M&A Deal Sourcing with AI: Finding Off-Market Targets

The problem: proprietary deal flow is manual work

Buy-side mandates — private equity, family offices, strategic buyers — live on proprietary deal flow. The most interesting targets aren’t on any platform: owner-managed SMEs facing succession, often without a maintained website, without a LinkedIn presence, without an M&A advisor. The DACH region makes this structurally worse: by common estimates, around 190,000 companies will face the succession question by 2027 — a large share of them will never issue a sale mandate, but will (if at all) respond to direct, discreet outreach.

Classically, this sourcing is expensive: analysts searching registers, maintaining lists, researching owners. This exact work can today be built as an AI pipeline. We run such a pipeline ourselves — here is how it’s built, including the honest numbers.

The 6-stage pipeline

1. Source        → Google Maps scraping by industry + region
2. Qualify       → AI research: ICP fit + succession signals, score 0-100
3. Find owner    → legal notice, commercial register, LinkedIn
4. Email         → enrichment waterfall + validation (only 'valid' gets contacted)
5. Copy          → LLM-personalized per owner, max. 90 words
6. Send          → cold email infrastructure with a clean deliverability setup

Step 1 — Sourcing via Google Maps

Industry keywords plus region (say, “mechanical engineering,” “CNC manufacturing,” “toolmaking” in Baden-Württemberg) yield hundreds of entries per search run, with name, address, phone number, website, category, and reviews. Why Maps instead of a company database: coverage of micro-SMEs without a web presence is markedly higher, and the data is more current — closed businesses drop out instead of haunting databases for years.

Step 2 — Qualifying with AI research

Each company runs through a research pass with a fixed schema:

• ICP fit: industry, estimated headcount, region, specialization
• Succession signals: owner age (commercial register, legal notice, LinkedIn), naming convention (“Inhaber,” “e.K.,” family name = company name), website age, “since 19xx” in the branding
• Disqualifiers: subsidiary of a group, insolvency, already sold

The result is a score from 0-100; above a threshold (60 for us) the target moves to the next stage. This is the same logic as in signal-based outbound — except here the signal is “succession likely” instead of “currently buying software.”

Steps 3 & 4 — Identifying the owner, validating email

For qualified targets, the owner is resolved by name (legal notice, commercial register via northdata, LinkedIn). The email address comes from an enrichment waterfall followed by validation — unverified addresses are never sent to. In our campaigns this makes the difference between 0.4% and 7.7% bounce rate; the latter ruins the sending domain. For micro-SMEs, the fallback is often info@domain with a personal salutation — unglamorous, but functional. Fundamentals in the deliverability guide.

Steps 5 & 6 — Copy and send

Outreach decides everything in this niche:

• Max. 90 words, no links, no tracking — a discreet note, not marketing
• The hook is the concrete succession signal (“Your business celebrates its 40th anniversary this year…”)
• Tone: discreet, at eye level, no broker talk, no valuation promises
• CTA: a no-obligation conversation — nothing more

Sending runs through a clean cold-email infrastructure (dedicated sending domains, conservative limits, warmup — details in the Instantly guide).

Real numbers: our buy-side campaign

Our own campaign in the DACH region, targeting PE firms and family offices (succession outreach), as of June 2026:

Metric	Value
Emails sent	1,756
Leads contacted	690
Unique replies	11 (1.6%)
Opportunities	5
Bounce rate	3.1%

The most important learning isn’t in the table: buy-side reacts sluggishly to cold email — PE firms and family offices are bombarded with deal offers daily. The sell-side (owners directly) is the more interesting lever within the same pipeline, because the timing signal — the impending succession — supplies the hook there, and practically no one approaches them systematically. And yet even the buy-side variant pays off: a single mandate carries the entire pipeline for years.

The cost reality

The pipeline is also an entirely different game economically than classic analyst sourcing: Maps scraping costs roughly half a cent per entry, AI research runs for cents per company, enrichment costs 10-30 cents per validated contact depending on the provider. A qualified, personally addressable list of 500 targets is thus produced for a fraction of a single analyst-day — the bottleneck shifts from searching to leading the conversations.

Conclusion

Off-market deal sourcing is no longer secret knowledge — it’s a process: public data sources, AI qualification by succession signals, validated owner contacts, and discreet, personalized outreach. The numbers are small and valuable rather than large and cheap — 5 opportunities from 690 contacts is a very good deal in this asset class. Anyone still sourcing exclusively through advisor networks and platforms in 2026 is competing for the deals everyone else already sees.

Next step

The complete pipeline — sourcing, AI qualification, owner research, and personalized outreach with human sign-off — can be tried directly as a system: Start your free trial — 4 weeks free, no credit card required.