Last week, I had an interesting conversation on LinkedIn about one of the hardest parts of government contracting: deciding if an opportunity is really a good fit.
Finding a contract is one thing. Deciding whether to pursue it is much harder.
You need to understand the requirements, evaluation criteria, customer, competition, your capabilities, and your gaps.
Then someone has to answer: Do we have a chance of winning this contract?
That requires human judgment. But a human doesn't have to do all the work to make that decision.
Don't ask ChatGPT or Claude if you should bid
You could upload a solicitation to ChatGPT or Claude and ask, "Should we pursue this contract?"
I wouldn't.
Instead, break the work into small tasks to answer that question and assign them to Digital Employees.
That's how I built the THINK School AI Capture Team.
I use three Digital Employees inside the team for this process.
Chet finds contracts that match the company's work and turns them into opportunity leads.
Kipp qualifies each opportunity against your company’s capabilities and qualification criteria. Instead of a generic fit score, Kipp presents the evidence in five categories: Fit, Partial, Unsupported, Unknown, and Risk.
If something isn't known, it stays unknown. That shows the human where more research or judgment is needed.
Ben uses the opportunity and qualification evidence to create a Capture Brief with a recommendation to pursue, investigate, or not pursue.
The human makes the final decision.
I built this workflow into the THINK School AI Capture Team. If you have a contract you're evaluating, give it to the team and see how Chet, Kipp, and Ben work through deciding fit.
How to build Digital Employees for your AI Capture Team?
A Digital Employee needs more than a name and a prompt.
Build each around four things:
Job: What work does this employee own? For Kipp, the job is to qualify an opportunity against the company's pursuit criteria.
Context: What does the employee need to know? Kipp needs information such as your capabilities, past performance, priorities, and qualification rules.
Skill: How should the employee perform the job? Kipp's process might be: read the opportunity, identify requirements, compare them against company context, collect evidence, identify gaps, and flag risks.
Output: What should the finished work look like? Kipp produces structured qualification evidence that another Digital Employee or human can use.
The model is:
Job + Context + Skill + Output = Digital Employee
Now connect the work
Digital Employees become a team when you connect their work.
Human Goal
↓
Chet: Find → Candidate Opportunity
↓
Kipp: Qualify → Qualification Evidence
↓
Ben: Brief → Capture Brief + Recommendation
↓
Human: Pursue | Investigate | Pass
The output from one job becomes the input for the next. That's orchestration.
My AI Team management model is:
Subject Matter Expert → Human Manager → AI Team → Digital Employees
The AI Team provides capacity. Digital Employees execute the work. The human provides judgment and remains in control.
Build a system, not prompts
Instead of starting with, "What should I ask ChatGPT or Claude?"
Start with, "What work needs to get done?"
Identify the jobs.
Give each Digital Employee the right context and skills. Define the outputs. Then connect the work.
That's how you move from using AI to running an AI Team.
It also gives you a record of why you pursued or passed on a contract.
Whether you win or lose, you can compare the outcome to the decision and improve your assessment next time.
And if you want to see what this looks like before building your own, give a contract you're considering to the THINK School AI Capture Team and let Chet, Kipp, and Ben work through the process with you.
Marvin