# Felt source alignment

October 31, 2026. Internal preparation record for the Felt brand review. This records what was actually read, what supports the proposed copy, and which boundaries matter. It is not a claim that a rebrand or new positioning has been approved.

## Controlling user correction after the source read

Aaron explicitly said: **“remove the AI recruiter framing. felt is the agentic platform for recruiting teams”.** Current accepted wording: **“Felt is the agentic platform for recruiting teams.”** This supersedes the earlier one-AI-recruiter interpretation for the rebrand. The Language of Alex finding below remains an accurate record of what that earlier source says, not an instruction to override the new user direction. Products remain distinct abilities/products within the platform; the resourceful crew expresses agents across the journey. Capability and rollout boundaries still apply.

## Latest user decision — October 31, 2026

**“Felt is the agentic platform for recruiting teams.”** The user explicitly directed removal of the AI recruiter master framing. This current instruction takes precedence over the older source positioning described below. The review now describes a platform whose agents work across the candidate journey. The established AI Interviewer product name remains. The task/process/operations ladder, implementation partnership and capability boundaries still apply. Source descriptions below are retained as evidence, not treated as current branding instructions where they conflict with this decision.

## Sources and freshness

Read [Product Home](https://app.notion.com/p/3bc49e23bd4b80f8b679e96fc01050b7), the [Alexandria Text index](https://app.notion.com/p/3ca49e23bd4b8195a2f8e95c7bdeaced), and the current [Alexandria database](https://app.notion.com/p/3d749e23bd4b80c088a8e993ae16aeb7), then opened the actual entries below. Where the older Text copy and newer database mirror differ, this review uses the newer database entry. Merely reading the index was not treated as product research.

These pages are GitHub mirrors. Their synced and edited dates describe the source snapshot, not fresh verification of production deployment. Top-level Notion page verification was unverified; individual customer-safe sections sometimes name Dan Kaufman as approver. Those are distinct states. “Uploaded,” “GA,” a named section approver, and universal availability are not interchangeable.

The parent agent freshly read the [supplied messaging Google Doc](https://docs.google.com/document/d/1CLXrUpeFLDwJwlib2wbjpfxWZxzrf4nMq9PaYKd_afY/edit?tab=t.0#heading=h.4nwyxmiwtgob). Its positioning is explicitly **proposed**, not an accepted final brand platform. The original document remains linked; the summary below records its relevant direction and scope.

## Actual entries read and what they support

| Source | Relevant support for the story | Boundaries and source status |
| --- | --- | --- |
| [The Language of Alex](https://app.notion.com/p/3d749e23bd4b8198baa3ddb826770d22), edited Oct 27 | Position one AI Recruiter. Describe completed work and value. Capabilities are connected skills of that recruiter. Distinguish the product's actions from the human implementation team's partnership. | Do not shrink the company to AI interviews, gender the product, or lead with new queues and artifacts for people to process. Avoid calling Alex's work an “assessment.” Name the customer system, not the connector vendor behind it. |
| [AI Interviews](https://app.notion.com/p/3d749e23bd4b81d1977bcd0dd0465b7e), edited Oct 27 | Role-specific interviews over supported voice and text channels; follow-up questions; results with transcript and reasoning; conversation context supports later work. | Listed GA; customer-safe core sections name Dan as approver. Interview guide and configuration matter. The broader brand should not imply universal hiring quality or that interview scores independently make the hiring decision. No customer proof or metrics are imported into this draft. |
| [Resume Screens](https://app.notion.com/p/3d749e23bd4b81529fead10559722c26), edited Sept 23 | Read applications against the customer's requirements; explain the evidence behind the review; write the result into the customer's system. | Listed GA with company/workflow configuration; source prose sections remain unapproved. This is application review, not a promise to parse every field or bypass stated credentials. The team and configured workflow determine next steps. |
| [Tasks](https://app.notion.com/p/3d749e23bd4b814da695e37b43a82547), edited Sept 18 | A defined goal, follow-up to obtain what is missing, and completion tied to observable evidence. Strong foundation for the task-ownership pillar. | Current database customer-safe core sections name Dan as approver; the older Text copy did not. Defined guides and supported actions constrain the work. Do not imply an unrestricted agent, unbounded persistence, invented goals, candidate selection decisions, or universal outreach to any person. |
| [Coordinator](https://app.notion.com/p/3d749e23bd4b81f3aef0c2551bddc84d), edited Sept 18 | Book a next meeting using availability and candidate selection; confirmations, reminders and rescheduling; carry prior context into the hiring manager's calendar invitation. | Listed GA; customer-safe core sections name Dan as approver. Calendar access and process configuration matter. Do not promise panel scheduling from this entry. Rich briefing depends on upstream information. |
| [Recruiter Notetaker](https://app.notion.com/p/3d749e23bd4b81ffa3c6c138f1649e20), edited Sept 18 | Capture a booked human video conversation, transcribe it, summarize it and attach context to the candidate record. | Listed GA; customer-safe core sections name Dan as approver. The entry explicitly distinguishes future functionality: it does not score the interview or complete an ATS scorecard. Meeting invitation is required; this capability does not cover phone calls. |
| [Workflows](https://app.notion.com/p/3d749e23bd4b819cb5fdff97d781e0d3), edited Oct 27 | Connect events, conditions, conversations, tasks, waiting and record updates into a configured recruiting process. Supports showing how work moves between steps. | WIP entry with unapproved prose. Supported steps are finite. Complex process changes involve implementation work and sometimes account flags. Do not promise any conceivable process, instant self-service transformation, or guaranteed deployment without setup. |
| [Single-Thread Hiring](https://app.notion.com/p/3d749e23bd4b812294afe98163f707c8), edited Oct 30 | An applied example of combining existing skills around a continuing candidate relationship. Useful evidence of the direction toward connected ownership. | Latest entry remains a one-customer pilot in India, with unapproved prose. It is not a separate product. Coordinator is not part of this pilot pattern. Third parties conduct and score their own tests; Alex follows up on completion. Do not claim the entire candidate journey in one conversation is generally available. |
| [Hiring Manager Intake](https://app.notion.com/p/3d749e23bd4b81988b13e8a640100674), edited Sept 10 | An applied example of getting role details through a conversation, then drafting job materials for a person's review. Demonstrates useful work beyond candidate interviews. | Feature-flagged, sellable in conversation, not GA; prose unapproved. It does not independently open the requisition. Passive intake notes and chasing panel feedback are described as future possibilities, not built capabilities. Excluded from the broad review-page promise. |
| [Identity](https://app.notion.com/p/3d749e23bd4b81f9b1c3c078d9e90532), edited Sept 18 | Documented identity checking using ID and live capture; a result that can be reviewed in the recruiting process. | Limited rollout in this source; confirm account availability before promising. The entry has mixed approval labels and flags some hero copy as drafted/unshipped. It does not provide background checks or establish right to work. Excluded from blanket capability promises. |

## Messaging document: proposed strategic direction

The Google Doc asks the brand to lead with the recruiting team's ambition and pain, rather than a product-first list. AI interviewing is part of the product, but is too narrow as the default company pitch. Show concrete work instead of declaring that the agent “does everything.” Make the implementation and change-management partnership visible.

The proposed ladder is:

1. **Own a task:** applications, interviews, and following up for missing information or documents.
2. **Own a process:** connect interviews, scheduling and follow-through; carry context forward.
3. **Transform operations:** work hands-on with the customer to define ownership, evolve roles and handoffs, and build toward business outcomes.

“Make hiring more human. For everyone.” is the proposed strategic frame: more opportunity for candidates to be heard, and more room for recruiting teams to do their best work. “Own as much as you want” describes the intended choice of delegation. In outward copy, show that choice through concrete examples and an agreed process rather than implying unrestricted capability.

## How this changes the Felt review copy

- **Buyer first:** the story begins with what recruiting teams make possible for people and businesses. The art direction sits in its own section.
- **Ambition with concrete work:** application review, interviews, missing-information follow-up and the next meeting booked are recognizable work. They substantiate the ownership story without an unbounded “everything” claim.
- **A clear expansion path:** the three pillars follow task, process and operations. The third pillar explicitly includes the human Felt team working alongside the customer's team.
- **Platform framing, corrected by Aaron:** Felt is the agentic platform for recruiting teams. Distinct products and abilities operate within it; the resourceful crew expresses agents across the journey. The previous one-product-actor interpretation is superseded. Do not infer a final navigation taxonomy or new product readiness from the illustration.
- **Human experience, evidence-based hiring:** emotional language concerns treatment, care and the experience of recruiting. It does not suggest that feelings replace role requirements, evidence or human judgment.
- **No invented proof:** the copy adds no metrics, customer names, performance guarantees or universal outcome claims. Pilot-only patterns and future functions stay out of the broad promise.

## Taxonomy note for further design work

The Alexandria index lists nine named capabilities: Talent Match, Resume Screens, AI Interviews, Proctor, Verify, Identity, Coordinator, Recruiter Notetaker and Tasks. The earlier seven-product icon scope followed the published website navigation. Tasks and Recruiter Notetaker are additional named capabilities in Alexandria, not proof that all nine should become separate public navigation items. Workflows and Knowledge Base are platform capabilities. Hiring Manager Intake and Single-Thread Hiring are applied examples, not new product slots. Confirm intended navigation and account availability separately from illustration coverage.

The current source read does not ratify any new Felt product names, master tagline, palette, or final identity system.


## Fresh alignment audit, October 31, 2026

Re-read tab t.0 of the exact supplied Google Doc through the Google Drive connector during round 19. The document still labels the proposed framing as proposed. Its central direction is “Make hiring more human. For everyone.” Its paired benefits are candidate opportunity to be heard and hiring-team time for their best work. It also explicitly calls for transformation partnership, task/process/operations ownership and concrete demonstrations instead of claiming everything.

The revised hero and rebrand thesis now state the paired human benefits. The story and ownership ladder make the human implementation team visible at task, process and operations levels. Platform is the initial product-family selection. The journey visualization shows candidate experience, agent actions and recruiter outcomes, with configuration-dependent optional checks and explicit human hiring decisions. Aaron’s newer agentic-platform label remains controlling.


## Product-line clarification, October 31, 2026

Aaron clarified that “Felt, at every step.” is specifically about being with the candidate throughout screening, interviewing, scheduling and the rest of the candidate journey. The task/process/operations ladder separately describes the work a recruiting team can choose to delegate. The agentic-platform positioning is unchanged.