AI or a New Website: Which Should Your SMB Pay for First?

If you’re a UK SMB owner with a finite budget and two live conversations (one with a web agency about a new website, one with an AI consultancy about automation) you need to pick one. They both sound valuable. They cost similar money. Both agencies are telling you their thing is the priority.

Here’s the honest framework we use with clients to work out which comes first. Spoiler: most SMBs get the order wrong, which is costly.

The fundamental question

Before either decision, answer this:

Are you losing more money to not being found / not converting website visitors, or to wasted operational time?

If the answer is “not being found / losing visitors”: the website comes first.

If the answer is “wasted operational time”: AI might come first, but only if the underlying website foundation is strong enough not to be a bottleneck.

If the honest answer is “both, equally”: website first. Always. Here’s why.

Why the website almost always comes first

1. The website is your shop front. Even if 90% of your enquiries come via word of mouth, the first thing people do when they hear your name is Google you. If they don’t find you, or find a 2016-era site that looks abandoned, you’ve already lost trust.

2. AI multiplies what your existing systems do. If your existing systems include “we rely on enquiries to walk in the door”, AI multiplies nothing. An AI enquiry triage agent with no enquiries to triage isn’t helpful.

3. A website pays back on the demand you’re already creating. Word of mouth, referrals, past customers, all of that lands on your website. A website that converts 3% of visitors versus 1% doubles your effective enquiry volume without you doing anything else differently.

4. AI projects depend on the website for execution. The AI chatbot lives on the website. The form that captures enquiries for the AI triage agent is on the website. The SEO work that drives traffic runs through the website. If the foundation is broken, every AI project on top of it inherits the problem.

When AI comes first

There are exceptions. AI can come first if:

Your website is already strong. Loads fast, ranks for your key terms, converts visitors at a decent rate, works on mobile. Then AI is the right next multiplier.

You’re operationally drowning. If the team is so stretched that you’re losing enquiries because nobody’s chasing them up or invoices are 60 days late or quality is slipping, fix the operational chaos first. A shiny new website that generates more enquiries you can’t handle just amplifies the pain.

Your business model is 100% operational efficiency. Some businesses make money purely on internal margin, a contract manufacturer on fixed-price work, a distributor with set supplier deals. There the website is almost incidental; the AI operational win is pure profit.

If any of these describe you, AI before website might be correct.

The honest “both” scenario

Some SMBs genuinely need both, and are better off doing one then the other in fast succession, not trying to combine the budgets into a weaker single project.

A realistic sequence:

1. Months 1 to 2: New website. £3k to £8k. Foundation laid.
2. Month 3: Website goes live, starts collecting data (enquiry sources, bounce rates, conversion points).
3. Months 4 to 6: First AI project, informed by what the website data shows you about demand and bottlenecks.

The reason to sequence rather than combine: you’d rather spend £5k on a great website and £5k on a great AI project, than £10k on a mediocre combined agency engagement that does neither really well.

Red flags on both sides

Watch for these on the web agency side:

• “We build AI-powered websites”. Most don’t. “AI features” in the brochure often means a generic chatbot plugin.
• “We guarantee page 1 on Google in 30 days”. Dishonest.
• Month-to-month retainers tied to the website they built. Vendor lock-in.

Watch for these on the AI consultancy side:

• Not asking about your website at all. If the consultancy doesn’t check whether your foundation is sound, they’re not thinking holistically.
• Starting with a chatbot demo. The demo-first approach is usually a sales tactic, not the right first project.
• No clear payback framing. “AI will transform your business” isn’t a business case.

A simple decision flowchart

Answer in order:

1. Does your website rank in the top 5 for your key services + locations on Google?
   • No → website first.
   • Yes → continue.
2. Does your website load in under 2 seconds on mobile?
   • No → website first.
   • Yes → continue.
3. Is your enquiry form working and tested in the last 30 days?
   • No → fix the form urgently (usually a sub-day job), then continue.
   • Yes → continue.
4. Is your team spending significant hours every week on repetitive, structured tasks that could be automated?
   • No → keep optimising the website, run some paid ads, come back to AI in 6 months.
   • Yes → AI project is worth the money.

This gets most SMBs to the right answer.

What happens when you get the order right

The pattern we see consistently:

• Website done well: enquiries rise 30-50% within 3 months of going live.
• AI project done well after: operational capacity rises enough to handle the new enquiries without hiring more staff.

Total cost: £5k website + £5k AI = £10k. Typical first-year return: £30k to £80k across increased enquiry conversion, extra revenue, and saved operational hours.

Pay-back timelines are usually 3 to 12 months across both.

What happens when you get the order wrong

Pattern we also see consistently:

• AI project before website: operational capacity rises, but enquiry volume stays flat because nothing’s changed at the front door. Team has spare capacity they don’t need. Project feels underwhelming.
• Website before AI but without any plan to invest in AI after: enquiries rise, team drowns, service quality dips, some of the new enquiries churn out because you couldn’t service them fast enough.

Both wrong orders leave money on the table.

Frequently asked questions

Can one agency do both? Some can. Most can’t. We’re in the unusual position of doing both well (web + AI), but we’re also honest when a client should split the work across specialists. Being good at both doesn’t mean forcing both into one contract.

What if my budget only covers one? Answer question 1 of the flowchart. If your website isn’t ranking and converting, everything else is a distraction. Website first, AI when revenue allows.

What about SEO and paid ads, where do they fit? SEO is part of a good website build (and ongoing). Paid ads work on top of either, as long as the website they land on is worth landing on.

How long before I should expect results from either? Website: measurable improvement in enquiries within 60 to 90 days. AI: measurable improvement in operational hours within 30 to 60 days of going live.

If you’re making this decision right now and want a straight answer for your specific business, book a 15-minute intro call. We’ll tell you which one to do first, based on actual data, not agency self-interest.