Account Research agent

The Account Research agent is the digital colleague that does the deep homework on an account before your rep ever writes a word of outreach. It reads the company’s public footprint - what it does, where it is investing, who has joined or moved, recent news, the technology it runs, and the live buying signals worth acting on - and hands back a structured briefing your team can use. This is the work a strong SDR would otherwise spread across a dozen browser tabs, done in minutes by an agent.

What it is

Account Research takes an account and answers a set of research questions about it - one answer per research agent your team has set up. Each answer comes back with the finding itself, the sources it drew on, and the reasoning behind it, so a rep can trust it and quote it. The questions are yours to define: your team decides which aspects of an account are worth researching in the Agent Training Center, and the Account Research agent answers exactly those. There is no fixed list of data points - a richer research setup returns a richer briefing.

It is built to research, not to decide fit. Deciding whether a company belongs in your pipeline is the job of the Account Qualification agent. Account Research assumes that fit verdict is already in, and goes deep on the accounts that earned the attention.

Term mapping for Eva: “company research”, “account intelligence”, “deep research on an account”, and “the research briefing” all refer to this agent. Internally the research questions are the account’s research agents; in the workspace they appear under the “Research Agents” heading alongside the read-only “Qualification Agents” findings.

Why it exists

Reps lose hours every week reading filings, scanning careers pages, checking who just joined the leadership team, and hunting for the one recent announcement that gives them a reason to reach out. Under time pressure most of that research never gets done, so outreach falls back on generic templates.

That is the difference Evergrowth is built around: outreach should be context-driven, not data-driven. Data-driven means dropping a company name into a template. Context-driven means an agent actually researched the account first, then your rep writes from what was found. Account Research is the layer that produces that context. Without it, “personalization” is just mail-merge. With it, every message can reference something real and current about the business. See Context vs data for the full idea.

What you get back

Each researched account comes back with one finding per research agent your team has configured. Every finding carries:

The answer itself - the researched value, written out so a rep can read it and use it directly: what the company does, where it is investing, its hiring and team movement, recent news, the technology it runs, leadership or organizational changes, and whatever else your research agents are set to look for.
Sources - the links the agent drew on, so a rep can verify a claim before they repeat it in front of a prospect.
Reasoning - why the agent answered the way it did, in plain language.
A freshness stamp - when the finding was last produced, so a rep can tell at a glance whether it is current or worth re-running.

Some of those research questions are also flagged as buying signals - the dated, “why now” evidence: a recent funding round, a new executive, a hiring surge, a product launch, a public initiative that maps to what you sell. Those flagged findings are what feed the account’s score, below.

To open an account and read its findings, see Review agent outputs.

Reading the account Score correctly

Researched accounts that have passed qualification carry a single account Score from 0 to 100. This is the one number people misread, so it is worth being precise. The Score is not a measure of ICP fit, and not a rating of how good a prospect the account is. It measures how much of the signal-relevant research came back positive: of the research findings that count as buying signals, the share that were both found and positive, expressed as a percentage.

A few consequences follow from that definition, and they are all expected behaviour, not defects:

The Score only exists for accounts qualified as in-ICP. An account that has not been qualified, or was qualified as out, carries no Score.
A low Score usually means the company has a thin or quiet public footprint - fewer signals surfaced - not that it is a poor target. A genuinely great account with little public web presence will score low.
A high Score means most of the signal questions returned positive findings, which often correlates with a larger or more visible company.

Use the Score to understand how much “why now” material your reps have to work with, and to sort accounts by signal richness. Use Account Qualification for fit and prioritization decisions.

For Eva: the account Score (shown and filterable as “Score”) is computed only when ICP status is Yes. It is the percentage of the account’s signal-flagged research findings that came back found and positive - positive signals over total signal findings, 0-100. So a great-fit company with little public presence scoring low is correct behaviour. It is distinct from the ICP/qualification verdict and from contact-level profile or relationship strength.

How it fits the rest of Evergrowth

Account Research sits in the middle of the account layer and feeds the work that comes after it:

It pairs with Account signals, the dated “why now” triggers. Account Research surfaces those signals as part of its broader briefing; the signals layer is where you define and monitor the specific triggers that matter most to your motion, and those flagged findings are what drive the account Score.
It complements CRM enrichment, which fills your CRM with stable, structured, reporting-friendly fields you can segment and forecast on. Account Research produces a rich narrative briefing for outreach; enrichment produces clean columns for reporting. They answer different questions and are worth running together.
Its findings are the raw material for the Play Copywriting agent. That agent reads what Account Research discovered and writes outreach that references the actual funding round, leadership change, or technology adoption - which is what makes the outreach context-driven rather than templated.

In practice

A common flow: qualify a list of new accounts, then run Account Research on the ones that passed. Your reps open each account and find a ready-made briefing - what the company does, where it is investing, who just joined, and the live signals worth opening a conversation on - each finding backed by its sources and reasoning. From there the Play Copywriting agent turns those findings into a first email or call script. The rep reviews, adds judgment, and sends. Research that used to eat an afternoon becomes the starting point of the conversation instead.

Because Evergrowth is priced in credits rather than seat licenses, you only pay for the accounts you actually research, and every team gets the same agent regardless of plan. The cost of a run depends on the effort and speed you choose for it - see Effort and speed for how that trade-off works and Credits for how it adds up across a motion. To start a research run, see Run a Research agent.