FundFit vs Atom Grants

The post this page is answering

On 23 January 2026 Atom Grants published “Atom Grants vs. Streamlyne Fund Fit 2.0: A Comprehensive Comparison”. It has been ranking unopposed since, which is the ordinary reason to write a reply.

The less ordinary reason is that it is a good post. It opens by listing nine capabilities both products share before it argues for anything, and it ends by naming three areas where FundFit is ahead — accurately. Most vendor comparisons do neither. So this is an answer rather than a rebuttal: their seven claims, taken one at a time, plus the part they could not have written, which is where we were wrong about them.

Three rows we had wrong

This page has existed since April, written from a feature matrix without their site in front of us. With the crawl in hand, three rows were false. They are corrected in the table above, and here is what each one said:

We said Atom had no ORCID integration. Their site says profiles are built entirely from public information — ORCID, Google Scholar, university and lab websites, and published papers — across 1.2 million US researchers. Not only do they use ORCID, they use more sources than we do.

We said their team building was internal-only. Their documentation describes a collaborator graph that “extends that search across institutions to surface complementary co-PIs”. That is the same reach we claim, and we were describing it as narrower.

We said they did not screen eligibility. Their site says the opposite in plain terms: criteria “such as citizenship, career stage, or institution type” are highlighted, and anything that may disqualify a researcher is flagged before they apply. Eligibility screening was one of our four headline differentiators. It is not one.

Publishing a comparison means accepting that the other company will read it. Leaving three false rows up after seeing their documentation would have been worse than getting them wrong in the first place.

Their seven claims, one at a time

1. AI-generated grant listings

Fair, and a real difference in kind. Atom generates structured listings by extracting eligibility, deadlines and amounts from PDFs and foundation sites into a consistent format. FundFit reads the same source material to match against, but does not restructure every opportunity into a uniform record. If your office spends its time normalising inconsistent solicitations by hand, that is a direct answer to it.

2. Interactive RFP chat

Fair, and we do not have it. An assistant trained on one specific solicitation, answering “what documents do I need” without anybody reading fifty pages, is useful in a way that is easy to demonstrate and hard to dismiss. Test it with a solicitation of your own — a badly written one, not the exemplar in the demo — and see whether the answers survive contact with the footnotes.

3. InfoReady integration for limited submissions

Fair. We do not integrate with InfoReady. If your internal competition process runs there, that is a concrete workflow they support and we do not, and it should count.

4. Deep research reports

Fair, with a question attached. Proactively generating a curated analysis of a researcher’s funding landscape is more than reactive matching. The question is how often a researcher reads a long generated report — the same question we would want asked about any long output. Ask them for engagement numbers, and ask us the same about our digests.

5. Natural-language email explanations

Fair, and the sharpest of the seven. Their alert email explains in plain language why each opportunity was sent, so the reasoning arrives with the message rather than behind a click. Ours puts the reasoning on the match. Theirs is better placed for a faculty member who will never open the tool — which is a real argument, and it is why the delivery point below is worth reading properly rather than as a rebuttal.

6. Administrative analytics

Fair for a standalone product. They have engagement dashboards inside the platform. FundFit institutions running Streamlyne Research report through Streamlyne Reporting, across proposals and awards as well as discovery — a broader answer for institutions that have it, and no answer at all for those that do not. If FundFit is your only Streamlyne product, take their point.

7. Grant forecasting

Fair, and genuinely theirs. Recognising that a programme runs annually and surfacing the anticipated cycle before it is announced is a capability we do not have and have not claimed. The question to put to them is what a forecast looks like when it is wrong: whether a researcher can distinguish an anticipated cycle from an announced one at a glance, because preparing against terms that then change is its own wasted fortnight.

Where they say we are ahead, and what we would add

Their post names three: ORCID-built profiles, Slack and Teams integration, and individual subscriptions. Two of those we would put differently.

ORCID is no longer a differentiator, and we should stop treating it as one. They build profiles from public signals too, from more sources than we use. What is left of the argument is narrower and still true: neither product should be asking your faculty to complete a questionnaire, and any vendor still doing so has not learned what kills adoption in this category.

Slack and Teams is the one that holds. They concede it and we can find no evidence of it on their site. The reason it matters is not the integration — it is that a funding email, however well written, competes with every other email a researcher gets. Their explanations-in-email approach is the best version of the email answer. Ours is not to send an email.

Individual subscriptions hold too. They sell to institutions. We sell to both. For a research office that is invisible; for a postdoc between appointments, or a researcher at an institution that will buy neither product, it is the entire question.

We would add a fourth they did not mention, because they had no reason to: HERD Visualizer PRO is included with a FundFit institutional subscription. That is NSF HERD survey data made navigable — where your institution sits against a peer set, by field, by year. It answers what comes after discovery, and it is a separate line item in most stacks.

What we would test if we were you

1. One real researcher, both products, one month. Count opportunities surfaced, opportunities your office judged worth pursuing, and opportunities the researcher actually opened. The third number is the one that predicts whether the purchase changes any behaviour.
2. Bring your own bad solicitation. Give both products a poorly structured call with the requirements scattered through footnotes. Their RFP chat should shine here or not at all.
3. Test delivery on a Tuesday. Ask each vendor what happens for a researcher who never logs in. One answer is an email with reasoning in it; the other is a message in the channel they already have open. Both are real; only your faculty can tell you which one gets read.
4. Try to break the eligibility screen. Both products claim to check it now. Pick someone structurally ineligible for a well-known programme and see whether either surfaces it anyway.
5. Ask about the forecast being wrong. And ask us about a match being wrong. The interesting answer is what the product does afterwards.
6. Price the whole stack. If you are buying benchmarking, count it. If you also run an eRA system, count whether the discovery tool comes from the same vendor or needs a connection built.

What we are not claiming

We are not claiming Atom Grants is a weak product. It is the strongest thing in this category besides ours, their post was more honest than we expected, and on four capabilities they are simply ahead.

We are not claiming the nine shared capabilities are secretly ours. They listed them accurately.

What we are claiming is narrower than this page used to claim: delivery into the tools researchers already use, subscriptions for people without an institution behind them, benchmarking included rather than bought, and one vendor if you also run an eRA system. Four things, not fourteen.

Everything above about Atom Grants comes from their own public site and documentation, crawled on the date at the top of this page. If we have something wrong — again — tell us and we will correct it and re-date the page, as we just did with three rows.