Prospect Research Engine

Inside this Project, Claude carries three things from chat to chat:

• Your ICP definition — so an account scored in week six is scored on the same criteria as one scored in week one.
• Your scoring rubric — including how you weight industry fit against trigger strength, so scores stay comparable.
• The shape of what has scored well — the patterns in accounts you have already researched, so the briefs get sharper as the Project accumulates work.

What it does not carry is the raw source material from previous accounts. Each brief is built from what you paste in that session. Re-paste anything you want re-analysed.

Reach for this Project when:

• You are working a target list and need every account researched the same way.
• You have a call booked and need a brief you can read in two minutes.
• You want to know whether an account is worth the outreach before you spend the effort.
• You need outreach angles anchored to something specific rather than a generic opener.

Use something else when:

• You need a single quick answer about one company — that is a chat message, not a Project.
• You are building a long-term picture of one named strategic account across many interactions — use Account 360 Builder, which is designed to accumulate rather than compare.
• You are writing the sequence itself — that is SDR Outbound Engine.

Per account, you provide:

• The company name, and the website URL if you have it.
• Pasted source material — website copy, recent news, LinkedIn company and people pages, job postings, funding announcements, earnings summaries, or filings.
• Optionally, the persona you intend to contact, if you want the angles aimed at a role.

The quality of the brief tracks the quality of what you paste. Three good sources beat a homepage alone.

Upload these once, when you create the Project.

The ICP definition is the one that matters most. Without it, scoring is guesswork dressed up as a number.

Claude accepts files up to 30MB each, and there is no fixed limit on how many you add, but everything in a Project's knowledge has to fit the context window. Fewer, sharper files beat a full archive.

Prepare files this way:

• Strip cover pages, legal boilerplate, and navigation chrome — they consume context and carry no signal.
• Prefer text formats. A clean export beats a scanned PDF.
• Name each file for what it contains, not when you exported it. Claude reads filenames.
• One subject per file. A merged reference document is harder to cite precisely.
• Date anything that goes stale, inside the file itself, so Claude can tell you when it is working from old material.

1. Create the Project. In Claude, create a new Project and name it for the territory or segment it covers — for example Prospect Research — Mid-Market Logistics. One Project per segment keeps the accumulated ICP context coherent.
2. Paste the Project Instructions. Copy the instructions from this page into the Project's custom instructions field.
3. Replace the bracketed fields. There are three. Fill all of them before you save — see the placeholder guidance below.
4. Upload the knowledge files. Start with the ICP definition and the product one-pager. You can add the rest later without disturbing what is already there.
5. Run your first session. Send the starting prompt with one account's source material pasted in, and check the brief against what you already know about that account. That calibration pass is what tells you whether your ICP definition is written clearly enough.

The bracketed fields below are placeholders you replace with your own details before you save the instructions. Leaving them unfilled produces generic output — that is the single most common reason a Project underperforms in its first week.

This Project has three:

• [FILL IN: Company Name] — your employer, not the prospect. It sets whose analyst Claude is acting as.
• [FILL IN: Your ICP Definition Summary] — two or three sentences naming the industries, company sizes, and signals that make an account a fit. The full document goes in knowledge; this summary keeps the criteria in front of Claude on every request.
• [FILL IN: Your Product One-Liner] — one sentence on what you sell and the problem it removes. Outreach angles are built against this, so a vague line produces vague angles.

Role

You are a senior B2B sales research analyst for [FILL IN: Company Name]. You turn raw company information into structured intelligence briefs that help a rep personalise outreach and prepare for calls. You work only from material the rep provides.

Operating rules

• Extract only facts explicitly stated in the provided materials. Never fabricate.
• Where information is missing, write Insufficient data. Do not guess.
• Flag uncertainty with [Unconfirmed].
• Never infer headcount, revenue, or funding from indirect signals without labelling the inference.
• Keep every output in the template structure so briefs stay comparable across accounts.

Module 1 — Company snapshot

Company name, HQ, founding year. Core business in two or three sentences. Employee count and revenue, each marked confirmed or estimated. Recent funding or acquisition activity with dates.

The full instructions continue with the ICP fit rubric, trigger detection, outreach angle ranking, the four commands, and the escalation rules.

This Project produces four deliverables, each in a fixed shape:

• Research brief — the full five-section template: company snapshot, ICP fit analysis with per-criterion scores, buying triggers detected, ranked outreach angles, and evidence-backed openers.
• ICP score — the fit analysis alone, when you want a fast qualify or disqualify decision before investing research time.
• Opener set — five cold email openers, each tied to a specific stated fact.
• Stakeholder map — named people with title, likely priority, and whether their relevance is confirmed or inferred.

A completed research brief. Meridian Freight Systems is fictional.

Company snapshot

Meridian Freight Systems. HQ Columbus, Ohio. Founded 2009. Mid-market freight brokerage software for regional carriers, focused on load matching and settlement automation. Roughly 240 employees per their careers page. Revenue not disclosed; Series C of $40M announced March 2026 [confirmed — press release].

ICP fit analysis

• Industry fit: 5/5 — logistics software, squarely in the named target vertical.
• Company size fit: 4/5 — 240 employees sits in the 150–500 band; slightly below the centre of the range.
• Tech stack alignment: 3/5 — job postings name Snowflake and dbt, so the data layer fits. No stated CRM, so integration fit is [Unconfirmed].
• Buying trigger signals: 5/5 — two strong, recent, independent triggers.
• Overall ICP fit: 17/20.

Buying triggers detected

• Series C closed March 2026, explicitly earmarked for "platform and data infrastructure" [confirmed — funding announcement].
• Six open roles on the data and analytics team, including a first Head of Revenue Operations [confirmed — careers page, retrieved from pasted material].
• A new VP Customer Success started in January 2026 [confirmed — LinkedIn].

Outreach angles, ranked

1. First RevOps hire. Hiring a first Head of Revenue Operations right after a raise usually means reporting is about to become someone's explicit problem. Strongest angle: recent, specific, and tied to a named role.
2. Post-raise infrastructure spend. The funding language names data infrastructure directly.
3. New CS leader. Weaker — a new VP is a real trigger but nothing in the material connects them to this problem, so the angle is inference.

Cold email openers

• "You're hiring your first Head of Revenue Operations — usually the point where someone finally owns the number that three teams have been reporting differently."
• "The Series C release called out platform and data infrastructure specifically, which is a more honest description of where the work actually goes than most funding announcements."

Insufficient data: current CRM, renewal cycle, incumbent vendor.