AI Marketing Client Results: Two B2B Engagements

What this page includes, and what it leaves out

This page presents only AfM's published case records for business-to-business clients. Each figure is given as the record states it, with its timeframe and what was measured. Three rules apply:

• • No totals. Figures from different engagements measure different things (cost per lead, leads processed, estimated savings), so adding them up would produce a number that means nothing.
• • Observed is not the same as caused. Where a record shows a change, this page says so. Where the record does not isolate what caused the change, this page says that too.
• • Client work only. AfM's founder-owned brand Harmonance is not a client and is not counted here.

Two client engagements qualify: an anonymised global technology platform (paid media, then lead-processing automation for the same client) and Allegiance Industries (an automated prospecting system). One timing note: the technology platform work (both halves) was delivered by AfM's founder as a freelance consultant, almost entirely before AI for Marketing Ltd was incorporated in August 2025. The Allegiance work was a one-off system build AfM delivered in 2026; it shows the kind of system work behind the managed service rather than a separate product on sale today.

The results at a glance

Published B2B case figures, as recorded

Figure	Engagement	Work	Timeframe	What was measured	Type of evidence
97% lower cost per lead: $270+ to $8.86	Global technology platform (anonymised)	Paid media, LinkedIn and Meta, DACH markets	Q4 2024 to Q3 2025, about 11 months	Cost per lead, first quarter (LinkedIn only) against final quarter (blended across Meta and LinkedIn)	Observed change; causes not separated
1,700+ total leads; 1,606 in Q3 2025	Global technology platform (anonymised)	Paid media	As above	Leads generated from ad forms	Observed volume
$7.12 best segment cost per lead	Global technology platform (anonymised)	Paid media	Q3 2025	Cost per lead in the best audience segment	Observed
$110K+ ad spend managed	Global technology platform (anonymised)	Paid media	About 11 months	Total media budget managed	Self-reported total; a cost, not a result
1,600+ leads processed in the peak quarter	Global technology platform (anonymised)	Lead processing and enrichment automation	2025, peak quarter	Inbound leads processed by the automated system	Self-reported; primary source not public
Response time: 1 to 3+ days before, same day after	Global technology platform (anonymised)	Lead processing automation	2025	Time from form submission to follow-up	Self-reported before-and-after
Up to 41 companies a day; 17 data fields per lead; 5 platforms connected	Allegiance Industries	Automated lead prospecting system	2026	System configuration	Configuration, not an outcome
Estimated 15 to 25 hours a week of manual prospecting replaced	Allegiance Industries	Automated lead prospecting system	2026	Sales team research time	Estimate
Client estimates savings of $9,000 to $10,000 a year	Allegiance Industries	Automated lead prospecting system	2026	Annual cost saving	Client's own estimate

Engagement 1: paid media for a global technology platform

The situation. The client's division for Germany, Austria and Switzerland needed to recruit agencies, entrepreneurs and digital professionals into a creator network programme. The first quarter of the engagement ran on LinkedIn only and produced leads at $270 or more each.

What was done. According to the published record: campaigns were run across LinkedIn and Meta, with a weekly optimisation cadence, regular creative testing, separate audience segments (agencies, freelancers, entrepreneurs, business owners and streamers), and budget split across the three countries on response data rather than population.

What was observed. Cost per lead fell from $270+ in the first quarter to a blended $8.86 in the final quarter, a 97% reduction. In Q3 2025 alone, 1,606 leads were recorded, and the best audience segment reached $7.12 per lead. Over the engagement, 1,700+ leads were generated on $110K+ of managed spend.

What the record does not show.

• • How much each change contributed. The starting figure is LinkedIn only; the final figure is a blend across Meta and LinkedIn, and the record says Meta delivered leads at a fraction of LinkedIn's cost. Much of the fall therefore reflects the change of platform rather than better LinkedIn performance. The record attributes the result to platform diversification, testing and segmentation together, without separating them.
• • Lead quality. No lead-quality or programme enrolment figures are published, so a low cost per lead should not be read as a low cost per customer.
• • A like-for-like baseline. The engagement built the account from its own first quarter; there is no comparison against what another approach would have achieved over the same period.

Engagement 1, continued: automating lead processing for the same client

As lead volume grew, manual handling of each form submission became the bottleneck. The automation record describes a system in which a form submission triggers an AI enrichment agent, the enriched record is written to the CRM with duplicate checks, and high-priority leads trigger an alert to the sales team.

What was recorded. Response time moved from 1 to 3+ days to the same day, and the system processed 1,600+ leads in its peak quarter. Both figures are self-reported and their primary source is not published, so treat them as the provider's account rather than independently verified measurements. This is the same client as engagement 1, not a separate win.

Engagement 2: automated prospecting for Allegiance Industries

The situation. A facility services company whose sales reps researched prospects by hand across two target sectors, manufacturing and education.

What was done. Per the Allegiance record, five platforms (Google Sheets, Make.com, a Relevance AI agent, Apollo and Zoho CRM) were connected into a daily prospecting run. It runs each morning without manual research; the sales team's only task is to add target companies to a sheet. It is configured to process up to 41 target companies a day and to map 17 data fields per lead into the CRM, with duplicate checks.

What was measured. The system is estimated to replace 15 to 25 hours a week of manual prospecting, and the client estimates it saves them $9,000 to $10,000 a year. Both are estimates, the second the client's own. This engagement was a system build rather than paid media, so there is no cost-per-lead figure for it.

Records that are not included, and why

• • ToastPal (a consumer AI product sold to individuals) and Mumshandmade (an e-commerce brand) are published client results but not B2B.
• • HEAVENSAKE is a consumer drinks brand; its record describes a content agent rather than measured marketing results.
• • Harmonance is AfM's founder-owned brand, not a client.

All are on the case studies page for readers who want them.

How to read any provider's case studies

Use the same tests on AfM and on anyone else:

1. • Is there a named or clearly described engagement, a timeframe and a stated measure?
2. • Is the baseline like-for-like, or does it compare different channels, periods or definitions?
3. • Is the figure observed in a platform, estimated, or the client's own estimate?
4. • Does the provider add up figures from different clients into one headline?
5. • Are founder-owned or internal projects presented as client work?

If you want to see what the same approach would look for in your business, the free written breakdown uses your public marketing and any figures you choose to share. For how the managed service runs day to day, see how it works.

FAQ

Why is the global technology platform anonymised?

The client is not named on AfM's site. The published record describes the sector, markets and programme so the work can be understood without identifying the organisation.

Are these results typical of what a new client should expect?

No single result predicts another business's outcome. Starting points, budgets, markets and offers differ. The records show what was done and measured in specific engagements, not a forecast for yours.

Why are the paid media results described as about 11 months?

The engagement ran from November 2024 (the fourth quarter of 2024) to September 2025 (the third quarter of 2025). Counted by calendar month that is 11 months, so both this page and the case record say about 11 months rather than a year.

Can I speak to one of these clients?

Ask during a call. Whether a reference is possible depends on each client's permission, which has to be requested case by case.