B2B AI Lead Generation

Why readiness comes before tools

AI lead generation does more of whatever your current process does. If targeting is vague, you reach more of the wrong people. If follow-up is slow, more leads go cold. If the data is messy, the mess spreads to every tool connected to it. So the useful first question is not which tool to buy but whether the foundations will hold the extra volume.

The assessment below covers nine foundations. Answer honestly for how things work today, not how they are meant to work.

The assessment

Assessment

B2B AI lead generation readiness

Answer each question for how your business works today. Yes means it is true and written down; partly means it is true but informal or inconsistent; no means it is not in place. Score 2 for yes, 1 for partly, 0 for no.

1. 1. Is your ideal customer profile written as rules someone else could apply, and have you counted how many companies match?

   AI can only target what is defined; the count tells you whether volume tactics are viable at all.

   YesPartlyNo
2. 2. Can you name the specific problem you solve and point to at least one proof of it (a result, a client, a demonstration)?

   Faster outreach with a weak offer produces faster rejection.

   YesPartlyNo
3. 3. Is your CRM reasonably clean: few duplicates, owners assigned, and a record of who has opted out?

   Every automated step reads from and writes to it; bad data multiplies.

   YesPartlyNo
4. 4. Do marketing and sales use the same written definitions of lead, qualified lead and opportunity?

   Without them you cannot tell whether AI improved anything.

   YesPartlyNo
5. 5. Does every new lead get a response from a named person within one working day?

   Extra leads are wasted if follow-up is already slow.

   YesPartlyNo
6. 6. Have you classified prospects as corporate or individual subscribers for UK email rules and recorded your lawful basis?

   Sole traders and some partnerships need consent or the soft opt-in; named contacts are covered by UK GDPR.

   YesPartlyNo
7. 7. Is your sending domain authenticated with SPF, DKIM and DMARC, with a working unsubscribe?

   Mailbox providers expect it, and automated volume exposes weak set-ups quickly.

   YesPartlyNo
8. 8. Do you have at least 90 days of baseline numbers for leads, qualified leads, meetings and opportunities by source?

   Without a baseline, any change is a guess.

   YesPartlyNo
9. 9. Is one person accountable for lead generation results and able to review AI output each week?

   AI needs an owner who checks its work and makes the calls it cannot.

   YesPartlyNo

Answered 0 of 9. Your result appears when every question is answered.

What each score range means

• 0 to 7: Fix the foundations first. AI would mostly accelerate existing problems. The gaps are in targeting, data or follow-up, which are cheaper to fix now than after scaling. Next step: Pick the two lowest-scoring questions and fix them before adding any AI tooling. Usually that means writing the ICP rules and setting a one-day response rule.
• 8 to 13: Ready for one focused pilot. The basics exist but some are informal. A narrow pilot will work if it avoids the weak areas. Next step: Choose one channel and one segment, automate research and enrichment first, keep a person approving every message, and compare against your baseline after a full sales cycle.
• 14 to 18: Ready to scale carefully. Targeting, data, follow-up, compliance and measurement are in place. AI can add capacity without adding chaos. Next step: Extend to more segments one at a time, keep weekly review of complaints and qualification rates, and move people time from research to conversations.

What the questions are really testing

The nine questions fall into four groups.

Targeting and offer (questions 1 and 2). These decide whether more reach is good news. A written, counted ICP also answers a question people rarely ask: how many prospects exist? If the answer is a few hundred, your approach should be closer to account-based selling than to automated sequences.

Data and definitions (questions 3, 4 and 8). These decide whether you can see what happens. A lead definition that changes every quarter, or a CRM where half the records have no owner, makes any improvement impossible to prove.

Follow-up and ownership (questions 5 and 9). These decide whether extra leads turn into conversations. In the inbound processing record, a 2025 build Jakub delivered before founding AfM, follow-up had slipped to "1-3+ days" as volume grew, and automation brought it back to "Same Day". That was a fix to question 5, and it mattered because volume had risen.

Compliance and sending (questions 6 and 7). These decide whether scale damages you. The ICO distinguishes corporate subscribers from individual subscribers such as sole traders, and UK GDPR still applies to named business contacts. Google's sender guidelines set authentication and spam rate expectations that automated sending can breach quickly. These questions tell you what to check; they are not legal advice, and a borderline case (a partnership, a contact collected long ago) is worth a qualified adviser's view.

If you scored low on one question only

A single "no" can matter more than the total. A "no" on question 6 or 7 is worth fixing before any sending, whatever the score, because the cost of getting it wrong falls on your domain and your reputation. A "no" on question 5 means the best first use of AI is routing and follow-up rather than generating more leads.

For the next steps after the assessment, the B2B lead generation playbook sets out the order of work, and automating without spamming covers questions 6 and 7 in detail.

FAQ

Why does the assessment not ask which AI tools we use?

Because tools are the easiest part to change and the least likely reason for poor results. The questions focus on the foundations that decide whether any tool helps.

We sell to very few accounts. Is AI lead generation still relevant?

Yes, but for research and preparation rather than volume. With a small market, use AI to brief people on each account and keep outreach personal.

How often should we retake the assessment?

Quarterly is sensible, and whenever you add a new channel or segment. Scores often drop when a team grows, because informal habits stop scaling.