AlchemyAlchemy — Monetize Iron Ideas

What an AI assessment covers, and what it doesn't

What an AI assessment covers, and what it doesn't
Before you pay anyone to tell you where AI fits in your business, know what a good assessment should hand you, what it cannot promise, and the questions to ask first.

"AI assessment" can mean almost anything. For some firms it is a free sales call. For others it is a months-long consulting project with a slide deck at the end. If you run a small or mid-sized business, you need something in between: a clear answer to where AI would actually pay off in your operation, and what it would cost.

Here is what a useful assessment should cover, what it cannot do, and how to tell a real one from a pitch.

What a good AI assessment should cover

1. How your business actually runs. Not your industry in general, but your quoting, scheduling, customer questions and back office. The wins are in the details: the calls that go to voicemail after hours, the same five questions your team answers every day, the invoices that go out a week late.

2. Specific tools, by name. "Use AI for customer service" is a category, not advice. A useful recommendation names the tool, says why it fits a business your size, and says why the bigger, better-known option does not.

3. Costs and returns, with the assumptions shown. Each recommendation should state what it costs to set up and run, what it should return, and the numbers behind that estimate, so you can check the maths against your own.

4. A ranking. Ten good ideas you cannot act on are worth less than three you can. Recommendations should be ordered by payback and effort, so you know what to do first.

5. The fixes that are not AI. Sometimes the best fix is a setting in software you already pay for, a simpler tool, or a change to a process. An honest assessment says so, even though "no AI needed" is less exciting to sell.

What an assessment cannot do

Be wary of anyone who promises these:

  • Guaranteed results. An assessment can estimate a return from your numbers. It cannot promise one. Your team still has to adopt the tools.
  • A dollar figure on everything. Some moves, like getting more reviews, pay off indirectly and over time. Putting a precise number on those is guesswork dressed as analysis.
  • Implementation. An assessment is a plan. Setting up the tools, migrating data and training your team are separate jobs, and should be priced separately.
  • A permanent answer. AI tools and prices change month to month. A plan is a snapshot, and the right tool in March may not be the right tool in June.

Questions to ask before you pay for one

  • What exactly will I receive, and when?
  • Will recommendations name specific tools and prices?
  • Can I see a sample report?
  • Is there a guarantee if the plan is not useful?
  • Do you also sell the software you might recommend? If so, how do you handle that conflict?

The last question matters. A firm that builds software has an incentive to recommend building software. Ask how they deal with it, and whether the plan is usable by someone else.

How we do it at Alchemy

The AI Assessment is one 45 to 60 minute call about how your business runs, from sales to back office. We find where AI wins back hours or margin, name the specific tool for each win, price it, and rank the list by payback. Your plan lands in 3 to 4 days. It costs $499, and if the roadmap is not worth the $499, we refund it.

Yes, we also build custom software. When a custom build is genuinely your best payback, it goes on the roadmap with its price next to it, but most roadmaps are mostly tools you can buy. The plan stands on its own, so you can hand it to any developer, including one who is not us.

You can see a sample assessment report for an invented business, built with the same method, tools and prices you would get. If you want an expert on call after that, there is ongoing AI advisory.

Not sure AI is worth looking at yet? Start with AI for small-business owners: what's worth it and what's noise.