RFPs Suck. AI's Going to Make Them Worse.

Oct 10, 2025

Song recco: (What's so funny 'bout) Peace, love and understanding. Elvis Costello & The Attractions

My tepid take: RFPs suck and are about to get a whole lot worse. This is part 1. Part 2 link at the end.

RFPs don’t work for nonprofits. Never have. And they’re about to get much, much worse. This tepid take is brought to you by a convo Ben Childers and Sam Landenswitch had, and to which I responded to -- briefly -- on Linkedin a few days ago. And in my 22+ years in this sector I've mostly been in biz dev and partnerships, done and led teams who responded to 100s of RFPs and written 100s of proposals.

RFPs were supposed to make things fair. But for nonprofits & the ecosystem (consultants, agencies, vendors), they’ve become a process that commodifies trust. RFPs aren’t actually the problem imo. They’re the symptom of how nonprofits, agencies, consultants, and vendors see and value each other.

When orgs treat vendors as “them,” versus "us", the result is a transactional process that strips out context, curiosity, and creativity. When they don't include (or don't identify or name) the complexities and context, they can't possibly get high quality, tailored, thoughtful responses. When respondents (agencies, consultants, vendors) treat orgs as “not as smart as us,” don't name the work and effort involved in responding, the result is boilerplate proposals that look fine on paper but have zero relationship behind them (and builds resentment). 

We call it fairness but to me it looks like bureaucracy in empathy’s clothing. It looks like missing the point of finding a great partner to do the best work within the (human) context of a dynamic, complex relationship. 

Then add AI to the mix: agencies are building agents to pump out RFP responses in minutes. Orgs are already talking about bots to detect those responses (reminds me of how school boards are responding to AI 🙄 ).

An automation race will = flooded inboxes with quantity and garbage. Not thoughtful, specific proposals. So yes, the RFP mess is about to get worse. Once both sides are using AI to pretend they’re listening to each other, we're running towards a cliff.

The way back isn’t better templates or smarter bots. We have to 1) identify and name the relationships we want with partners, recognize that this is an ecosystem, that most vendors / agencies / consultants are not "them", there is no "us" in opposition to them, and the false binary is BS (and corporate, capitalist BS at that). Then 2) re-imagine what “fairness” actually means in a dynamic, complex system. 

If you want the how: the tangible, human ways to make this whole thing suck so much less, that’s in post 2: How to Make RFPs Suck Much Less (and More Human) https://lnkd.in/gHg748Zj

***

This is part 2: how to Make RFPs Suck a Whole Lot Less (and More Human)

Part 1 was RFPs are about to get a whole lot worse: https://lnkd.in/gjvxaNTv. 

RFPs aren’t broken. That assumes they broke along the way but started off whole or functional. They feel so soul-sucking because they strip out what actually makes great work happen: trust, context, and the reality of humans working together in changing environments with fluctuating needs/goals/challenges and evolving knowledge.

So what do we do instead of sprinting toward that AI-generated cliff? First we name what we’re really doing when we issue or respond to an RFP. We’re defining a relationship. We’re building (or breaking) trust before the work even starts.

Want to ignore that every project is actually about people, with different power, emotions, and egos? Want to flatten complex relationships into line items and budget cells? Want to keep pretending that “transparency + transactionality = fairness”? Then keep doing RFPs the old way.

Or you try this. Actionable tips to try:

1 - Start with a conversation (not a templated form)

Before you issue an RFP, host discovery calls or open briefings with some trusted partners. Use it to pressure-test your problem statement and hear how others would frame it.

2 - Share what’s unknown (not just what’s decided)

Please stop pretending you have perfect clarity. Naming the uncertainties is so much more productive. And human. (“We’re not sure if this is a tech issue or a workflow issue.”) It builds trust and invites better ideas.

3 - Pay for scoping

If you’re asking respondents to help define the problem, call that what it is: work. A small paid discovery phase beats ten unpaid hours of guessing.

4 - Focus on the fit (aka the partnership) 

Replace 20 pages of boilerplate with a few real questions:

- How do you approach work like this?

- What do you need from us to succeed?

- What does success look like to you?

- How do you propose we work together?

5 - I beg of you, use AI for clarity (not volume)

Let AI help you tighten your questions or summarize responses. Please don’t use it to generate RFPs or RFP responses or to evaluate responses / proposals. 

None of this will “fix” RFPs. They weren’t built for the kind of work nonprofits actually do. They can't be fixed into effectiveness. 

But every time we choose curiosity over control and relationship over transaction, I promise we make the system a little more human. And honestly, that’s the only kind of fairness worth building toward.

If you missed the first post, the why behind all this, read: RFPs Are About to Get a Whole Lot Worse.