Last updated: July 2026
A useful AI interview-prep output needs four things: the real role context, safe inputs, verified sources, and your own judgement on top. Give it those and it saves you hours. Skip them and you get advice so generic it could apply to a barista in Perth or a CFO in New York.
I use AI every day to research companies and prepare for conversations. The single biggest lever on whether it helps is context. Type "help me prepare for a sales interview" and the tool guesses. Tell it the product, the buyer, the route to market and the territory, and it starts working on your actual role.
This page is the method for both sides of a building products sales interview. Hiring managers use it to write sharper questions. Candidates use it to prepare properly. The complete copy-paste prompts live in the AI interview-prep toolkit. This page is the thinking around them.
What context makes an AI interview prompt specific to building products?
The failure I see most often is simple. Without enough context, AI writes questions for a designer, a builder or an engineer rather than for a salesperson working in building products. It tests the wrong job. The person you are hiring, or being hired as, sells the product. The prompt has to make that clear.
Give the model these fields and the output changes completely:
- The role and what it is responsible for: business development, account management, or specification.
- The product and its technical threshold: cladding, insulation, lighting, facade systems, flooring.
- The buyer and the route to market: specification through architects and designers, direct to builders and contractors, or through merchants and distributors.
- The project stage the role owns: creating, defending, converting or fulfilling a specification.
- The territory, the sales cycle length and the support structure around the seat.
Those distinctions matter because the roles are genuinely different jobs. If your example needs to separate a BDM, an Account Manager and a Specification Manager, my guide to technical sales roles in building products sets out what each one actually does.
What should you never paste into a public AI tool?
Confidential information should never go into a public AI tool. That rule is plain and it comes before any prompt.
Do not paste this:
- Candidate names, contact details or a full CV without approval.
- Protected characteristics, health information or anything about someone's personal circumstances.
- Reference notes, salary or package details, and internal briefs.
- Customer names, pricing, and non-public pipeline information.
Do not ask AI to decide this:
- Whether a candidate is suitable, honest or a "culture fit."
- A person's personality or motivation read off a profile.
- The final hiring decision.
AI can structure evidence and surface areas to probe. It cannot assess a person, and it should not be asked to. Judge people against observable role behaviours and working conditions, not inferred traits.
On research, treat every web result as a claim, not a fact. Require a source link and a date. If the tool cannot find something, the honest answer is "not found," not a guess. Company size, market position and "growing or struggling" labels are often stale, private or inferred, so check them against the company's own site, the job brief and the actual conversation before you rely on them. Employee-review posts are individual reports, not proof of what a workplace is like.
Hiring managers stay responsible for privacy, fairness, accuracy and the decision. Your organisation's policies, and Australian privacy and employment obligations, may require stricter controls than a public tool offers. If in doubt, use an approved tool and account setting, and keep people's data out.
The five-step workflow for both sides
- Define the role. Product, buyer, project stage, route to market, territory and technical threshold. State plainly that the person is a salesperson in the industry, not a designer or installer.
- Strip the inputs. Remove personal, sensitive and confidential information. Use a tool and account setting your organisation approves.
- Research public facts only. Ask for sources and dates. Accept "not found." Confirm anything important against first-party sources.
- Generate or rehearse. Write questions against observable role criteria, or practise answers out loud.
- Apply judgement. Check the output, remove anything biased or irrelevant, and decide what a human still has to ask and judge.
Same five steps whether you are hiring or interviewing. The value is that AI does the preliminary digging and lays it out quickly, which frees your time for the conversation, the relationships and the parts of the decision only a person can make.
For hiring managers: from a generic prompt to a useful one
The point of AI here is to customise questions to the actual role, so you can dig deeper and spot gaps a standard question set would miss. Compare two prompts.
Generic: "Give me ten interview questions for a BDM."
Context-rich: "I am hiring a Territory Manager selling cladding into commercial projects in Victoria. The route to market is specification through architects, then conversion with the builder. Generate interview questions that test how the candidate builds specifications with designers, defends a spec at tender, and manages a long project cycle. For each question, tell me what evidence a strong answer contains and what needs a follow-up. This is a salesperson in building products, not a designer or a builder."
The first gives you questions any manager could ask. The second, in my experience, is where AI earns its place. It probes the specific commercial work and shows you where to push. Three questions the context-rich prompt tends to produce:
- "Walk me through a specification you created from scratch. Who did you influence, and how did you hold it through to the builder?"
- "Tell me about a spec you lost at tender. What happened, and what would you do differently?"
- "How do you keep a project alive across an eighteen-month cycle without the pipeline going quiet?"
Read the answers against evidence, not polish. When someone claims specification experience, listen for whether they created and defended a spec or simply fulfilled one that already existed. Keep the follow-up areas neutral and evidence-based rather than framing them as red flags. For the substantive human question framework, my building products interview questions guide is the companion to this, with model-specific help in ChatGPT for building products sales and Claude for building products sales.
For candidates: research, rehearse and stay honest
Same method, other side of the table. Give the model the role and the company's public facts, ask it for the questions you are likely to face, and use it as a mock interviewer that asks one question at a time and gives direct feedback on each answer before moving on.
Two rules keep it useful. Use accurate, non-confidential achievements you can actually explain, not numbers you cannot stand behind. AI can organise the truth of what you have done. It should never invent it. And verify what the tool tells you about the company against their own website and the job brief before you repeat it in the room.
Which AI tool should you use, and how current is this?
Do not pick a tool by reputation. Pick one that can use the sources or files you need, under settings your organisation approves, and check its current documentation before you rely on a feature. Search, file access, connected sources and privacy controls change often and vary by product, plan and account. What a tool could do six months ago is not a safe guide to today.
Tool guidance checked: July 2026. Because this moves quickly, treat any tool-specific detail as something to re-check every six months.
Where to get the full prompts
The complete, copy-paste prompts for both audiences live in the AI interview-prep toolkit. If you want to build a wider AI workflow for the rest of your sales or hiring work, the AI toolkit is the hub. Use this page for the method and the safeguards, and the toolkit for the prompts themselves.
If you are hiring building products sales talent, or preparing for your own next move, get in touch. I run sales recruitment for the Australian building products industry. Happy to help, no obligation.
James