How to write an automation project brief that gets you accurate quotes

Why most automation briefs fail

The typical automation brief sent to agencies looks something like this: "We need to automate our lead intake process. Currently it takes too long and we lose leads. Please send us a quote." That brief produces three predictable outcomes — none of them good.

First, it produces wildly different quotes. One agency quotes $8,000 because they assume a simple Zapier workflow. Another quotes $45,000 because they assume a custom integration with your CRM and a routing engine. Both quotes are technically responsive to the brief. Neither is comparable to the other. You can't evaluate them against each other because they're solving different problems.

Second, vague briefs invite risk premiums. Agencies hate ambiguity. When they can't see what they're building, they pad estimates 30-50% to absorb unknown complexity. Your $20,000 project becomes a $30,000 project before any work starts, just to cover the agency's uncertainty about what you actually want.

Third, vague briefs lead to mid-project scope fights. The agency starts building. You see the first demo and realize they misunderstood half of what you wanted. Now you're in change-order territory. Every change adds time and cost. By the end of the project, you've spent 60% more than the original quote and the relationship is strained.

Good briefs prevent all three of these outcomes. A good brief is specific enough that two competent agencies will quote within 25% of each other on the same scope. It's detailed enough that the agency can size the work accurately. And it's clear enough that scope changes during the project are rare and isolated, not constant and pervasive.

The brief is also the document that protects you. When something goes wrong mid-project — and something always goes wrong — the brief is what you reference. "We agreed in the brief that X was in scope. Y was explicitly out of scope. This change request falls under Y, which is a change order." Without a specific brief, every disagreement becomes a he-said-she-said about what was promised.

What an agency actually needs from you

Agencies need seven things from you to write an accurate quote. Most briefs include one or two of them. The gap between briefs that get accurate quotes and briefs that don't is usually whether these seven things are present.

1. The business problem in measurable terms. Not "we want to automate lead intake." Instead: "We currently get 200 leads per month from our website. Our average response time is 6 hours during business hours and 18 hours overnight. We close 12% of leads. We believe faster response would improve close rate to 18-20%." The agency now understands what success looks like in numbers.

2. The current process documented step-by-step. Walk through exactly what happens today. When a lead submits the form, where does the data go? Who sees it first? What do they do with it? What systems do they touch? What decisions do they make? Where does the lead go next? A two-page process walkthrough saves 20 hours of agency discovery time.

3. The systems involved and how the agency can access them. List every tool the automation will touch — CRM, email platform, calendar, payment processor, billing system. For each, note who has admin access, what API plan you're on, and whether you're willing to give the agency access during the build. "We use HubSpot Professional, Twilio with API access, Stripe with admin keys available" is much better than "we use HubSpot and some other stuff."

4. The success criteria. What does the automation need to do for you to consider the project successful? "Reduce response time to under 5 minutes" is a success criterion. "Generally improve our process" is not. Three to five specific criteria are ideal.

5. The budget range. Yes, share it. Agencies hear "we don't have a budget in mind, just send your best price" and treat it as a signal you're fishing for low quotes. Telling them "we've budgeted $15,000-$25,000 for this project" lets them scope appropriately and tells you whether they can deliver within your range. If three agencies all respond with quotes 3x your budget, you know your budget is unrealistic and you can adjust scope.

6. The timeline expectation. When do you need this live? Real deadlines (regulatory, contractual, seasonal) are different from preference deadlines. Tell the agency which is which. "We need this live by July 1 because that's when our compliance deadline hits" gets different treatment than "we'd like to have this done in Q3."

7. What's explicitly out of scope. This is the most-skipped item and the one that prevents the most disputes. "We want lead intake automation. We do not want CRM cleanup, sales sequence automation, or website form redesign in this project. Those may be future projects but are not part of this engagement." Now the agency knows where the fence lines are.

The seven questions that determine cost

Automation project cost is driven by seven specific factors. If you understand them, you can write a brief that produces a cost-appropriate quote. If you don't, you'll be surprised by quotes that seem high or low for reasons you don't understand.

1. How many systems are involved? A 2-system integration (web form → CRM) is dramatically different from a 7-system orchestration (form → CRM → enrichment tool → calendar → email → Slack → analytics). Each system added increases complexity non-linearly. Three systems is roughly twice the work of two. Five systems is roughly five times the work of two.

2. Are the systems API-friendly or do they require workarounds? Modern SaaS tools (HubSpot, Salesforce, Stripe, Twilio) have mature APIs. Legacy systems, custom-built tools, or on-prem software often require workarounds — screen scraping, CSV imports, RPA tools. Workarounds add 3-10x the development time of clean API integration.

3. How complex is the business logic? "When a lead comes in, send a welcome email" is simple logic. "When a lead comes in, check if they're from an existing customer account, route to the assigned account executive if so, otherwise score the lead based on company size, industry, and behavioral signals, then route to the appropriate team based on score and territory" is complex logic. Complex logic adds significant build time and significant testing time.

4. What's the error handling and exception requirement? "Send notification when something fails" is basic. "Automatically retry, fall back to manual queue, alert specific team based on error type, log to audit trail" is enterprise. Enterprise-grade error handling can be 30-40% of total build cost.

5. Does it need real-time or can it run in batches? Real-time automation is expensive — webhooks, queueing infrastructure, idempotency handling, race conditions. Batch automation (runs every 15 minutes, runs nightly) is much cheaper. Many use cases that "feel" real-time can actually tolerate 5-15 minute delays without business impact.

6. What's the data volume? Automating 200 records per month is different from automating 200,000 records per day. Volume affects infrastructure choices, error handling needs, and monitoring complexity. Be honest about current and expected volume.

7. Who owns it after launch? If the agency builds and walks away, that's one cost. If the agency builds and provides ongoing maintenance, that's another. If the agency builds, then trains your team to maintain it, that's a third. Each model has different upfront costs and different total costs of ownership. Decide which you want before getting quotes.

When you address these seven questions explicitly in your brief, you give the agency everything they need to write an accurate quote. When you don't, the agency makes assumptions — and their assumptions usually drift toward higher cost to protect themselves.

The brief template (copy this)

This is the structure that produces comparable, accurate quotes. Copy it, fill in your specifics, and send the same document to every agency you're considering.

Section 1: Company context. Two paragraphs. Who you are, what you do, who you serve, how big you are. This isn't marketing copy — it's context that helps the agency understand your operational scale and complexity.

Section 2: The business problem. What's broken today, quantified. Include current metrics (response time, conversion rate, time spent on manual work, error rate) and target metrics. This is the success criteria framing.

Section 3: Current state walkthrough. Document the existing process step-by-step. Include screenshots if helpful. Note where humans make judgment calls vs. where the work is mechanical. Note what systems touch the process today and how.

Section 4: Desired future state. What does the process look like after automation? Walk through it the same way — step by step, system by system. Be specific about where humans still touch the process and what they do.

Section 5: Systems and access. List every system involved. For each, note: vendor name, your plan tier, API access status, who has admin credentials, whether the agency can have access during the build. Include integration platforms you already use (Zapier, Make, n8n) if applicable.

Section 6: Volume and performance requirements. Current volume, expected 12-month volume, peak volume scenarios. Latency requirements (real-time vs. batch). Uptime requirements (best-effort vs. monitored vs. SLA).

Section 7: Out of scope. What you're explicitly NOT asking for. This protects both parties from scope creep.

Section 8: Success criteria. Three to five measurable outcomes. These become the acceptance criteria at project end.

Section 9: Budget range and timeline. A range, not a single number. Real deadline vs. preference deadline.

Section 10: How you'll evaluate proposals. Tell them what you're looking for in the proposal. Reduces back-and-forth and signals seriousness.

Send this same document to every agency. Don't customize it per agency — uniform input is the only way to get comparable output. The cleaner version of the brief lives in a single Google Doc that you share read-only with each agency.

What to leave out (and why)

Bad briefs include things that pollute the agency's ability to scope the work. Three things to leave out:

1. Your proposed technical solution. If you write "we want this built in Make.com using these specific modules," you've eliminated the agency's ability to recommend a better tool. Maybe Make is the right choice. Maybe n8n is. Maybe a custom integration is. Let the agency recommend the stack as part of their proposal. You're hiring them for expertise — let them apply it.

The exception: if you have a hard constraint (existing investment in a tool, IT-mandated stack, compliance requirement), state the constraint and why. "We must use AWS-based services because our compliance program requires US-hosted infrastructure" is a constraint. "We want to use Zapier because we've heard of it" is a preference disguised as a constraint, and it's expensive to enforce.

2. Sales pitches about why your company is great. Agencies don't need to be sold on you. They need information to scope the work. A two-paragraph company context section is enough. Multi-page company history slows the agency down without helping them write a better quote.

3. Vague aspirational language. "We want best-in-class automation." "We want a future-proof solution." "We want something that scales." These phrases mean nothing to an agency trying to scope the work. They're either redundant (every agency assumes you want quality) or they're your way of avoiding specifics (which forces the agency to make assumptions). Replace aspirational language with measurable criteria.

A common failure mode: founders include 10 pages of aspirational vision, then 2 paragraphs of actual requirements. The agency now knows what kind of company you want to become, but not what they're actually building. Flip the ratio. Keep aspiration brief; make requirements detailed.

How to evaluate the quotes you receive

Once your brief is out, you'll get proposals back. The instinct is to compare total prices and pick the cheapest one that seems credible. That's the wrong instinct. Cheapest quote often means highest total cost — through scope creep, rework, and missed deadlines.

Evaluate proposals on five dimensions:

1. Did they understand the brief? Read the proposal's "understanding of requirements" section. Did they accurately summarize what you asked for? Did they identify the actual problem, or did they reframe it into something they're comfortable building? If a proposal says "we understand you want X" and X isn't what you asked for, that's a discovery problem that will compound through the project.

2. Are their assumptions visible? Good proposals state assumptions explicitly. "We assume your HubSpot is on Professional plan with API access enabled" or "we assume the existing form will not be modified as part of this project." Visible assumptions are negotiable. Hidden assumptions become change orders.

3. Is the breakdown specific? Vague proposals say "Discovery & Design: $5,000. Development: $15,000. QA & Launch: $5,000." Specific proposals break down the actual work — what gets built, how long each component takes, who does what. Specific breakdowns let you negotiate scope. Vague breakdowns prevent meaningful negotiation.

4. What's the post-launch story? The proposal should address what happens after launch. Who owns the automation? Who maintains it? What happens when something breaks? What's the warranty period? An automation that works for 3 weeks then quietly breaks is worse than no automation at all.

5. Are the people on the proposal the people who will do the work? Agencies sometimes do "bait and switch" — senior people sell the project, junior people execute. Ask directly: who will be the lead architect? Who will write code? What's their availability during the project? If the agency won't commit to specific staffing, that's information.

If you're comparing two or three credible proposals and one is dramatically cheaper, the cheaper proposal is usually missing something. Find what it's missing before choosing it on price alone.

Red flags in agency responses

Agencies signal a lot in how they respond to briefs. Watch for these red flags:

No discovery questions. A good agency will read your brief and come back with five to fifteen clarifying questions before quoting. If an agency reads a multi-page brief and immediately produces a quote with no questions, they're either (a) not actually reading the brief, (b) using a template they apply to every project, or (c) so eager to close that they're skipping due diligence. None of these are signs of a good build partner.

Fixed-bid quotes for vague scope. If your brief has ambiguity (and most do, even when written carefully), a fixed-bid quote is the agency taking on the risk of that ambiguity. They usually price it in. If an agency offers fixed-bid quotes regardless of scope clarity, they're either pricing high or planning to use change orders aggressively.

Aggressive timeline promises. "We can have this live in 3 weeks" for a complex multi-system automation is usually fiction. Either they're going to cut corners on testing and discovery, or they're going to miss the deadline. Agencies that promise unrealistic timelines either don't understand the work or assume you'll forgive missed deadlines.

No reference customers in your stack. If you're building HubSpot-Stripe-Twilio automation and the agency has never done a project with that combination, you're paying for their learning. Either their quote should reflect that learning curve, or they should disclose it.

Reluctance to share IP and code arrangements. Who owns the automation code after the project? Who has access to it? What happens if you fire them mid-project — do you get the work in progress? Good agencies answer these questions clearly. Bad agencies obfuscate. Code ownership clauses matter — get them in writing before signing.

Pressure to sign quickly. "This pricing is only valid this week" or "we have one slot opening that I can hold for 48 hours" are sales tactics, not technical realities. Good agencies have backlog and can wait for you to make a careful decision. Agencies that pressure you to sign are usually either desperate for revenue or trying to prevent you from comparing alternatives.

Common mistakes to avoid

Three mistakes pop up repeatedly in automation procurement:

Mistake 1: Sending the brief to 10+ agencies. More quotes does not mean better information. After 3-5 quotes, you'll see the price range and understand the work clearly. Beyond that, you're adding cycles without adding signal. You're also signaling to agencies that you're shopping aggressively, which makes the best agencies less interested in working with you.

Mistake 2: Choosing the cheapest credible quote. The cheapest quote almost always means missing scope, junior staffing, or aggressive change-order tactics. The cheapest quote often becomes the most expensive project. The best value usually comes from the middle of the price range — competent agencies pricing the work honestly.

Mistake 3: Skipping the contract details. Once you pick an agency, the contract matters more than the quote. Get clear answers on: payment schedule (avoid 100% upfront), milestone-based payment ties, IP and code ownership, what happens if they miss deadlines, what happens if you cancel, warranty period after launch, ongoing maintenance terms. Lawyers who specialize in agency contracts are inexpensive compared to the cost of getting these wrong.

The best automation projects start with a tight brief, end with a clear contract, and have a competent agency in between. Each of those three elements matters. A great agency can't save a vague brief. A great brief can't save a bad agency. And a great agency with a great brief still needs a clear contract to align expectations through the project.

Most operators skip the brief work because it feels like overhead. It's not. The 8-15 hours you spend writing a tight brief saves you 80-150 hours of project friction. That ROI is among the best you'll find anywhere in your operation.