• Features
  • API & MCP
  • Blog
  • Pricing
  • Log in
  • Sign up

AI account research for B2B sales

How to run account and contact level research

Better pipeline comes from accurate messaging. Accurate messaging comes from understanding your account and contact. Twain researches prospects in real-time to build qualification and messaging strategy before writing a single outreach message

Research ProspectInspect Example
  • Sources remain attached
  • Unknowns remain visible
  • Your team reviews before delivery
Public-data exampleUltra
Johan Thelin*Director, Customer & Sales Enablement · Reuters News Agency
What Twain found

Johan highlighted Reuters’ MCP launch for AI newsroom workflows.

LinkedIn post • Jul 24
Recommended angleLead with seller readiness, not cold outbound.

His role is enablement; the research does not show that he owns prospecting.

*Randomly selected from public information. Neither Johan Thelin nor Reuters News Agency is affiliated with Twain, and neither is a Twain customer.

Twain interprets each signal against your campaign

A post, job change, or technology choice matters when it changes who you contact, what you say, or whether you should write at all.

  1. 01
    Company knowledgeYour ICP, offer, proof, and language your team already uses.
  2. 02
    Campaign briefThe campaign goal, audience, and the standard each message must meet.
  3. 03
    CRM contextHistory, ownership, and the facts your team already knows.
  4. 04
    Live web researchFresh evidence that changes who you contact and what you say.

One account brief your team can inspect

Twain turns scattered context into evidence your team can question, refine, and use.

A public-data example, researched in Ultra

Inspect the evidence and decisions behind the message

Move between Sequence, Research, Thesis, and Strategy to see how each decision follows from the available evidence.

Johan Thelin*Reuters News Agency
Ultra
Generated opening

Reuters MCP server

Hey Johan,

Your post on the Reuters MCP server launch describes the same bet we made at Twain. Our research agents run over MCP, so a sales team can pull deep prospect research into the tools they already use. Seeing an enablement director explain this kind of plumbing publicly made me think you'd want to look at the selling side of the same idea.

I'd bet account research and the first draft before a customer conversation still sit with each rep, and how good they are depends on how much time each rep had that day. Twain builds prep for them, a structured profile from CRM data and live web signals with a draft written from it, and it works inside your existing stack.

If that's worth a look, I'll build a profile and a draft for one of Reuters' target accounts so you can put it in front of your sellers and judge the output yourself.

Research

Signals that change the message

Johan leads customer and sales enablement at Reuters and shared the company’s MCP and AI video work.

  • Johan highlighted Reuters' launch of its Model Context Protocol (MCP) server, designed to integrate Reuters content directly into newsrooms' and media organizations' AI-driven workflows.
    LinkedIn post • Jul 24
  • Johan shared updates on AI-powered video capabilities at Reuters, including translation across seven additional languages alongside tools for speaker identification and time-coded transcription.
    LinkedIn post • Jul 24
  • Johan promoted hiring for a Commercial Data Analyst within the Reuters Commercial Excellence team in Gdansk, focusing on Salesforce, ETL workflows, and Power BI.
    LinkedIn repost • Jun 29
Limits that change the angle

What the research did not find

  • No evidence Johan owns an outbound or prospecting motion. His titles, bio, and activity show enablement, content, coaching, readiness, and onboarding. The material does not show a sequencer, campaign, or pipeline target under his name.Do not pitch his outbound campaigns. Frame research briefs, call openers, and pre-meeting preparation as tools for his sellers.
  • No evidence Reuters News Agency runs cold outbound. The snapshot shows a subscription and marketplace business with an account-heavy ICP, but no SDR teams or cold campaigns.Keep cold-outbound language out of the opener. Lead with research quality, seller preparation, and relevant messages.
  • No further evidence clears the bar, and Johan's stack beyond Salesforce is unknown.Do not name a specific sequencer. Mention Salesforce only as a cautious compatibility point.
What the evidence says together

Prospect narrative

  • Johan moved from retail sales in Norway to sales representative, trainer, and team manager at 3C Online in London. He joined Reuters in December 2021, became Senior Manager in April 2024, and was promoted in April 2026 to Director, Customer & Sales Enablement. He owns readiness, coaching, frameworks, onboarding, and content for a global sales organisation.
  • Reuters is productising AI through its MCP server and AI video translation across seven languages. Johan promoted both developments and supported a Commercial Excellence buildout involving Salesforce, ETL, and Power BI.
  • The overlap: Johan is championing MCP so newsrooms can bring Reuters content into AI workflows, while Twain's research agents use MCP with tools such as Claude and Cursor. His sellers now need strong preparation for increasingly technical, AI-focused customer conversations.

Persona fit

Persona fit — partial: enablement problem fits, SDR/BDR leader seat does not. His bio names tailored content, coaching, onboarding, and sales readiness, but the material does not show ownership of SDR or BDR prospecting targets, campaign preparation, qualification, outbound channels, or budget.

Why now

The strongest current angle is Reuters News Agency's Model Context Protocol server alongside the Commercial Excellence buildout. Lead with the MCP post, then connect Twain's research briefs and call openers to seller readiness.

Why us

Twain's Research Agent pulls CRM and live web evidence into a structured prospect profile, then writes from that evidence. The relevant outcome for Johan is better seller preparation for technical customer conversations.

Level alignment

Company-level and contact-level signals do not fully align. Reuters' strategy is not Johan's to set, and there is no evidence he owns outbound campaigns or tooling. Lean on Johan's own activity and enablement remit.

Risks

Johan judges seller communication for a living, so a templated or breathless AI-tool pitch will fail. The note must demonstrate research quality through his MCP post. Do not presume Reuters runs cold outbound, quote an unsupported reply figure, or name a sequencer based on inference.

Channel preference

LinkedIn is the strongest channel, with high confidence. Email is viable later, but LinkedIn is where he clearly stays active. He judges seller communication on this channel, so the note must work as the product demo.

Language and tone

  • Use enablement terms such as readiness, coaching, frameworks, onboarding, sellers, and outcomes.
  • Avoid growth-hacking language such as volume, blasts, spintax, deliverability, or booked meetings on autopilot.
  • Name specific outputs such as structured insights, why-now context, and call openers instead of saying AI-powered personalization.

Sequence strategy

  1. Lead with Johan's MCP server post. Connect Reuters' use of MCP in AI workflows to Twain's research agents, which use the same deployment pattern.
  2. Move to his expanded Director remit and enablement mandate. Frame Twain around research briefs, call openers, and pre-meeting preparation that his sellers use.
  3. Use the question to ask whether his reps will actually use the output. Name Batuhan Hazir and his report that Twain 'gets the messaging right instantly' without heavy prompting.
  4. Close with a researched profile and tailored draft to show his sellers, so a short conversation can be judged against the enablement roadmap.

*This prospect was selected at random. Twain generated the research and recommendations in this example from publicly available information. Neither Johan Thelin nor Reuters News Agency is affiliated with Twain, and neither is a Twain customer.

Choose how much research Twain does for each lead

Base researches and drafts, High adds qualification, and Ultra adds the Strategy that decides what to say, avoid, and lead with.

  • Fast

    Base

    Researches the account, contact, and relationship, then writes a sequence from your instructions.

    Sequence
    Included
    Research
    Included
    Thesis
    Not included
    Strategy
    Not included
    Best for
    High-volume email campaigns where every lead still needs a defensible reason to reach out.
  • Broader

    High

    Adds LinkedIn and YouTube activity, Absent signals, and a Thesis that explains why now and why you.

    Sequence
    Included
    Research
    Included
    Thesis
    Included
    Strategy
    Not included
    Best for
    Multi-channel outreach where the team needs more than a plausible opener before acting.
  • Deepest

    Ultra

    Runs the most searches and adds a Strategy for what to say, what to avoid, and how to order the sequence.

    Sequence
    Included
    Research
    Included
    Thesis
    Included
    Strategy
    Included
    Best for
    High-stakes campaigns where the draft needs a real point of view, not a generic prompt.
What Twain producesFastBaseResearches the account, contact, and relationship, then writes a sequence from your instructions.BroaderHighAdds LinkedIn and YouTube activity, Absent signals, and a Thesis that explains why now and why you.DeepestUltraRuns the most searches and adds a Strategy for what to say, what to avoid, and how to order the sequence.
SequenceIncludedIncludedIncluded
ResearchIncludedIncludedIncluded
ThesisNot includedIncludedIncluded
StrategyNot includedNot includedIncluded
Best forHigh-volume email campaigns where every lead still needs a defensible reason to reach out.Multi-channel outreach where the team needs more than a plausible opener before acting.High-stakes campaigns where the draft needs a real point of view, not a generic prompt.

Delivery

Move approved research and messages through the stack your team already uses

Twain imports context from your CRM and sources, then sends approved work through native integrations, API, MCP, and workflows.

  1. 01

    Twain

    Twain researches and qualifies the prospect, builds the Strategy, and gives your team a draft to review.

    The research layer

  2. 02

    Your CRM & sources

    Bring lead and account context in from systems your team already uses.

    Import from

  3. 03

    Your sequencer & destinations

    Send approved research and messages to your sequencer, CRM, or workflow.

    Export to

Customer results

The output should sound like someone did the homework

  • Reply rate

    12% positive

    Replies on cold outbound

    Before Twain our outbound outreach was very generic and did not tackle individual pain points. Now we send 1:1 deep researched emails and we see 12% positive reply rates, whilst we send highly specific messages on scale.

    In my 12 years of doing outbound, I have never seen this.

    Florian Teufel

    Demand Gen Lead @ Xentral

  • Message quality

    Sounds human

    Not a template, speaks to real pain point

    Most AI outreach tools just make it faster to send worse emails.

    Twain's contextual research actually digs into a prospect's situation, so the draft already speaks to their real pain points and sounds like a person who did the work, not a template.

    It's the only kind of AI worth using.

    Madeleine Cooper

    Founding GTM @ Cotiss

  • Time to launch

    25 minutes

    For an ABM campaign that took 6 hours

    Since using Twain, we have cut our ABM preparation time by up to 90%. Now it takes us 25 minutes instead of 6 hours to launch a campaign.

    Before, our emails were like any other cold outreach message. Now, every message is unique and hyper-personalized. They sound human and help us book demos with bigger logos.

    Jernej Samide

    Head of Global Sales @ Ingestro

Questions before you try Twain on a real account

Yes. Twain shows the sources, Signals, Absent signals, Thesis, Strategy, and draft so your team can review the work before it moves into the rest of your stack.

No. Twain researches, qualifies, and writes. Your CRM keeps the context, and your sequencer sends after your team approves the message.

That becomes part of the answer. Twain keeps missing evidence visible, narrows the angle, avoids unsupported claims, or recommends not contacting the lead yet.

Yes. New users get 500 free credits at signup, with no credit card required. The credits expire after 30 days, so you can test Twain on real prospects first.

Yes. Research one prospect, inspect the sources, fit, strategy, and draft, then decide what should move into your wider workflow.

Twain checks the account against your campaign, shows what it found and could not verify, then writes a message for your team to review.

See whether a prospect is worth contacting — and what to say

Research Prospect
Ask AI about Twain
Features
Pricing
API docs
Integrations
Help center
Blog
Privacy policy
Terms of service
Imprint
© Twain Technology GmbH