200 of cash collected, 312 references verified. The buying conversation moves from feature comparison to ROI maths. SaaS vendors are not built to sell this conversation. Agentic vendors are built for nothing else.

4. Why the UK, why now, why me

The UK has 5.5 million SMEs, an under-invested SaaS market, and a finance community that already buys on payback period rather than feature richness. It is the right market for outcome-based pricing. The window to claim the category labels — AutoEmails, AutoSettle, AutoTill, Competable — is roughly eighteen months. After that, the labels are taken and the moat shifts from naming to distribution.

Which is where Creative Agent comes in. The agency is not separate from the product portfolio. It is the distribution layer. Every Autothink product is pressure-tested against real UK clients before it is named, so the launch is not a press release — it is a working customer base on day one. The product feeds the agency new categories to sell. The agency feeds the product real-world data and revenue. The loop compounds.

Twelve years and three hundred clients in, this is the most founder- shaped problem I have ever seen. The technology is finally ready. The buyer is finally educated. The incumbents are finally defendable targets. The category labels are finally available. There is no version of this thesis that involves waiting.

5. What this means for you, depending on who you are

If you run a UK business — audit your software stack and write down, for each line item, the outcome the tool was bought to produce. Wherever the answer is "the tool helps a person produce the outcome", that line item is a candidate for replacement before 2027. Email, credit control, checkout, references, address quality, social, CI — start there.

If you are a founder — stop building features inside saturated SaaS categories. Pick one workflow, ship the agent that owns the outcome end-to-end, price it on outcome, and use the saved seats and hours as the sales narrative. Distribution is the moat. The technology is not.

If you are a journalist or analyst — the story is not "AI inside SaaS". The story is "outcome-based pricing breaks the SaaS P&L". That is the headline by the end of 2026, and the UK is the cheapest market in the developed world to watch it happen.

6. The next twenty-three pieces

This is week one of a twelve-week series. Across the next twenty-three pieces I'll go inside each Autothink product, each pillar of the AI-first marketing thesis at Creative Agent, the founder operating cadence behind shipping seven products in parallel, and the broader industry architecture I think the UK ends up with by 2030. The full schedule lives at /insights.

If you want to apply any of this to your own business, the most useful thing you can do is book a call. If you want the data behind the thesis, /research is the chronicle. If you want the marketing engine that delivers it, that is what Creative Marketing Group is for.