Inside our free AI business analyser

The BespokeWorks Instant Business Analyser is a free AI tool that audits any company website and returns a personalised three-phase automation roadmap. You enter a URL, answer a few optional context questions (industry, company size, primary goal, biggest challenge, tools currently used), and a few minutes later you get a report describing exactly which AI and automation projects your business could implement in the next two weeks, two months, and six months, with realistic ROI framing for each.

It is free, runs without a signup wall, and is built on the same analysis stack we use internally when scoping client engagements. You can try it at bespokeworks.ai/instant-analysis.

When a founder approaches us asking “where should we start with AI?”, the honest answer is that nobody can tell them until someone has spent a few hours reading the website, looking at reviews, understanding the customer journey, and mapping the tech stack. That’s a consulting call, and consulting calls cost money. Most businesses that would benefit from AI never get to that conversation because there’s no cheap way to preview what it would even look like.

We wanted to compress that first consulting hour into something a founder could run in five minutes for free. Not a generic scorecard, but an actual roadmap specific to their business. So we took the analysis workflow our team uses internally, automated the repeatable parts, and shipped it as a public tool.

Most free website audit tools (HubSpot Website Grader, Neil Patel’s SEO Analyzer, Woorank, Seobility) run a rule-based checklist over on-page SEO, Core Web Vitals, mobile-friendliness, and a handful of technical signals. The output is a number out of 100 and a list of things to fix. That’s useful if you want to improve your title tags, but it has nothing to say about your actual business.

The BespokeWorks analyser is different in nine meaningful ways:

The short version: other tools grade your website. This one reasons about your business.

The reasonable comparison is not how long it takes the tool, but how long the same analysis would take a human doing it properly. We sat down and timed it, stage by stage, the way a junior analyst would actually work through a single client.

Independent UK business and digital strategy consultants typically charge between £100 and £250 per hour, depending on seniority. (Sources: Glassdoor UK consultant salary data, Payscale UK 2024 freelance rates, IPSE freelance pay survey.) That gives a simple formula:

That is the realistic outlay if a founder commissioned the same depth of analysis from a human. The free tool delivers it in roughly five minutes for £0. The point is not that humans are slow; the point is that the tool front-loads the parts of the work that are mechanical (crawling, structured extraction, signal aggregation) so a human’s time is freed up for the parts that actually need judgement (the conversation that follows).

A note on the formula: the time per stage is our internal estimate based on running this exact workflow manually before automating it. It will vary by site complexity and consultant experience. The hourly rate range reflects published 2024 UK consultant rates and is conservative at the low end. We have deliberately not used inflated “Big Four” hourly rates (typically £400 to £1,000+), which would make the saving look unreasonable. Take the formula, not the number; plug in whatever rate matches your reality.

The analyser runs a six-stage pipeline. Each stage produces a structured signal that feeds into the next. We’ve deliberately described the “what” below rather than the specific prompts, weightings, or playbook contents. Those are the parts that took months of iteration and are what makes the final output actually useful.

What makes each stage non-obvious

• Multi-page crawl with relevance filtering. Beyond about 20 relevant pages, extra crawling adds noise and latency without changing the output. The value is in which pages, not how many.
• Customer journey mapping. Every page is tagged for which part of the customer journey it serves (awareness, consideration, decision, conversion, retention) and how it links to the others. The analyser then spots missing stages, dead-end pages, and the spots where visitors drop off. Most audit tools have never looked at your site this way.
• Multi-signal business profile. Four independent sources (what’s on your site, the tools you’re using, your search rankings, your public reviews) are merged with an honesty rating per field, so the recommendation stage knows what’s confident and what’s inferred.
• Industry-specific playbooks. A generic AI tool will suggest “add a chatbot.” Ours consults playbooks for e-commerce, software, healthcare, professional services, hospitality, and B2B services. An e-commerce store gets abandoned-cart recovery and inventory alerts; a clinic gets appointment reminders and records workflows.
• Digital maturity scoring. The score decides whether the roadmap leans toward quick-win automations or deeper transformation, so a solo operator and a 200-person company get very different plans.
• Personalised 3-phase roadmap. The final stage names each automation, the problem it solves, what kind of tool you’d need, and a realistic cost-and-payoff framing, for quick wins (1 to 2 weeks), foundation (1 to 2 months), and transformation (3 to 6 months).

The combination matters more than any single stage. A journey graph on its own tells you where friction is but not what to do about it. An industry playbook on its own gives generic advice. Digital maturity on its own is a number. Together they produce a report that reads like a junior consultant spent an hour on your business, except it took about five minutes and cost nothing.

Several things that sounded obvious going in turned out to be wrong once we tested them against real businesses.

• More pages is not always better. Beyond about twenty relevant pages, extra reading adds noise and slows the run down without changing the recommendations. The value is in which pages, not how many.
• One-size-fits-all advice underperforms playbooks. We tried a single “universal” recommendation approach. It kept returning the same four suggestions regardless of industry. Splitting into industry-specific playbooks tripled how useful the output was.
• What customers say matters more than how many say it. A small business with twelve detailed reviews often surfaces sharper pain points than a large company with two thousand generic five-star reviews. We weight by specificity, not count.
• Better context beats a fancier model. Adding the twelve-question context form moved the needle on recommendation quality more than any AI upgrade we tried. Most free tools ask for a URL and stop there, and that’s why their output is generic.
• Speed is a feature. Users abandon analyses that take over three minutes. Every speed-up we made (running stages in parallel, only reading the pages that matter, asking the AI for structured answers) was driven by keeping the full analysis under five minutes end-to-end.

The analyser is designed for anyone who needs a fast, concrete answer to “where should my business start with AI automation?”

• Founders and operators scoping their first AI project and wanting a second opinion before committing budget.
• Consultants and agencies generating a structured initial point of view on a prospect before a discovery call.
• Operations leads building the internal business case for automation spend and needing an external reference point.
• Investors and analysts quickly assessing the digital maturity and automation headroom of portfolio or target companies.

It is not a replacement for a custom engagement. It is how we start a lot of them. If the output lines up with what you already suspected, you have confirmation. If it surfaces something you hadn’t considered, you have a lead worth exploring.