Impact Bridge resources
Guide
Before you post the role: 10 questions
Most hiring mistakes start way before the interview. They start when the role gets written around a title and everyone hopes the right person will make the problem clearer. Answer these first.
- What business outcome should this hire move in the next 12 months? If you can't name one, you may be hiring for comfort, not for growth.
- What problem are we actually solving? Too much work, unclear ownership, a broken workflow, and a missing skill all look the same from far away. They need different answers.
- What will this person own? Write three to five outcomes, not a list of tasks.
- What should be different 90 days after they start? If nobody can say, nobody will be able to tell whether the hire is working.
- Does someone already own part of this work? Overlap is where a lot of new hires get stuck.
- Is a manager still holding on to work they should have handed off? Sometimes the fix is delegation, not headcount.
- Which tasks need a person, which need a person working with AI, which could be automated, and which should stop? Use the Person, AI, or both worksheet.
- What level does the work really need? Match the title to the work. Inflated titles create problems that show up a year later.
- Who will onboard and lead this person? A great hire with an unprepared manager can still fail.
- If we didn't hire, what would we do instead? If the answer is "nothing changes," look harder at question 2.
The rule of thumb: if you can't answer 3, 4, and 7, you're not ready to post the role yet. That's not a failure. It's the cheapest point to catch the mistake.
Tool
Person, AI, or both?
Before you write a job description, break the role into tasks and decide what each one actually needs. The example below is filled in. Replace it with your own.
How to sort each task
- Human: needs judgment, influence, relationships, compassion, or a specialty only a person brings.
- AI-augmented: AI can drive it, but a person still owns it and checks it.
- AI or automation: once it's built and trained, it can run on its own.
- Stop doing: it doesn't serve an outcome the business needs.
Don't forget the hidden work. Every task you hand to AI still needs someone to monitor it, correct it, and keep it running. That work is never zero. Count it when you size the role.
Guide
Is it a role, or is it just work?
"We have work that needs to get done" and "we need another full time job" are two different sentences. Most hiring mistakes happen when a founder treats them as the same one.
1. Important doesn't mean full time
A task being important proves a need exists. It doesn't prove you need a person dedicated to it. Some critical work takes a few hours a month. Size the role by how long the work actually takes, not by how much it matters.
2. Size the role after you decide what AI will do
Don't size a role on how long the work takes today. Decide which tasks a person keeps, which AI can support, and which can run on their own. Then size the human work that's left.
3. Count the work of managing the AI
If AI takes on part of the role, someone still has to monitor it, correct it, and maintain it. There's also the work of setting it up and covering the load until it works. Leave that out and your plan will look cheaper than it is.
4. Some work needs a fast answer, not a full time person
A task can take little time and still need someone to respond quickly. Ask what happens if the response is slow or wrong. If the answer is serious, a person needs to respond. If it's low risk, AI can be the first response.
5. Decide what the work needs before you decide who does it
Outsourcing, contracting, and hiring are sourcing choices. Make them after you know what the work needs. The full list of options is wider than most founders use:
- Hire, full time or fractional
- Develop someone you already have
- Redeploy the work to the right owner
- Contract outside expertise with a clear scope
- Automate with a human in the lead
- Eliminate work that doesn't serve an outcome
A founder once asked us what a senior operations hire should cost. When we mapped the work, the strategic part was already covered by a more senior leader who was coming in. What was left was operations work. That's a different role, a different level, and a different price. Before you benchmark the salary, figure out the work.
Template
The 120 day integration plan
Hiring gets the person in the door. What happens in the first 120 days decides whether they stay and deliver. Write this plan before the person starts.
Before day onePull the success outcomes from the scorecard you hired against. Write what good looks like at 30, 60, 90, and 120 days. Name who they need to meet and what their manager will do in the first two weeks.
Day 15Ask: Are you clear on what you own? What's confusing so far? Who haven't you met yet that you should?
Day 30Ask: Does the role match what we described? Where are you stuck? What do you need from your manager that you're not getting?
Day 60Look for the first visible wins. Ask: What's getting in your way? What would you change about how work gets done here?
Day 90Review against the 90 day outcomes you wrote before they started. Adjust the scope if the role has changed. Say so out loud.
Day 120Have a stay conversation. What would make you stay? What would make you leave? Can you connect your work to where the company is going?
For founders and leaders
Don't let what you hear in these conversations sit in a file. Pull the themes across every new hire and look at them together. If three people tell you the same thing, it's not a people problem. It's the system.
For one founder, the theme that came up early was that new hires didn't understand the vision. That's fixable, if someone tells the founder in month two instead of in the exit interview.
Template
Interview scorecard
Most managers are expected to interview well. Very few are trained to. They ask a question, get a partial answer, and fill in the blanks themselves. A scorecard fixes that.
Build it before the first interview
| Part | What goes in it |
|---|
| Role outcomes | The three to five outcomes this person will own. Pull them from "Before you post the role." |
| Competencies | The capabilities each outcome needs. Keep it to the ones that actually predict success in this role. |
| Questions | One or two questions per competency, asked the same way to every candidate. |
| Evidence | What the candidate actually did, in their words. Not how you felt about them. |
| Rating | 1 = no evidence. 2 = some evidence, gaps. 3 = clear evidence. 4 = strong evidence at the level the role needs. |
| Decision | Hire, don't hire, or what you still need to learn. With the reason. |
Rules that make it work
- Probe. When you get a partial answer, ask what they did, what happened, and what they would do differently.
- Score before you talk. Each interviewer scores alone before the debrief, so the loudest opinion doesn't set the result.
- Evidence over chemistry. Confidence and likability are not the same as capability. A good interview predicts performance. It doesn't just confirm chemistry.
- Gut can inform the decision. It shouldn't be the whole decision.
Checklist
Is your manager ready for this hire?
You can hire great people. But if the manager is unclear, unavailable, inconsistent, or unprepared to lead, the hire can still fail. Check these before the offer goes out.
- Is the role clear? Can the manager explain what this person owns in two sentences?
- Is the manager prepared to lead? Do they have time on their calendar for this person in the first 30 days?
- Is onboarding structured? Is there a written plan, or is it "shadow people and figure it out"?
- Are expectations defined? Does the new hire know what good looks like at 30, 60, and 90 days?
- Will feedback come early? Is there a set time to check in, or will feedback wait for the first review?
If the answer to any of these is no, the hire could fail because of the system, not the person. Strong people can still fail inside weak systems.
Article, by Chichi Okpara
Before you hire the person, figure out the work
I see this all the time.
A company says, "We need to hire."
And sometimes they do.
But sometimes the real issue is that nobody has stopped to figure out what the problem actually is. Is there really too much work? Is the role clear? Does someone else already own part of it? Is the manager still holding on to things they should have handed off a long time ago?
Because if the role itself is messy, hiring someone into it usually doesn't make things better. Now you just have a new person trying to make sense of the same confusion.
Before you hire, you should be able to answer: What problem are we actually solving? What does this person need to own? What should be different 90 days after they start? And is hiring even the answer?
Sometimes it is. Sometimes the real issue is unclear ownership, a broken workflow, or a manager who needs more support.
Hiring is expensive. Do the thinking before you add the headcount. Because hiring someone into confusion usually just gives the confusion a new employee.
Article, by Chichi Okpara
You automated the mess
Companies are buying AI before fixing the way work actually gets done.
And I keep thinking, so what exactly are we automating?
Because if the process is already broken, AI doesn't magically fix it. You automated the mess. Congratulations. Now you have a faster mess.
AI strategy has to start with understanding the work first. Where are people getting stuck? What's unclear? What's taking too long? What actually needs to change?
Getting clear takes front loaded work. Thinking, redesigning, processing. Then you bring in the AI and map the workflow well. AI can make you less efficient if you're not clear on the work.
Then bring in the AI. But you still need the human.
Article, by Chichi Okpara
Human in the lead
Everyone keeps talking about keeping the "human in the loop" with AI.
At Impact Bridge, I emphasize something a little different with my team: human in the lead.
Because to me, "in the loop" can start to sound like the human is just there to review what AI already did.
But AI only knows what we give it. It doesn't know the history. The personalities. The nuance. Why something worked before and why it might not work now. That context comes from us.
So yes, use the AI. Let it make us faster. Let it help us think. Let it surface things we may have missed.
But the human still has to lead.
Article, by Chichi Okpara
The hire was right. The system wasn't.
Sometimes the hire wasn't the problem at all.
The role was unclear. The manager was not prepared. The onboarding was too loose. Success expectations were not defined. Feedback came too late.
Most people will tell you an A player will be an A player anywhere. I don't agree. Put an A player in a system with no structure, no accountability, and no clear outcomes, and they will cave to it.
So before you decide the person was the problem, ask: Was the role clear? Was the manager prepared to lead? Was onboarding structured? Were expectations defined? Did feedback come early?
If the answer is no, the hire probably failed because of the system, not the person. Strong people can still fail inside weak systems.
