Everything here is in the product now

Governance for how you ask for money.

We check what you are claiming and sharpen the language you claim it in — not the prose, but the terms, the grammar and the framing a reviewer scores against.

How it runs

From your own files to an award, with a compliance check on each side.

Fig. 1 — The pipeline, stage by stageIllustrative
SourcesEmailDriveCalendarScrubunder set rulesEvidencePast performanceRevenueCapabilitiesfilled, a fact on file. dashed, a gap.Pre-awardeligibilityMatchopen fundingPost-awardwhat is duethe record

You connect the sources.

From your accounts, it reads only the ones you connect: your email, your documents, your calendar.

A governed scrub reads what is connected.

It reads those sources for evidence under set rules, and every gap it finds is left visible as a gap.

The evidence is sorted into three kinds.

Past performance, revenue and capabilities are the facts a funder asks for. A filled square stands for a fact on file and a dashed square for a gap.

Eligibility is checked before any money is pursued.

The first compliance check sits ahead of the match, so you do not spend effort on funding you cannot claim.

The evidence is matched to funding that is open.

Only evidence that passed the check reaches the match.

A second check follows the money after it lands.

It tracks what is due, what is overdue and what is missing. A clean record is the strongest thing you can hand the next funder.

The squares illustrate facts found and gaps left, and are not counts.
One system

Eleven capabilities, one profile, nothing gathered twice.

Fig. 2 — The method runs as a cycle, and each cycle adds to the facts on fileIllustrative
the recordcycle1facts on file01Foundationverified facts02Discoveryopen funding03Responsethe submission

Foundation

The work starts from the facts that only you hold.

Goes in
Past performance, numbers and registrations, read from the sources you connect.
Comes out
A narrative built from verified facts, with every gap left visible as a gap.

Discovery

Those facts are mapped to funding that is open now.

Goes in
The verified facts already on file from the first step.
Comes out
Grants, contracts and cooperative agreements you can pursue.

Response

The submission is drafted from the same facts.

Goes in
One opportunity that the facts on file can support.
Comes out
A response drafted from those facts, with a person approving every step that reaches a funder.

Renewal

The award and its record feed the next cycle.

Goes in
The award, and the record of what was due and what was delivered after it.
Comes out
One more fact on file that the next funder can check.
The figure shows the order of the work, and the lines in the middle are not counts.
A record, as the product holds itExample
SBIRCloses in 9 days
Phase I — Health Technology
$275,000Grant instrument
Eligibility checked before you saw it. Flagged: your SAM registration has to be current before this one can be filed.
Why it is built this way

The research, and where it shows up in the product.

On data and decisions

Firms that decide from data show output and productivity 5 to 6 percent higher than expected.

Brynjolfsson, Hitt & Kim. “Strength in Numbers,” 2011.
In the product

One profile, read by every capability.

  • The research becomes your opportunity records, the eligibility call reads the certifications you confirmed at the start, and the outcomes you gave us turn into the evidence inside a narrative.
  • Your documents, stories and strategy are reused by everything that writes.
  • Nothing is gathered twice, and nothing is handed over a wall.
On what gets through

Of 101 basic-science findings published with clear clinical promise, five were licensed for clinical use about twenty years later.

Contopoulos-Ioannidis, Ntzani & Ioannidis. The American Journal of Medicine, 2003.
In the product

Two gates before a word is generated.

  • Discovery starts from your profile — your codes, your certifications, your service areas — so the list is short because it is filtered.
  • Eligibility Analysis gives a GO or NO-GO with a confidence level, the factors that qualify you, and the factors that put you at risk.
  • A section schema you confirm before anything is written.
Two clocks

A deal takes a blink. Trust takes years.

Speed governs the operational layer; trust governs the relational layer. Optimizing one does not require degrading the other — but it does require building both deliberately.

So the product is fast where the work is mechanical, and it stops for you wherever a claim is being made in your name.

Fig. 3 — The same transaction, on two clocks9 cleared while trust builds
A transaction 0.1 s — done Deep institutional trust 3–5 years Not to scale
Fast, where the work is mechanical
  • Ask in plain language. It searches SAM.gov, Grants.gov and state portals, filtered by your profile.
  • A forwarded link comes back as a structured record. Two minutes instead of a browser tab you meant to come back to.
Yours to check, where a claim is made
  • Anything the agent could not verify stays marked in the text for a human to resolve.
  • On each section you Approve, Revise, Edit or Skip. Nothing leaves the screen on its own.
  • Document Rating scores against a named rubric. It will not invent criteria and hand you a number.
A response within about 0.1 second feels instantaneous: Nielsen, 1993, after Miller, 1968. Trust is built from assessed ability, benevolence and integrity, and its slower part accrues through repeated interaction: Mayer, Davis & Schoorman, Academy of Management Review, 1995; McAllister, Academy of Management Journal, 1995. Three to five years for deep institutional trust is our own estimate; the literature supports a multi-year span without fixing the figure. Not to scale.
The workspace

Where the eleven capabilities actually live.

The skills below are not eleven separate tools you log into. They run inside one workspace, against one profile, and the chat is always on alongside whatever you are looking at — so starting a skill is a keystroke rather than a navigation.

People stand before a large blue screen showing the sea
01

Dashboard

Where things stand.

02

Opportunities

The pipeline, by priority and deadline.

03

Calendar

Deadlines, as dates rather than a list.

04

Knowledge

Your documents, stories and strategy, reused by everything that writes.

05

Profile

The working set everything reads from.

06

Analytics & Billing

What you have spent, and on what.

01 · Foundation

The part nobody sells you, because it is not glamorous.

Most applications fail before they are written. Not on the prose — on a missing registration, an entity that cannot hold the award, or an outcome nobody wrote down at the time. This is that work, and everything downstream reads from it.

A woman walks through the archway of a blue building

Funding Readiness

Start here Funding Readiness: a guided conversation on the left, and the organization profile it builds on the rightSee it in detail

A guided conversation that builds your organisation profile: what you are, what you deliver, which registrations and certifications you hold, and whether you are chasing grants, contracts, or both. It ends at that choice, because the rest of the product branches on it.

Drop in whatever you already have — a capability statement, an old application, your registrations — and it reads them instead of asking you to type out what is already written down somewhere.

Called grant_intake inside the product. We call it readiness here because nobody has ever wanted intake — they want to know whether they are fundable.

Strategic Planning

Strategic Planning: a planning session in progress, with the strategy document building beside itSee it in detail

A planning session that produces positioning, your funding mix, a certification roadmap — 8(a), WOSB, HUBZone and the rest — and a twelve-month pursuit plan.

It writes back to your profile as a strategy document rather than leaving the conclusions in a chat you will not find again.

Story Collection

Story Collection: an interview for an impact story, with the story library and its readiness weight beside itSee it in detail

Interviews you for impact stories, testimonials and outcomes you can put a number on. Each one is tagged by program and handed to the writing step as evidence.

This is the difference between a narrative that asserts and one that demonstrates. A reviewer scores the second.

02 · Discovery

A short list, because it is filtered — not because the database is.

Searching Grants.gov or SAM.gov yourself returns thousands of things you are not eligible for. Discovery starts from your profile instead: your codes, your certifications, your service areas.

01

Grant Search

Ask in plain language. It searches SAM.gov, Grants.gov and state portals, filtered by your profile, and the results land as opportunity cards in your pipeline rather than as a list you have to re-key.

Grant Search: a search for STEM education grants in Mississippi, with results matched to the profileSee it in detail
02

Contract Search

The same, for the contracting side. The product tracks the full range as distinct types rather than lumping them together: RFP, RFQ, RFx, CMAS and ITN on the contracting side, NOFO, LOI, SBIR and STTR on the grant side. SBIR and STTR sit with the grants because that is how most agencies award them, though the instrument changes by agency and the record carries whichever it is.

Set-asides are first-class too, so DVBE and SB work is distinguishable from open competition at a glance instead of being buried in the text.

03

Add Opportunity

Somebody forwards you a link. Paste it, or drop the solicitation document in, and it comes back as a structured record: funder, value, deadline, codes, set-aside and attachments.

Two minutes instead of a browser tab you meant to come back to.

Add Opportunity: a pasted link extracted into a structured opportunity recordSee it in detail
04

Funder Language

Reads the solicitation and that funder’s past awards for the vocabulary they reward: the priority phrases, the evaluation verbs, the terms to mirror and the terms to avoid.

It is applied as a style layer over the writing, which is how a draft stops reading like a template and starts reading like it was written for that program.

Funder Language: a funder’s priority phrases, evaluation verbs, terms to mirror and terms to avoidSee it in detail
05

Partner Finder

A great many opportunities are out of reach alone and straightforward with a partner — because of size, past performance, or a set-aside you do not hold. This finds the organisations worth teaming with.

06

Partner Outreach

And then approaches them, which is the part people put off. Teaming conversations have to start weeks before a deadline, not in the final week.

What a match looks like when it landsExamples
NOFOCloses in 14 days
Community Development Block Grant — Public Facilities
$250,000 – $1.2MCDBGSB set-aside
Matched on your NAICS code and the certification you confirmed at intake. Eligibility checked before it was shown to you.
RFPCloses in 31 days
Workforce Services — Regional Delivery Partner
$1.8M over 3 yearsRFPOpen competition
Past performance in your profile covers two of the three required references.
03 · Response

The structure is agreed before a word is generated.

The failure mode of every AI writing tool is a confident draft in the wrong shape. The response path puts two gates in front of that: a decision about whether to bid at all, and a section schema you confirm before anything is written.

Eligibility Analysis

GO / NO-GO Eligibility Analysis: a GO recommendation with a confidence level, the qualifying factors and one riskSee it in detail

A recommendation with a confidence level, the factors that qualify you, and the factors that put you at risk. It reads your profile, the opportunity and your certifications.

A NO-GO is as useful as a GO, and considerably cheaper. It names what would have to change, which is often a single registration.

Opportunity Response

Opportunity Response: the solicitation analysis, ahead of the two steps a person decidesSee it in detail

The starting point on any opportunity. It analyses the solicitation, derives the sections the funder is actually asking for, and shows you that schema to confirm before generation begins.

Then it hands off to the writing agent in the right mode — grant or contract — because those are different documents with different rules, not one document with different words.

The writing agent

Works section by section The writing agent, in grant mode: the sections of a draft with a state on each, and the document beside themGrant mode, in detail The writing agent, in contract mode: sections numbered against the solicitation, unmet requirements and the rate tableContract mode, in detail

Not a prompt box that returns a document. An agent working a structure it agreed with you: it takes one section at a time, writes against the solicitation and your knowledge base, and stops. On each section you Approve, Revise, Edit or Skip, and a weak paragraph costs you that paragraph rather than the whole draft.

Anything it could not verify stays marked in the text for a human to resolve. It does not smooth over a gap to make the prose read better, which is the behaviour that gets people into trouble with a funder.

In contract mode the same agent speaks contract: section-number references against L and M, an inventory of requirements you do not yet meet, and the rate table.

The draft, section by section

A draft is a list of sections with a state on each one, not a wall of text. Word counts per section, against the page limit the solicitation sets. Approved sections lock, so a later run cannot quietly rewrite something you already signed off.

The agent stays open beside the draft rather than being something you invoke and dismiss: ask it to revise a section, start another skill, or challenge what it wrote, without leaving the document.

What comes out

A document with version history, a preview, sharing, and export — including straight to Google Docs, because the last mile of every proposal is somebody else wanting it in their own tool.

Drafts can be promoted when they are ready, so there is a difference between something you are working on and something you would send.

A draft, and the state of every section in itExample
Approved
Executive summary
Signed off. Locked unless you reopen it.
In review
Statement of need
Scored 82 against the funder’s own rubric.
Drafted
Project design
Waiting on you. Nothing is submitted from here.
Drafted
Outcomes and evaluation
Pulled the three numbers you gave us at intake.
Not started
Budget narrative
Needs the figures that are still open on your checklist.
Nothing leaves this screen on its own. A section is yours to accept, send back, or rewrite, and the price of generating one is quoted before it runs.
04 · Quality

Scored against a rubric, or not scored at all.

Document Rating: a draft scored against a named rubric, criterion by criterionSee it in detail

A score with no stated basis is worse than no score. It feels like information.

Document Rating scores any document — a draft, a capability statement, a narrative — against a named rubric: the funder’s own evaluation criteria, your internal QA standard, or one you supply.

If there is no rubric on file it asks you for one before scoring. It will not invent criteria and hand you a number.

Fig. 4 — A rating, and what every point rests onIllustrative
Document
Statement of need, in review
Rubric
The funder’s evaluation criteria, as published in the solicitation
Need is shown with dataThree local figures, each with its source. One is from 2019.
27 / 30
Fit with the funder’s prioritiesTwo of the three stated priorities are addressed by name. Workforce is not mentioned.
21 / 25
Who is served, and whereService area given by census tract. Eligibility is stated once, in the fourth paragraph.
18 / 20
The gap in existing servicesAsserted, not shown. No other provider is named.
10 / 15
Clarity and length412 words over the page limit.
6 / 10
Total82 / 100
Not scored
Cover letterNo rubric on file for this document. It asked for one, and scored nothing.
An example, not a client’s document. The criteria and weights stand in for a funder’s own.
What it costs to run

Bring your own key, and build on your own dime.

The AI in this product is Claude. You can run it on our account, or you can put in your own Anthropic API key and pay Anthropic directly.

Your key

In Settings, add an Anthropic API key from console.anthropic.com. It is stored encrypted for your organisation, and you can remove it at any time.

Model usage then goes on your Anthropic bill rather than through us. What we charge is separate and visible, instead of being folded into a number you cannot break down.

You also choose the model used for proposals and generation, rather than being handed whatever we picked.

Priced per section, quoted first

Pay per call

You are charged for a section when that section is written, and you see the cost, your balance, and what is left before you agree to it. Ask for a section again and it is priced again in front of you.

Every document is agreed to on its own. Nothing runs up a bill in the background, and there is no monthly surprise from work you did not watch happen.

Why we offer it

Because the alternative is a per-seat price with our margin hidden inside your model usage, and you having no way to tell the two apart.

With your own key the cost of the AI is a line on your Anthropic bill that you can read, and what you pay us is for the product and the work — which is the thing you should be judging us on.

If you would rather not hold a key, the organisation account works exactly the same way and tops up with funds.

One system

The report is not a deliverable. It is the state of your pipeline.

Most funding consultants hand you a PDF and leave. Ours is a live document at a permanent link that says which version you are reading and what changed since the last one — because the programs move, deadlines close, and a report that was right in March is wrong by June.

And the profile behind that report is the profile everything else writes from. The research becomes your opportunity records, the eligibility call reads the certifications you confirmed at the start, and the outcomes you gave us turn into the evidence inside a narrative. Nothing is gathered twice, and nothing is handed over a wall.

If you would rather read the engineering than the pitch, the skills architecture is published — every capability on this page as an entry with its inputs, where it is entered from, and what it hands on.

Money moves.
So should you.

Book a call →