Back to vendors
F

Floworks

Also known as: Floworks AI, Alisha AI SDR, ThorV2, One Floworks Technologies

Visit site
Entry pricePRO pay per meeting (sales) · ULTRA $498/mo (1,000 contacts researched, 7,500 reachouts) · ADVANCED custom · CRM integration on annual plansFull pricing detail

AI sales employee suite led by Alisha, an autonomous AI SDR on the proprietary ThorV2 model that researches prospects, personalizes email and LinkedIn outreach, and books meetings.

Floworks is a Y Combinator backed sales automation company building a suite of AI employees for B2B revenue teams.

Its flagship is Alisha, an autonomous AI sales development representative designed to perform the full outbound SDR job: defining the ideal customer profile, finding prospect lists, researching each prospect, writing personalized emails, following up on prospect behavior and engagement, handling replies and booking meetings before updating the CRM.

Around Alisha sit other role based agents, including Jesse for research and lead sourcing across more than 700 million contacts on intent data, Linda for LinkedIn outreach, and Sam for phone outreach, which the vendor lists as coming soon.

The technical differentiator is ThorV2, Floworks' proprietary function calling model, which the company describes in an Arxiv paper and positions on reliability and reduced hallucination. At campaign launch the customer uploads an FAQ and describes the product, and the agents work from that material and a generated playbook rather than inventing answers. Prospecting uses a chat style interface where an operator can request a precise persona in plain language and the research agent assembles the list.

Outreach is omnichannel across email and LinkedIn, so if a prospect does not reply to email, Alisha follows up on LinkedIn. Rather than writing every email from scratch, it starts from a library of proven cold email recipes and personalizes each copy per lead, and it books meetings with warm slot suggestions rather than a bare scheduling link. Unlimited email warmup is included.

Floworks publishes three plans: PRO, billed per meeting through sales for agencies; ULTRA at $498 a month, self serve, with research on 1,000 contacts and 7,500 email and LinkedIn reachouts; and ADVANCED, custom for teams of ten or more SDRs. CRM integration requires annual billing. The compliance page heads SOC 2 Type 2 and ISO 27001 as audits currently under way, so a buyer needing an attestation should confirm its status directly.

Vendor details

Canonical URL

https://www.floworks.ai

Category

GTM / revenue agent

Subcategory

AI SDR (suite of sales AI employees)

Funding status

Independent, Y Combinator backed (One Floworks Technologies, Inc.). Builds a suite of AI sales employees led by Alisha, an autonomous AI SDR, powered by its proprietary ThorV2 function calling model, which the company benchmarks above GPT-4 and Claude 3 on function calling accuracy.

Company status

independent

Use cases & customers

Primary use cases

AI SDR outboundprospect researchpersonalized email and LinkedIn outreachobjection handlingmeeting scheduling

Target customers

startups and SMB sales teamsB2B SaaSmid-market

Deployment options

SaaS

Integrations

Built on the ThorV2 function calling model and integrates with email (Gmail, Outlook), calendars (Calendly, Google Calendar), HubSpot CRM, and LinkedIn, with a contact database of hundreds of millions of records for prospecting.

In practice

You are a founder with no SDR team. You upload your sales docs and connect your inbox, and Alisha researches prospects, writes personalized emails, answers their questions from your materials, and books meetings 24/7.

You need a precise lead list fast. You ask in plain language for VPs of Sales at UK healthcare SaaS that recently signed an NHS deal, and the research agent assembles it.

A prospect ignores your emails. Alisha automatically follows up on LinkedIn, and when they reply with an objection, it handles the response and proposes warm meeting slots rather than a cold scheduling link.

Agentic Index coverage score

5.5 / 14 capabilities · 39%

Integrations & Tool Calling Partial

A short fixed list of connectors with no way to add your own. Named: Gmail and Outlook for sending, LinkedIn for social outreach, Calendly and Google Calendar for booking, and HubSpot with Salesforce for CRM, the latter gated to annual plans. The agents take authenticated action rather than reading alone: they send from connected mailboxes, act inside LinkedIn, book into calendars and write back to the CRM. No public API, no webhook and no custom connector is documented anywhere on the site, so a buyer whose stack is not on the list has no documented path.

Sourcefloworks.ai/how-it-works and site navigationread 2026-09-01

Workflow Orchestration Partial

The platform carries a documented five-step motion end to end: connect mailboxes and LinkedIn with warmup, define the ICP and generate leads, research prospects and write personalized messages, launch the campaign with an uploaded FAQ and generated playbook, then run automatically sending messages and booking appointments.

Several named agents divide the work between them, with Jesse researching, Alisha and Linda writing and sending across email and LinkedIn, and Dan handling CRM updates.

The control flow itself is not documented; no branching or conditional paths, no retry or error handling and no routing surface the customer can inspect or configure appears on any page, and the reply-triggered escalation from email to LinkedIn is described in outcome terms rather than as configurable logic.

Sourcefloworks.ai/how-it-worksread 2026-09-01

Knowledge Grounding & RAG Full

The buyer points the agent at their own material through a documented ingestion step. At campaign launch the customer uploads an FAQ and describes their product, and the platform produces an AI-generated personalized playbook the agents then work from, so prospect questions are answered from material the customer supplied rather than from the model alone. Prospect-side research runs separately across the web and a contact database. The FAQ upload is the only documented ingestion route, and no other document types are named.

Sourcefloworks.ai/how-it-works campaign launch and personalization stepsread 2026-09-01

Human Oversight & Guardrails Not documented

Autonomy is the stated design and no approval step sits in front of it. The vendor's own five-step workflow ends by telling the customer to let the system run automatically, sending messages and booking appointments, and the product is marketed as a fully autonomous AI SDR. What the customer sets is upstream: the ICP, the product description, the uploaded FAQ and the generated playbook. Nothing lets a buyer mark actions as requiring sign-off, name who signs off, or hold a message pending review, and no draft queue or approval surface appears in the workflow or the site navigation.

Sourcefloworks.ai/how-it-works automate and monitor stepread 2026-09-01

Security, Identity & Governance Partial

One attestation the vendor does not contradict, and no access controls: the compliance page shows ADA CASA Tier 2 as certified. SOC 2 Type 2 and ISO 27001 are disputed between the vendor's own pages, the compliance page heading them as audits currently under way while the home page calls the company SOC 2 Type II, ISO 27001, GDPR and HIPAA compliant, so neither can be relied on yet. No SSO, roles or permissions the buyer operates are documented.

Sourcefloworks.ai/complianceread 2026-10-01

Observability & Auditability Partial

Campaign-level monitoring is documented as the fifth step of the product's own workflow, automate and monitor, and the site reports open rate, response rate, click-through on opened emails and scheduling rate as the measures a customer watches. The Sage agent is positioned for sales data analysis, pipeline forecasting and campaign optimization. That is aggregate reporting; nothing documents a per-run trace showing which sources the research agent consulted, why a particular message was written, or what each agent did in sequence, and no execution log or audit surface is documented.

Sourcefloworks.ai/how-it-works and homepage metricsread 2026-09-01

Memory & State Persistence Partial

State persists across a campaign rather than within an exchange: the customer uploads an FAQ and product description at campaign launch and the agents draw on it thereafter, lead lists are uploaded or synced and held, an AI-generated playbook is produced and reused across the campaign, and reply history carries the follow-up sequence forward from email to LinkedIn. No memory layer with a stated scope or lifetime is documented; nothing says what is retained between campaigns, for how long, or whether it is scoped per campaign or per workspace.

Sourcefloworks.ai/how-it-works campaign launch stepread 2026-09-01

Deployment & Data Residency Not documented

Delivered only as a cloud native SaaS platform with no self host, on premises, or in VPC deployment option documented.

Sourcefloworks.airead 2026-10-01

Prebuilt Agents, Templates & Packs Partial

Packaged assets ship in two forms. Named role agents aimed at named jobs: Alisha for outbound sending, Jesse for research and lead sourcing, Linda for LinkedIn, Sage for RevOps analytics and forecasting, Dan for CRM hygiene and inbox management, with Sam for phone calls listed as coming soon. And an Email Playbooks section published in the site navigation, alongside an AI-generated playbook produced per campaign. The agents are the product's components rather than a catalog a buyer selects from, and no adoptable templates are documented inside the Email Playbooks section.

Sourcefloworks.ai navigation and how-it-worksread 2026-09-01

Triggers & Channel Coverage Full

No person has to start each step. Jesse finds prospects on intent signals, and Alisha automatically follows up based on prospect behavior and engagement, with a missed email reply moving the prospect to a LinkedIn touch.

Sourcefloworks.airead 2026-10-01

Model Flexibility & Routing Not documented

Runs on the fixed proprietary ThorV2 model with no user facing model choice, bring your own key, or routing gateway.

Sourcefloworks.airead 2026-10-01

APIs, SDKs & MCP Extensibility Not documented

No public API, SDK, or MCP server is documented; ThorV2 is described in a research paper but not exposed as a developer platform.

Sourcefloworks.airead 2026-10-01

Testing, Debugging & Optimization Not documented

No user facing agent testing, debugging, or evaluation product is documented; the Sage agent provides campaign analytics rather than an evaluation framework.

Sourcefloworks.airead 2026-10-01

Browser & Computer Use Partial

A dedicated LinkedIn agent, Linda, runs connection requests, messages and meeting scheduling inside LinkedIn alongside the email channel, with LinkedIn connected and warmed at setup. That is the agent acting inside an interface the vendor does not control. The vendor's own pages do not state how those actions are executed. Everything else runs over connected mailboxes, calendars and CRM, which are programmatic routes.

Sourcefloworks.ai/how-it-works setup and personalization stepsread 2026-09-01

The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded

Pricing

PRO pay per meeting (sales) · ULTRA $498/mo (1,000 contacts researched, 7,500 reachouts) · ADVANCED custom · CRM integration on annual plans

monthly subscription on ULTRA, pay per meeting on PRO, custom on ADVANCED

Included quota

ULTRA: research on 1,000 contacts and 7,500 email and LinkedIn reachouts a month. ADVANCED: 5,000 plus contacts and 40,000 emails and LinkedIn prospecting a month, dedicated account manager. CRM integration on annual plans.

What is public

The ULTRA price ($498 a month) and its allowances are public; PRO is pay per meeting and ADVANCED is custom, both through sales.

Billing mechanics

Monthly or annual billing on ULTRA, with CRM integration only on annual; PRO is billed per meeting; ADVANCED is custom.

Cost watchouts

CRM integration on paid plans requires annual billing, and PRO charges per meeting at a rate set with sales.

Variable cost rationale

The Ultra plan is a flat monthly subscription, so cost is predictable, but a per-meeting (pay-per-meeting) component on at least one tier means cost can scale with booked meetings, and CRM sync requires an annual commitment. Unlimited warmup is bundled, but data verification (to avoid wasted volume) is a separate operational cost.

Additional watchouts

CRM sync requires an annual commitment, and only ULTRA carries a published price; the PRO per-meeting rate and ADVANCED pricing come from sales.

Overage / add-ons

Plans scale by channel coverage and seat/volume; a per-meeting (pay-per-meeting) component applies on at least one tier.

Sales call required

No, self serve available

Free / trial

No free plan or trial on the pricing page

Lowest paid plan

ULTRA $498 a month

Commercial notes

Differentiated by the proprietary ThorV2 model and positioned for startups on ULTRA and agencies on PRO.

Key ambiguities

Only ULTRA carries a published price; the PRO per meeting rate and ADVANCED pricing are set with sales.

Cancellation / refund

Ultra can run month-to-month; CRM sync requires an annual commitment. Other terms not detailed.

Support SLA / resale

24/7 support reported; demo/intro-call onboarding; claims a working setup in about a day.

Missing data

The PRO per-meeting rate, ADVANCED pricing and detailed plan quotas are not published.

Agentic Index verified 2026-10-01

Alternatives to Floworks

The closest documented capability profiles to Floworks among GTM and revenue agents tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.

  • Aomni5.5 / 14Adds documented Model Flexibility & Routing and APIs, SDKs & MCP Extensibility
  • AnyBiz4.0 / 14Fuller documented coverage on Triggers & Channel Coverage
  • Regie.ai7.0 / 14Adds documented Human Oversight & Guardrails and Testing, Debugging & Optimization
  • BambooBox5.5 / 14Fuller documented coverage on Workflow Orchestration and Prebuilt Agents, Templates & Packs
  • Closely6.5 / 14Adds documented Human Oversight & Guardrails and Deployment & Data Residency, among others
  • Connected-Stories5.5 / 14Adds documented Human Oversight & Guardrails and Testing, Debugging & Optimization

Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded

Head to head

Contact us

Found a vendor we missed? Have feedback on the index? We'd love to hear from you.