Every proposal we send at The Kendall Project now takes 80% less time to produce and it comes out right the first time. We got there by applying the Kendall Framework to our own work: we started with the problem, built the context around it, then packaged that context into a proposal skill that Claude (or any generative AI) uses every time.
It Started With a Problem
Our proposals were taking too long to build. Each one pulls together brand standards, product descriptions, payment terms and the specifics of the client's need, and assembling all of it by hand took most of a working day.
That time had a real cost. Proposals reached clients later than we wanted, and when a client is ready to move, a late proposal can mean losing the opportunity entirely.
So we did what the Kendall Framework is built around: we started with the problem and built everything else from there.
Step 1: Build the Context Around the Problem
Once the problem was clear, every other piece of context had a purpose. We asked a few questions, each anchored to that problem:
- Who builds our proposals, and what are they responsible for?
- What does solved look like? For us, a proposal that is correct the first time and reaches the client on time.
- What rules can a proposal never break?
We captured the answers as Context Blocks, modular, standardized units of organizational knowledge used to make AI inputs reliable and reusable. We started with Role, Problem, Objective and Rule. Keeping the problem at the center meant we only built context that helped solve it.
Explore AI Foundations Training →
Step 2: Add the Rest of the Context the Proposal Needs
That gave us the foundation. From there, we added whatever other context a complete proposal requires. The test was simple: if a reviewer would ever have to correct it, the AI needed to know it up front.
That included:
- Brand standards: our colors, fonts, logo use and document formatting
- Products: what we offer, how each offering is described, and how they fit together
- Payment: terms, structure and the exact language we use
- Proposal structure: the sections every proposal contains and the order they appear in
- Anything else that recurs: the details that showed up in every proposal and used to be retyped each time
Each piece became reusable context with a clear owner, so when something changes, we update it once and every future proposal reflects it.
Step 3: Package the Context Into a Proposal Skill
With the context assembled, we built a proposal skill in Claude. A skill is a reusable set of instructions and context the AI draws on whenever a certain kind of task comes up. Ours holds everything a proposal needs, so the AI applies our brand, products, payment terms and formatting automatically.
This is what we mean by an AI assistant. It is not a clever one-off prompt. It is structured, governed context delivered to the AI every time, so the output is consistent no matter who on the team asks for a proposal.
Now, preparing a proposal starts with the client's specifics: who they are, what they need, and which offering fits. The skill supplies everything else. The first draft comes back on brand, with the right products and the right terms, ready for a final read rather than a rewrite.
What Could an AI Assistant Do for Your Organization?
The Result: 80% Less Time on Every Proposal
Proposal creation time is down 70%, which gives us back countless time on every proposal. Proposals now reach clients on time. The time came from two places: assembling the document by hand, and the review cycles that followed to fix brand, product and payment details.
Because the skill gets those details correct the first time, review now focuses on the substance of the offer for each client. The team spends its time on the thinking that wins work, and the assistant handles the details that used to consume the afternoon.
The value also compounds. Every proposal draws on the same certified context, so quality stays consistent as volume grows, and an update to a single Context Block improves every proposal after it.
Where to Start With Your Own
Pick a document your team produces often, where accuracy matters and the same details appear every time. Proposals, statements of work and client reports are all good candidates.
Then start where we did, with the problem. Name it clearly, including what it costs, and build the context around it: who does the work, what solved looks like, and the rules it must follow. Add the rest of the context the document needs, and package it into a skill so it is applied the same way every time.
The context work pays off even beyond the assistant. Writing down your brand standards, product descriptions and payment terms in one governed place brings clarity your team benefits from either way. The AI assistant is simply how that clarity shows up in every proposal you send.
Ready to Turn AI Into Real Business Value?