CDL Masterclass 0%
Welcome

Becoming the Must-Hire Candidate

Chloe — this course was designed specifically for you. Your presentation is polished and ready. But the 5-7 minute presentation is the warm-up — the 50 minutes of probing questions that follow is where you win or lose the job.

The good news: you already think like a senior CDL more than you realise. Your Fixflo experience — the notification bug investigation, the pushback on custom fees, the marketplace growth from zero to £90k MRR — is exactly the kind of thinking Kraken wants. This course helps you connect what you already do to what they're looking for, and fills in the gaps where the case study demands new skills.

"This is the difference between we hope she gets the job and she's definitely getting the job."
— From your coaching session

Before we start, a scenario:

It's Thursday, 5:15pm. Your phone rings. The Domain Owner for Commercials at United Energy is on the line, panicking. "20,000 customers are on the wrong product. This is a Kraken issue and we need it fixed immediately."

What do you do first?

A Immediately escalate to the Transformation Director and loop in Engineering
B Set up a cross-functional meeting for Monday to discuss the issue
C Explain that this is likely a Systems Integrator issue, not a Kraken problem
D Acknowledge their concern, commit to investigating and calling back by 7pm with facts

If you chose D, you're already thinking like a senior CDL. If you chose A, B, or C — that's exactly why this course exists. Over the next 7 modules, you'll learn the thinking process that separates "competent" from "must-hire."

How this course works

  • 1
    7 teaching modules — each covers a core CDL concept with the WHY behind it, not just the what
  • 2
    Quizzes after each module — scenario-based questions that test understanding, with immediate feedback
  • 3
    Interview tips throughout — connecting each concept to questions they'll actually ask you
  • 4
    Final assessment — a multi-step scenario that requires applying everything you've learned
  • 5
    Interview rehearsal mode — practice answering questions out loud with a timer, then self-assess against key points

Your progress is saved automatically. You can close the browser and pick up where you left off.

Module 1

First Question: "Is This Real?"

When a panicked stakeholder calls with a crisis, every instinct tells you to act immediately. Escalate. Get people involved. Start fixing.

That instinct is the single biggest trap for a Client Delivery Lead.

You Already Do This at Fixflo

Remember when a contractor reported they'd stopped receiving notifications? You didn't just log a ticket and escalate — you tested it yourself first. You ran test scenarios with the contractor, reproduced the issue, and narrowed it to specific notification types before raising it with engineering. That's validation-first thinking. The case study demands the same instinct at a bigger scale.

A panicked phone call from a domain owner is a claim, not a confirmed incident. The worst thing you can do is trigger a chain of escalation across three companies based on unvalidated information.

Why validation matters

Think about what happens if you escalate immediately:

If it's not real

  • !
    You've mobilised three companies for nothing
  • !
    You look reactive in front of your boss and the client
  • !
    You've caused an expensive fire drill based on one phone call

If you validate first

  • ✓
    You know the real scope before involving others
  • ✓
    You escalate with data, not a phone call
  • ✓
    You protect both Kraken and the client from overreaction

The three things to do on that call

  • 1
    Acknowledge genuinely — "I hear you, this is serious and I'm taking it seriously." They need to feel heard before they can hear you.
  • 2
    Commit to a specific timeline — "Give me 90 minutes. I'll come back to you with what I know by 7pm." Not "I'll look into it" — a concrete commitment.
  • 3
    Don't speculate or defend — Don't say "that doesn't sound right" or "it might be the SI's fault." Just commit to finding out.
"The job of the client delivery lead here is to say: how can I quickly validate what is real here? What the real problem is. Because you've just basically called and started shouting. How do I work out what the truth is to work out the next step?"
— From your coaching session

Ask them to hold off escalating too

This is subtle but powerful: ask the Domain Owner to hold off escalating to their own leadership until you call back at 7pm. Why?

  • •
    It protects them from looking reactive to their bosses
  • •
    It prevents the political fire from starting before you have facts
  • •
    It shows you're thinking about their interests, not just Kraken's
They WILL ask: "Walk us through the first 90 minutes after that Thursday afternoon call." Your answer should demonstrate the validation-first mindset. Lead with "the first thing I do is listen" and "I don't immediately escalate." If they probe further, bridge to Fixflo: "This is exactly how I handle urgent contractor issues — I validate the scope myself before escalating, because mobilising engineering based on incomplete information wastes everyone's time." This is the #1 way to differentiate yourself.

Remember This

  • A panicked call is a claim, not a confirmed incident
  • Acknowledge, commit to a timeline, don't speculate
  • Ask the Domain Owner to hold off escalating until you have data
  • You protect the company AND the client by validating first
  • This thinking process is what separates senior from junior CDLs

Check Your Understanding

Scenario: The Domain Owner says 20,000 customers are on the wrong product and wants you to immediately brief the Transformation Director. You...
A Brief the Transformation Director immediately with what the Domain Owner told you
B Acknowledge the concern, commit to calling back by 7pm with validated data, then investigate
C Set up a war room with the SI, Kraken engineering, and UE's technical team
D Tell the Domain Owner to calm down and that it's probably not as bad as they think
During the initial call, the Domain Owner asks: "Is this Kraken's fault?" How do you respond?
A "Based on the data flow, this is most likely a Systems Integrator issue, not ours."
B "Yes, it's likely our fault and we'll take full responsibility for fixing it."
C "I understand why you'd think that — the products do live in our platform. Let me trace the full data flow so we can work out together where this went wrong."
D "I can't comment on fault until we've completed a full investigation."
Why do you ask the Domain Owner to hold off escalating to their own leadership?
A To buy Kraken more time before leadership gets involved
B It protects them from looking reactive to their bosses, and ensures leadership conversations happen with data, not panic
C To prevent the information from spreading before you can control the narrative
Module 2

Rapid Investigation — 90 Minutes

You've acknowledged the Domain Owner, committed to 7pm, and bought yourself 90 minutes. Now you need to work out what's actually happening. This is where the real CDL skill shows.

Your Fixflo Investigation Pattern

You already have an investigation instinct. When the email issue came in at Fixflo, you didn't just pass it to engineering — you tested whether the contractor could receive emails outside the system, dug into the root cause, and found that someone who'd left had the wrong email address configured. That's the same structured approach this module teaches, just applied to a bigger dataset. The difference at Kraken: you use AI to do in 90 minutes what would normally take a data analyst a day.

The shape of the problem determines the response. Is it 20,000 or 200? Have any been billed? One cohort or systemic? Every answer changes what you do next.

The 4-Step Validation Sequence

These four checks run in order because each one informs the next. (Look at slides 3-4 of your presentation — the investigation dashboard and impact assessment are built from these steps.)

  • 1
    Pull the product mapping table — This translates legacy product codes into Kraken product IDs. If a mapping is wrong, it tells you the scope immediately. This is where AI accelerates: feed it the mapping table and it flags anomalies across 20,000 accounts in seconds.
  • 2
    Check billing status — Have any of these customers actually been billed on the wrong product? This is the critical question because billed = financial harm to customers. It changes the urgency of everything that follows.
  • 3
    Cohort analysis — Is this the latest migration batch only? Historical too? One product mapping or multiple? This tells you whether you're dealing with an isolated error or a systemic problem.
  • 4
    Sample check 50 accounts — Manually verify a random sample against legacy records. "Wrong product" in a migration often just means "differently named in the new system." You need to check actual rates — are customers being charged differently from what they agreed to?
"Finding out that truth as a client delivery lead tells you how you would lead the rest of the project. If no one's been billed and no one's being billed for 28 days, that billing team needs to be notified — but not tonight at 11pm. If half the people have been billed, we got to pause everything right now."
— From your coaching session

The Fork: What your data tells you

After 90 minutes, you know three things: how many accounts are actually affected, whether anyone's been billed, and the financial exposure. Now you're at a decision point:

Confirmed & Significant

Full escalation. Brief your TD, the billing CDL, the comms CDL. This is the rest of your presentation — contain, correct, prevent.

Not Confirmed or Overstated

Share the evidence transparently. Invite the client to verify your findings. Protect the relationship while protecting Kraken.

In the case study scenario, your investigation reveals: 12,847 confirmed affected (not 20,000), 3,214 already billed, and £847K financial exposure. It's real and significant — but the Domain Owner's number was wrong by 36%. That matters.

They WILL ask: "What if this had happened on a Friday evening — no one's available until Monday?" Your answer should show urgency calibration: pause migrations and confirm billing status before the weekend (these are time-critical). The detailed investigation can wait until Monday — UNLESS billing is imminent, in which case you make the uncomfortable Friday night calls.

Remember This

  • Four checks in order: mapping table, billing status, cohort analysis, sample 50 accounts
  • The shape of the problem determines the response
  • Billed vs unbilled changes EVERYTHING about urgency
  • Your validated data may tell a different story than the Domain Owner's claim (12,847 vs 20,000)
  • AI accelerates the investigation but you make the judgement calls

Check Your Understanding

Put these investigation steps in the correct order:
Check billing status — have any customers been billed?
Sample check 50 accounts against legacy records
Pull the product mapping table
Cohort analysis — one batch or systemic?
Correct order: 1) Product mapping table (defines the scope), 2) Billing status (determines urgency), 3) Cohort analysis (determines breadth), 4) Sample check (verifies the reality of actual rates vs agreed tariffs). Each step informs the next — you can't assess billing urgency until you know the scope, and you can't sample-check until you know which cohorts to look at.
Your investigation reveals 200 affected accounts, none billed, all from one migration batch that ran yesterday. How does this change your response compared to 12,847 with 3,214 billed?
A Same response — any number of wrong accounts requires full escalation
B Much smaller response — fix the mapping, correct the 200 accounts, add a validation check, and brief the Domain Owner with the good news that it's contained
C Tell the Domain Owner they overreacted and it's not a real problem
Your investigation shows the claim is overstated — only 47 accounts are affected and none have been billed. The Domain Owner is still insisting it's a major crisis. What do you do?
A Tell the Domain Owner their initial assessment was wrong and the issue is minor
B Escalate anyway since the Domain Owner is still concerned
C Share your evidence transparently, walk them through your methodology, and invite them to verify your findings
Module 3

Root Cause Analysis & Data Flow Thinking

Your investigation confirmed 12,847 accounts are affected and 3,214 have been billed. Now you need to find where it went wrong. Acting on the wrong root cause is worse than not acting at all — if you fix the wrong thing, you could make 12,847 accounts wrong a second time.

You've Done This Before

The Fixflo notification bug is a perfect parallel. A contractor reported "notifications aren't working" — that was the symptom. You didn't just say "notifications are broken, fix it." You traced the data flow: which notification types were affected? Where in the system were they triggered? You found it was a webhook misconfiguration — a specific technical root cause, not a vague system issue. That's root cause analysis. At Kraken, you're tracing a different data flow (legacy → SI → Kraken API) but the thinking is identical.

You're not here to assign blame. You're here to find the failure point with data, so you can fix it with confidence.

The 5-Stage Data Flow

Every customer account in Kraken got there through this pipeline. The error entered at one of these stages. (This maps directly to slide 5 of your presentation — the root cause hypotheses.)

  • 1
    Legacy Systems — The source data. Was the legacy data already wrong? If customers were on the wrong product in the old system, the SI faithfully extracted bad data. Owner: United Energy + SI
  • 2
    Systems Integrator — Mapping Table — This translates legacy product codes into Kraken product IDs. A single wrong row affects every customer in the cohort. Most likely failure point. Owner: SI + United Energy
  • 3
    SI Pipeline Logic — Even if the mapping table is correct, there could be a bug in the transformation logic that applied it wrongly. Owner: SI
  • 4
    Kraken API & Product Config — Were the products configured correctly in Kraken? If Product A and Product B were set up with swapped details, the mapping could be right but the target is wrong. Owner: United Energy + Kraken
  • 5
    Customer Account — An API processing anomaly at Kraken's end. Least likely, but check API request logs against customer accounts. Owner: Kraken Engineering

Why you need hypotheses, not conclusions

Notice that the presentation deliberately presents five hypotheses impartially. It doesn't say "it's the SI's fault." Why?

Diplomatic reason

Walking into the client conversation saying "it's your SI's fault" before you have data will make United Energy defensive and the SI uncooperative. The structured investigation approach gets you to the same answer but with evidence everyone has to accept.

Technical reason

Acting on an assumption with 12,847 customer accounts is dangerous. If you're wrong and it's a Kraken product configuration issue, a correction based on the wrong root cause makes things worse — now you've modified 12,847 accounts incorrectly twice.

"She needs to prove that this is the way she thinks. She thinks about validating assumptions rather than just presenting assumptions."
— From your coaching session

But you DO have an instinct

Having hypotheses doesn't mean having no opinion. The SI mapping table (hypothesis 2) is the most probable because the pattern matches: 12,847 accounts, all on the same wrong product, all from the latest migration batch. A single wrong row in a mapping table would produce exactly this pattern.

The senior CDL approach: have a strong instinct but hold it back deliberately. Present the structured investigation, then if asked "which is most likely?" you have a reasoned answer ready.

They WILL ask: "Which of those five hypotheses do you think is most likely?" Best answer: "If I had to rank them, I'd put the SI's mapping table first. A single wrong row would affect every customer consistently — which matches the pattern. But I deliberately didn't lead with that because acting on an assumption with 12,847 accounts is dangerous, and pointing fingers before having data makes the SI uncooperative."

Remember This

  • Trace the 5-stage data flow: Legacy → SI Mapping → SI Pipeline → Kraken Config → Customer Account
  • Each stage has an owner and a verification method
  • Present hypotheses, not conclusions — for diplomatic AND technical reasons
  • Have an instinct (SI mapping table is most likely) but hold it back until asked
  • The goal is to find the failure point with data, not to assign blame

Check Your Understanding

The SI insists their mapping table is correct. What evidence do you request to verify?
A Ask them to double-check their work and report back
B Request the mapping table itself, raw API payloads for the affected cohort, and a sample of legacy source records to cross-reference all three
C Escalate to the SI's programme manager to apply pressure
Why did the presentation deliberately NOT name the most likely root cause during the investigation section?
A Because you genuinely don't know which hypothesis is correct
B Both diplomatic (avoids making SI defensive) and technical (wrong root cause = wrong correction on 12,847 accounts) reasons
C Purely to avoid alienating the Systems Integrator
You discover that Kraken DID contribute to the problem — a product was misconfigured in the platform. How do you handle this?
A Focus the investigation on the SI's mapping table and downplay Kraken's contribution
B Immediately apologise to the client and accept full responsibility
C Present the findings honestly to all parties, own Kraken's part proportionally, and show what you're changing to prevent it
Module 4

AI as a Skill Multiplier

This module is arguably the most important for differentiating yourself. Every candidate at Kraken will mention AI. The question is how you mention it.

Chloe — Let's Be Honest About Your Starting Point

Right now, you use AI mostly for polishing emails and drafting communications. That's the "AI for speed" column below — useful, but every candidate does this. The case study presentation shows you using AI for investigation: building dashboards, analysing account data, scenario modelling. You don't need to be an expert at this yet — but you need to articulate the vision credibly and show you understand the difference. Your presentation already demonstrates it; this module helps you defend it when they probe.

Everyone will say "AI helps me with comms." Everyone. That doesn't differentiate you. The question is: how do you use AI to make yourself better at your job — not just faster?

Speed vs New Capability

There are two fundamentally different ways AI helps you:

AI for Speed

Makes you faster at things you can already do:

  • •
    Drafting emails and internal briefings
  • •
    Writing customer service scripts
  • •
    Creating presentation drafts

Useful but not differentiating. Everyone does this.

AI for New Capability

Gives you skills you don't currently have:

  • ★
    Data analysis & SQL generation to investigate incidents
  • ★
    Building dashboards to visualise account data
  • ★
    Scenario modelling for financial exposure

This is what makes you indispensable.

"If you can do the investigation yourself with AI, I need to hire you. If you're just going to call Sartac and he's already busy, I don't need you. I can call Sartac myself."
— From your coaching session

Show, don't tell

The second critical principle: don't have an "AI section." Weave AI throughout your work naturally.

In your presentation, AI appears everywhere — and the reveal on slide 8 only works because they didn't see it coming:

  • 1
    The validation dashboard — built with AI in ~20 minutes. SQL generation, data analysis, and visualisation. This IS the investigation — it's not a separate "AI" thing.
  • 2
    The client email — AI-drafted, then refined for tone and regulatory compliance. Mentioned naturally in the comms section, not as an "AI showcase."
  • 3
    The investigation itself — AI flagged anomalies across 20,000 accounts in seconds. This is what makes you capable of doing the investigation solo in 90 minutes instead of needing a data analyst.
  • 4
    The presentation itself — AI-assisted throughout. Only revealed at the end.

The Reveal

"I don't talk about AI. I just use it. You've already seen AI throughout this presentation. You didn't notice because it just looked like competent delivery. That's the point." — This is the most powerful slide in the presentation. It reframes everything that came before.

They WILL ask: "How do you use AI day-to-day?" Don't start with email drafting — that's everyone's answer. Instead: "I've been evolving my AI usage from communications into investigation and analysis. This case study is a good example — I used AI to build the investigation dashboard, categorise affected accounts, and model the financial exposure. That's the shift I'm most excited about: using AI to give myself capabilities I didn't have before, not just to do existing work faster." This is honest (you ARE evolving) while positioning you at the frontier.

Remember This

  • Everyone will mention AI for comms. That doesn't differentiate you.
  • AI for speed = faster. AI for new capability = indispensable.
  • Show, don't tell: weave AI throughout, don't have a separate "AI section"
  • The investigation dashboard is the killer example — a skill you gained through AI
  • "I don't talk about AI. I just use it."

Check Your Understanding

Classify each AI use case: does it make you faster (speed) or give you a new capability?
Building an incident dashboard from raw SQL data
Drafting a client email
Analysing 20,000 accounts for anomalies in seconds
Creating a customer service call script
Scenario modelling financial exposure
Polishing presentation slides

Click each chip to toggle: amber = Speed | cyan = New Capability

Speed (amber): Drafting emails, creating call scripts, polishing slides — you can already do these things, AI just does them faster.

New Capability (cyan): Building dashboards from SQL, analysing 20,000 accounts, scenario modelling — these are skills you gain through AI that you didn't have before. This is what makes you a CDL who can investigate independently instead of waiting for a data analyst.
In the interview, they ask how you'd use AI day-to-day. Which answer is stronger?
A "AI helps me draft stakeholder communications, create presentations, and write internal briefings. It saves me about 2 hours a day on writing tasks."
B "The biggest value is in investigation. When a client calls with a data issue, I can pull account data into AI, generate analysis, build a dashboard, and have the shape of the problem within 90 minutes — without waiting for a data analyst. AI also helps with comms, but honestly that's the least interesting use case."
Why does the presentation NOT have a dedicated "AI" section or slide?
A To avoid drawing attention to the fact that AI was used
B Because the case study didn't ask about AI
C Because AI woven throughout is more powerful than AI in a separate section — "I don't talk about AI, I just use it"
Module 5

The Context Layer — Building Informed AI

Module 4 covered what AI can do for you. This module covers how to make AI do it well. The difference between a mediocre AI output and a genuinely useful one is context.

You Already Build Context — Just Not for AI Yet

At Fixflo, you use Pendo for in-app research and customer questionnaires, and Miro for customer journey mapping and workflow documentation. You build up knowledge about your contractors and agents over time. That's the exact same principle as the context layer below — accumulating structured knowledge about your clients and their processes. The difference is feeding that accumulated knowledge into AI so it can work with your context, not from cold.

When AI drafts arrive informed, they arrive good enough to refine rather than rewrite. When they arrive cold, you spend more time fixing them than you would have writing from scratch.

The Four Context Files

You maintain four living documents that build up over time. Every week, you add to them. When you need AI to help with something, you feed it the relevant context first.

Incident Log

Running record of issues, resolutions, and patterns. When a new incident occurs, AI can reference similar past incidents and how they were handled.

Stakeholder Profiles

Who are the key people, what do they care about, how do they communicate? When AI drafts an email to the TD, it knows their preferences and concerns.

Comms History

Previous emails, Slack messages, meeting notes. AI can match your established tone and reference prior conversations naturally.

Decision Log

What was decided, by whom, and why. When AI prepares you for a meeting, it knows the history of how you got here.

Cold vs Informed: The Difference

Here's what the same AI draft looks like with and without context:

Cold Draft (no context)

"Dear Domain Owner, I wanted to update you on the product mapping issue. We are currently investigating and will provide further updates as they become available. We take this matter seriously and are working to resolve it promptly."

Generic, corporate, could be about anything. Requires a full rewrite.

Informed Draft (with context layer)

"Hi Sarah, following up on our call — I've completed my initial investigation. The picture is narrower than initially reported: 12,847 confirmed affected (not 20,000), with 3,214 already billed. I've identified the likely root cause in the SI's product mapping and am preparing a remediation plan. I'll have a full brief for your leadership meeting on Tuesday. Happy to jump on a call before then if you'd like to walk through the dashboard I've built from the data."

Specific, personal, references prior conversation. Needs refinement, not rewriting.

"The way I work with AI is I build up a context layer. Every week we write everything we know about the clients. So then when AI is helping me write to that client, it knows who they are. I'm not a ChatGPT copy-paste idiot saying 'help me make this email more professional.'"
— From your coaching session
They WILL ask: "How do you use AI for client communication?" Your answer should describe the context layer system. Explain that you build up knowledge about each client over time — stakeholder profiles, comms history, decision logs — so when AI helps you draft, it arrives informed. The key phrase: "good enough to refine, not rewrite."

Remember This

  • Four context files: Incident Log, Stakeholder Profiles, Comms History, Decision Log
  • Build them up over time — they compound in value
  • Cold AI drafts need rewriting; informed AI drafts need refining
  • This system is what separates "copy-paste from ChatGPT" from genuine AI-native working
  • Be ready to explain WHY this approach works in the interview

Check Your Understanding

You're drafting an urgent email to the Domain Owner about the investigation findings. Which context files would make the AI draft most useful?
A Just the Incident Log — it has all the technical details
B Stakeholder Profiles (who they are and what they care about), Comms History (previous tone), and Incident Log (current facts)
C Incident Log and Decision Log only
An interviewer asks: "Why not just use ChatGPT to write your emails?" What's the best response?
A "I prefer to write my own emails because they have more personality."
B "I do use AI for drafting, but the quality depends on context. I maintain stakeholder profiles and comms history so the AI knows who it's writing to. A cold draft needs rewriting. An informed draft just needs refining."
C "ChatGPT doesn't know the client or the situation, so the outputs aren't useful enough."
Why is "good enough to refine rather than rewrite" the goal, not "perfect first draft"?
A Because AI isn't good enough for perfect drafts
B Because the human refinement layer is where judgement, tone calibration, and political sensitivity come in — AI gets you to 80%, you close the last 20%
C Because it saves more time than trying to get the prompt perfect
Module 6

Stakeholder Communication — Same Facts, Different Tones

Communication determines whether this incident damages the relationship or strengthens it. The facts don't change between audiences — but how you present them changes everything.

You Already Manage Multiple Audiences

At Fixflo, you're the bridge between contractors (who need operational clarity), property managers (who need confidence the system works), engineering (who need precise bug reports), customer success (who need product knowledge), and leadership (who need revenue and progress updates). You already calibrate your tone for each audience. The case study has the same five-audience challenge — the names change but the skill is identical. Look at slide 7 of your presentation where you have the different email previews for each audience.

Each stakeholder has different needs, fears, and decision-making authority. They need to hear the same truth through a different lens.

The Five Audiences

1. Domain Owner (United Energy)

Tone: Calm & Empathetic — "I hear you. Here's what I'm doing right now."

Why: They're panicking. They need to feel heard before they can hear facts. De-escalate first, then brief. If you lead with data before acknowledging their stress, they won't process it.

2. Transformation Director (United Energy)

Tone: Factual & Structured — "Here's the situation, the likely cause, and my plan."

Why: They need to brief their board. Give them a structured summary they can forward. Situation, impact, root cause hypothesis, plan, timeline. No waffle.

3. Systems Integrator

Tone: Collaborative — "We need your data. Let's investigate together."

Why: They're most likely the key contributor. If you make them defensive, they slow down. You need them moving fast and sharing data freely. Frame it as a joint investigation, not an accusation.

4. Customers (if billed)

Tone: Apologetic & Transparent — "We're sorry. We've fixed it. You won't be out of pocket."

Why: Customers don't care about Kraken, SIs, or data mapping. They care about their bill. No jargon, no system names. Proactive — get ahead of complaints.

5. Kraken Leadership

Tone: Confident & Solution-focused — "Contained. Plan in place. Relationship managed."

Why: Your TD needs to know the right person is handling this. They need the unvarnished picture: status, numbers, recommendation. Fast, searchable, async — Slack, not a formal email.

Channel matters too

The internal update goes via Slack — your TD needs it fast, searchable, async. The client update goes as an email — it's a record, it shows formality, and they can forward it to their leadership.

They WILL ask: "How would you communicate this to United Energy's customer service agents?" This tests whether you think beyond leadership. Your answer: work with the Customer Journeys CDL to prepare a briefing note (what happened, plain language), a call script (what agents say if customers call), and an escalation path (who to go to if the agent isn't sure). AI drafts all three in minutes.

Remember This

  • Five audiences, five tones: empathetic, factual, collaborative, apologetic, confident
  • De-escalate the Domain Owner before diving into data
  • The SI needs collaboration, not accusation — you need them moving fast
  • Customer comms: no jargon, proactive, and sent by United Energy (their relationship)
  • Internal comms via Slack (fast/async), client comms via email (formal/forwardable)

Check Your Understanding

Match the message to the correct audience. This message was written for whom?

"Following up on our call — I've completed my initial investigation. The picture is narrower than initially reported. I'll walk you through the dashboard I built from the data. Can we talk before your 7pm?"
A Domain Owner (United Energy)
B Transformation Director (United Energy)
C Systems Integrator
D Kraken Leadership
This message was intended for the Transformation Director, but the tone is wrong:

"Hi, just wanted to give you a heads up — we've found some product issues in the latest batch. Nothing too crazy, probably about 12,847 accounts. Let me know if you want to chat about it!"

What needs to change?
A Just make it more formal with "Dear" and "Regards"
B Restructure as: Situation (12,847 confirmed, 3,214 billed), Impact (£847K exposure), Root cause hypothesis, Remediation plan with timeline. Remove the minimising language.
C Add more technical detail about the root cause
The Systems Integrator is getting defensive. Their technical lead says: "Our mapping table is fine, this must be a Kraken configuration issue." Your next message should be:
A Escalate to their programme manager and request a formal review
B Share your dashboard showing the SI mapping is the most likely root cause
C "Understood — let's rule that out. Can you share the mapping table so we can cross-reference it with UE's product catalogue? We're checking our side too. Let's eliminate possibilities systematically."
Module 7

What Separates Senior from Standard

This final teaching module ties everything together. Every concept you've learned feeds into one question the interviewers are silently asking: "Does this person think like a senior CDL?"

Chloe — You Already Think This Way

When that property manager wanted custom works management fee percentages, you didn't just say no. You investigated the commercial case, assessed ROI against development effort, and framed the pushback in terms they could accept. When enterprise customers later requested reporting features, you built the business case because the data justified it. That's senior CDL thinking: validate the request, assess proportionality, act on evidence. The case study demands the same pattern — just with a crisis instead of a feature request.

Anyone can follow an escalation path. A senior CDL decides IF and HOW to escalate. The thinking process before execution is the entire differentiator.

The Junior vs Senior Pattern

Across every module, there's a consistent pattern:

The Standard CDL

  • ✕
    Gets the call → escalates immediately
  • ✕
    Presents assumptions as facts (20,000 customers)
  • ✕
    Tells the SI it's their fault before having evidence
  • ✕
    Mentions AI as a separate section: "and then I also use AI for emails"
  • ✕
    Uses the same tone with every audience

The Senior CDL

  • ✓
    Gets the call → validates first, escalates with data
  • ✓
    Distinguishes claims from facts (12,847 confirmed, not 20,000 claimed)
  • ✓
    Presents hypotheses diplomatically, has an instinct ready if asked
  • ✓
    Uses AI throughout naturally — "I don't talk about AI, I just use it"
  • ✓
    Calibrates tone, channel, and detail level per audience

Urgency Calibration

One of the most senior skills: understanding that the investigation determines the urgency level, and different findings demand different responses.

The same incident, two realities:

Reality A: No one billed, billing in 28 days

Notify the billing team — but not tonight at 11pm. This is urgent but not emergency. Schedule the correction for the coming week. The Domain Owner gets a reassuring call. Your TD gets a structured plan.

Reality B: 3,214 already billed, billing run tonight

Pause everything NOW. This is a Friday night phone call to engineering. Every hour of delay is more customers being billed incorrectly. The Domain Owner gets an immediate callback. Your TD gets an out-of-hours escalation.

"Everybody else will just go straight into execution mode. This is the number one way to differentiate yourself. The number two way is to show AI making you better, not just faster."
— From your coaching session

Protecting the Organisation

A senior CDL protects the organisation from two things:

  • 1
    Unnecessary fire drills — Mobilising three companies based on unvalidated information wastes time, money, and credibility. Validation is protection.
  • 2
    Wrong corrections — Acting on the wrong root cause with 12,847 accounts makes things worse, not better. Hypotheses are protection.
They WILL ask: "What makes you different from other candidates?" Your answer should weave together the validation-first mindset, AI-powered investigation, and urgency calibration. Something like: "I don't jump to execution. I validate first, which protects both the company and the client. I use AI to give myself investigation capabilities that let me act independently. And I calibrate my response to what the data actually shows, not what someone told me on the phone."

Remember This

  • The thinking process before execution is the entire differentiator
  • Senior CDLs validate, then act. Standard CDLs act, then hope.
  • Urgency calibration: same incident, different findings, different response
  • You protect the organisation from fire drills AND from wrong corrections
  • This is what makes you "must-hire" vs "one of a few decent people"

Check Your Understanding

Two CDLs respond to the same incident. Which response is senior-level?
A CDL A: "I immediately briefed my TD, looped in the billing CDL, contacted the SI, and set up a Monday morning war room. I've started scoping the bulk correction for 20,000 accounts."
B CDL B: "I acknowledged the Domain Owner's concern and committed to calling back by 7pm. In 90 minutes, I used AI to analyse the account data and found 12,847 confirmed affected — not 20,000. 3,214 were already billed. I called back with facts and a tiered response plan based on the actual scope."
No one has been billed and the next billing run is 28 days away. The issue is confirmed — 12,847 accounts on the wrong product. Do you call the billing team tonight?
A Yes — 12,847 accounts is serious enough to warrant immediate out-of-hours escalation
B No — send a Slack message to the billing CDL tonight flagging the issue, then have a structured conversation Monday morning about the correction timeline
C No — wait until Monday to tell anyone about the billing implications
The interviewer asks: "What makes you different from other candidates?" Which answer best captures the senior CDL mindset?
A "I have 3+ years of client-facing delivery experience managing complex stakeholder relationships across technical and commercial domains."
B "I use AI extensively for stakeholder communications, which lets me respond to clients faster and with more polished outputs than most people."
C "I don't jump to execution. I validate first, which protects both the company and the client. I use AI to give myself investigation capabilities — building dashboards, analysing data — so I can act independently. And I calibrate my response to what the data shows, not what someone told me on the phone."
Final Assessment

Putting It All Together

Chloe — this assessment tests whether you can apply all seven modules in a realistic scenario. Each question builds on the previous one, just like a real incident unfolds. Think of it as a dress rehearsal for the probing questions after your presentation.

The Scenario

It's Wednesday, 4:30pm. You're the CDL for Commercials on the United Energy programme. The Migration Lead (not the Domain Owner this time) messages you on Slack: "Hey, we've spotted something odd in the latest batch. Looks like a chunk of customers might have landed on the wrong product. The Domain Owner doesn't know yet. Can we chat?"

Question 1 of 5

The Migration Lead has flagged something before the Domain Owner knows. What's your first move?
A Call the Domain Owner immediately to give them a heads-up
B Get on a call with the Migration Lead immediately to understand what they've found, then run your rapid investigation before involving anyone else
C Schedule a meeting for tomorrow morning with the Migration Lead and your TD

Question 2 of 5

Your investigation reveals: 8,400 accounts on wrong products, 0 billed (next billing run is in 12 days), all from the same migration batch that ran Monday. The SI's mapping table has 2 incorrect rows. How would you describe the urgency level?
A Critical — out-of-hours escalation, pause everything tonight
B Urgent but contained — fix the mapping this week, correct accounts before the billing run, brief stakeholders tomorrow morning with data and a plan
C Low priority — no one's been billed, so handle it next week

Question 3 of 5

You need to tell the Domain Owner. But unlike the main case study, you're bringing the news TO them (they didn't discover it). How does this change your approach?
A Lead with the positive: "Good news — we caught something before it became a problem"
B Downplay the numbers to avoid causing the same panic as in the main scenario
C Be transparent with the full scope (8,400 accounts, 2 mapping errors, 12 days to billing) and present your remediation plan. The proactive discovery IS the good news — let it speak for itself.

Question 4 of 5

The SI's technical lead pushes back: "Those mapping rows are correct — the client changed the product catalogue last week and didn't tell us." If true, this shifts ownership to United Energy. How do you handle this?
A Accept the SI's explanation and tell the Domain Owner that UE's product catalogue change caused the issue
B Dismiss the SI's claim — you've already identified them as the likely root cause
C Verify the claim: "That's an important detail. Can you show me the mapping table version from before last week? And I'll check with UE whether a product catalogue change was made. If true, we need to fix the communication process between UE and the SI."

Question 5 of 5

The incident is resolved. In the post-incident review, the TD asks: "What would you change to prevent this?" What's the strongest answer?
A "The SI should do more thorough testing before each migration batch"
B "Three things: a pre-migration validation gate where UE signs off that product mappings match their catalogue before a batch runs; a formal change notification process so catalogue updates trigger a mapping table review; and automated post-migration reconciliation that compares accounts against expected products before billing."
C "Kraken should build commercial logic validation into the API so wrong products are rejected automatically"
Interview Rehearsal

Practice Out Loud

Chloe — the presentation is 5-7 minutes. The remaining 50 minutes is the real interview. This mode lets you practice answering questions out loud with a timer, then self-assess against the key points.

How it works: Read the question, take 5 seconds to think, then answer out loud for 2-3 minutes. When the timer ends, reveal the key points — these now include your specific Fixflo examples and STAR-format talking points. Rate yourself honestly.

General rule: Keep answers to 2-3 minutes. If they want more, they'll ask. Better to be concise and prompted for depth than to ramble and get cut off. Use "In this scenario..." and "In my experience at Fixflo..." to toggle between the case study and your real experience — this shows range.
Question 1 of 20

2:30

Key Points to Hit

    Things to Avoid

      How did you do?