Agency Client Onboarding Playbook
A repeatable, seven-step playbook for getting any new client set up and delivering first insights within five working days.
The first week sets the tone for every client relationship. A structured onboarding process means you deliver real data faster, avoid rework, and give clients a concrete reason to stay engaged from day one.
What you'll learn
- The seven steps to complete in the first week (the timeline is a guide; actual pace depends on client availability and keyword volume)
- What data to collect before touching Outercite
- How to frame the first insights report so clients understand what they are looking at
Why a repeatable process matters
Onboarding done ad hoc creates problems downstream. Missing keywords mean the first month of data is incomplete. Undefined competitors mean the baseline comparison is useless. A messy workspace structure means reporting takes longer than it should.
A repeatable playbook fixes all of that. You run the same seven steps for every client, adapt the specifics, and arrive at the first review with clean data and a clear story.
Outercite tracks citations across six AI engines: ChatGPT, Claude, Perplexity, Google AI, Grok, and DeepSeek. Each engine has different citation behaviour. Perplexity is the strongest citer at 58%, Google AI the hardest at 38%. Knowing this before you start helps you set accurate expectations with clients.
The seven-step onboarding
Step 1: Client intake call (suggested: Day 1)
Before you open Outercite, spend 30 minutes with the client covering four questions:
- What does their customer actually ask an AI engine? Think about the buying journey: research queries, comparison queries, local queries. A dental practice needs different prompts to a B2B SaaS company.
- Who are their top three to five competitors by customer perception? Not just market leaders. Who do their customers actually compare them to?
- Which geographies matter? Local, national, or global. This affects how you frame prompts.
- What is the one business outcome they care about? More bookings, more leads, more brand recognition. This becomes your north-star metric.
Record the answers. You will use them in Step 3.
Step 2: Create the client workspace (suggested: Day 1)
Go to /clients/add and create a new client workspace under your parent organisation. Name it clearly: use the client's trading name, not an abbreviation only you will recognise.
Assign the workspace to the right team members. If you use white-label reports, configure the branding now rather than before the first report deadline.
Step 3: Set up prompts and keywords (suggested: Day 2)
This is the most important configuration step. Prompts are the questions Outercite will use to check whether the client appears across all six engines.
Start with 20-30 prompts. A good set covers:
- Informational queries: "What is the best [category] in [city]?" or "How do I choose a [product type]?"
- Buying intent queries: "Best [product/service] for [use case]"
- Comparison queries: "[Brand A] vs [Brand B]"
- Branded queries: The client's own business name, to confirm it appears when someone asks for it directly
Add prompts via /prompts. Group them logically. You can always add more once the first verification cycle runs and you see which categories produce the most activity.
Step 4: Add competitors (suggested: Day 2)
Add the three to five competitors identified in the intake call. Outercite tracks their citation rates alongside your client's, so you can show share of voice and competitor threat scores in every report.
Go to /competitors to add them. Use the exact brand name as they appear in AI answers, not just their domain.
Do not add too many competitors upfront. Five is a practical ceiling for a first setup. A crowded competitor list makes the baseline comparison harder to read and can distract from the core story in early client reports.
Step 5: Let the verification pipeline run (suggested: Days 2-3)
Once prompts and competitors are configured, Outercite begins running checks. The verification pipeline classifies each prompt by intent, runs it across the six engines, and applies two-model cross-verification to rule out false positives.
Give it at least 24 hours before drawing conclusions. You want enough checks across all six engines to see a meaningful pattern. The aggregate citation rate across the platform sits at about 48%. Your client's early numbers may differ; an initial baseline is a starting point, not a final verdict.
Step 6: Build the first insights summary (suggested: Day 4)
With 24-48 hours of verified data, you have enough to build a first summary. Cover three things:
- Where they stand today. Citation rate per engine, overall visibility score, any immediate wins (engines where they already appear well).
- Where the gaps are. Engines where their citation rate is low, prompt categories where they do not appear at all.
- What to watch. Set up surge alerts at /settings/notifications for keywords where a competitor appears but the client does not. This means you will know immediately if the landscape shifts.
Keep this summary to one page or one section of your report. It is a baseline, not a deep analysis. Frame it as: "This is where you stand. Here is what we will do about it."
Step 7: First client call (suggested: Day 5)
Walk the client through the summary. Avoid jargon: explain citation rate and visibility score in plain terms before you show the numbers. The goal of this call is not to impress with data. It is to get the client to understand the baseline and agree on what you are trying to change.
End the call by confirming:
- The prompts you have set up accurately reflect what their customers ask
- The competitors you have added are the right ones
- The one or two metrics you will report on monthly
Once the client confirms those three things, the engagement is properly scoped. See the monthly client review template for how to structure ongoing reporting.
A note on expectations
Early data is a starting point. A citation rate of 20% in week one is not a failure. It is a measurement. The value you deliver is moving that number over time and explaining why it is moving. Set this expectation clearly on the Day 5 call, and clients will not be disappointed by an imperfect baseline.
Try this in Outercite
Start your next client onboarding from the agency portfolio view at /agency. Create the workspace, add 20 prompts, and let the first verification cycle run overnight. By tomorrow morning you will have verified baseline data to bring to your intake meeting.
Related
Pitching AI search to clients
How to present AI search visibility to prospects and walk away with a signed brief.
How to onboard a new client
Step-by-step UI walkthrough for setting up a new client workspace in Outercite.
Monthly client review template
A structured approach to monthly reviews: what to show, how to narrate it, and which metrics matter.
Schedule client reports
How to automate monthly or weekly reports so they land in your client's inbox on time.
Pitching AI Search Visibility to Clients
How to present AI search visibility to clients who have never heard of it, and walk away with a signed brief.
Monthly Client Review Template
A structured template for running monthly AI search reviews with clients: what metrics to show, how to narrate the story, and how to present wins clearly.
