Using the Portal
A plain-English guide to running clients, audits, and monitoring in Sextant.
This is the operating guide for your Sextant portal β how to actually run the thing, day to day. It covers adding a client, filling in their intake, running an audit, working the fix list, turning on continuous monitoring, collecting leads with the free audit tool, and giving your client their own login. It does not explain how to interpret an audit report β that has its own guide, "How to Read This Report," under Resources. This one is about the buttons, not the numbers. You don't need to read it end to end. Skim the headings and jump to the job you are trying to do.
1. Getting oriented β the sidebar and your Portfolio
Home β Portfolio (this is where you land after signing in)
What it is
Portfolio is your home base: every client you manage, who needs a look right now, and what has run recently. The sidebar groups everything else into Pipeline (free-audit leads), Agency (your own settings, team, and the free-audit tool), and Resources (this guide, the report guide, and the roadmap).
How to use it
Portfolio reads top to bottom in a deliberate order. "Clients" is your full roster β click any name to open that client. "Needs attention" appears below it and surfaces monitored clients worth checking on; when nothing qualifies it simply says you are all caught up, and if no client has monitoring turned on the section is hidden entirely. "Recent audits" at the bottom shows the latest completed and in-flight work, with links to view each report or download it as Word or Markdown.
- Use the Active / Archived toggle above the client table to switch views.
- On a wide screen, "Columns" lets you hide columns you do not use; your choice is remembered in that browser.
- "Sort by movement" appears only once at least one client is on a monitoring plan.
- A client showing an amber "β Complete Intake" chip has not had their intake filled in yet β see the intake section below.
On a phone the sidebar becomes four bottom tabs β Home, Pipeline, Agency, More β and the client table becomes cards. Everything is reachable, it just stacks differently.
2. Why your screen may not match a teammate's
Everywhere β it depends on the role assigned to your login
What it is
Sextant hides what you are not permitted to use rather than showing it grayed out. That means two people looking at the same client can genuinely see a different set of tabs, and neither screen is broken.
How to use it
Inside a client, the Portal, Intake, and Settings tabs only appear for people whose role includes "Manage Clients." Without it you will see Overview, Audits, Monitoring, Recommendations, and Insights and nothing further β the three management tabs are not dimmed, they are absent. The same rule applies in the sidebar: Team only appears with "Manage Users," Roles only with "Manage Roles." Within monitoring, read-only users can see every chart and table but the buttons that change things β Add prompt, Accept, Save, Generate β are not rendered.
So if a colleague says "click the Intake tab" and you have no Intake tab, you are not looking in the wrong place β you need the Manage Clients permission. Whoever administers your roles can add it under Agency β Roles.
3. Adding a client
Portfolio β "+ Add client" (top right)
What it is
Creating the client record. This is the minimum Sextant needs to run a first audit: who they are, where they are, and who they compete with.
How to use it
Click "+ Add client" and fill in the short form. Brand name, Website URL, Industry, and Location are all required; Competitors is optional and takes a comma-separated list. Industry and Location are not decorative β they are written directly into the questions we put to the AI engines, so "Denver, CO" and "Colorado" will produce different results. Use the phrasing a real buyer would use.
- Portfolio β "+ Add client".
- Enter Brand name and Website URL.
- Enter Industry (e.g. "SaaS, Marketing") and Location (e.g. "Denver, CO").
- Optionally list competitors, comma-separated.
- Click "Create Client".
After you create a client the dialog closes and the list refreshes β it does not jump you into the new client. Click the name in the table to open their workspace. Your next stop should be Intake.
4. Inside a client: what each tab is for
Portfolio β click a client name
What it is
Every client has their own workspace with a fixed set of tabs. Knowing which tab owns which job saves a lot of hunting.
How to use it
Overview is a quick read of the latest score and its five dimensions. Audits is the full history of every audit run, with a link into each report. Monitoring holds the continuous-tracking features and has its own row of sub-tabs. Recommendations is the fix list generated from the last audit. Insights collects your proven wins. Portal, Intake, and Settings are the management tabs and only appear if your role permits.
- Overview β latest composite score out of 100 and the dimension breakdown.
- Audits β every audit, its status, and a "View" link into the report.
- Monitoring β sub-tabs for Overview, Prompts, Sources, Competitors, Alerts and Setup.
- Recommendations β the generated fixes, and where you record what shipped.
- Insights β the "Proof of Impact" cards for recommendations you have measured.
- Portal β invite the client to their own white-labeled login.
- Intake β the brand questionnaire that grounds every audit.
- Settings β edit brand details, archive, export, or delete the client.
"Trigger Audit" sits in the client header, so it is available from every tab, not just Overview. The Insights tab shows nothing at all until you have measured at least one recommendation β a blank Insights tab is expected on a new client, not a fault.
5. The intake questionnaire β the highest-value 20 minutes you will spend
Client β Intake (requires Manage Clients)
What it is
A structured profile of the brand: who they are, how they position themselves, who they really compete with, what proof they can point to, and β most importantly β the actual questions their buyers ask. The screen sums it up in one line: structured brand context that feeds every audit prompt, where more detail means more accurate AI scoring.
How to use it
Work through the five groups: Company, Positioning, Competitors, Proof, and AI question targets. Every field has a one-line explanation underneath telling you what it is used for. The list fields accept one item per line or comma-separated values. Click "Save intake" when done β you can come back and add more at any time.
- Company β aliases the AI might use, founded year, headcount, client count, service areas.
- Positioning β category, one-liner, ideal customer, services, differentiators, certifications.
- Competitors β the real firms they compete with, name plus optional domain.
- Proof β review profile URLs (G2, Clutch), notable clients, awards.
- AI question targets β use cases, problems solved, tasks, known issues, error messages.
- Click "Save intake".
Intake does two things nothing else can. First, it unlocks deeper audit questions: the "AI question targets" lists and the Positioning category are what let the audit ask specific, buyer-shaped questions instead of only generic ones β leave them blank and those questions are skipped entirely. Second, it stops the report inventing things. Fields like founded year, headcount and client count exist so the write-up quotes a fact you supplied rather than guessing. A client with a thin intake gets a thinner, vaguer report. Fill it in before the first audit if you possibly can.
6. Running an audit and what to expect
Client β "Trigger Audit" (in the header, on any tab). Also "Run Audit" on each Portfolio row.
What it is
An audit puts a batch of real questions to five AI engines β Claude, ChatGPT, Perplexity, Gemini and Grok β crawls the client website, scores what came back across five dimensions, and generates the fix list. It is the core deliverable.
How to use it
Click "Trigger Audit". There is no confirmation dialog β the button reads "Queuingβ¦" for a moment and the job is away. The audit then runs in the background for several minutes; it is not instant, because it is genuinely waiting on five different AI services to answer.
- Open the client and click "Trigger Audit".
- Go to the Audits tab β a new row appears with status "pending", then "running".
- Wait a few minutes, then reload the page to see the status change.
- When the row reads "completed", click "View" to open the report.
The page does not refresh itself while an audit runs. If the status still says pending, reload the browser rather than assuming it is stuck. Reports are also downloadable as Word or Markdown from the "Recent audits" list on Portfolio, which is the easy way to get an editable copy for a client deck.
7. Reading the report with a client
Client β Audits β "View", then "How to read this report β" in the top right
What it is
The audit report is the document you walk a client through. It has its own dedicated plain-English companion guide, so this guide does not repeat it.
How to use it
Open any completed audit and click "How to read this report" in the header β it opens in a new tab so you can keep it beside the report during a call. It explains every section in order, from the composite score to the methodology notes. The same page is always available in the sidebar under Resources β Report guide.
That guide contains no client data, so it is safe to leave on screen or share while presenting.
8. Working the fix list
Client β Recommendations
What it is
After an audit completes, Sextant generates recommendations for the gaps it found β and each one contains the actual deliverable, not just advice. A schema recommendation contains the finished code. A blog-title recommendation contains the rewritten titles.
How to use it
Each card shows the fix, a badge for its track (On-site, Off-site or Profile), a status badge, and a one-line "why this matters." Click "View details" to expand it. Inside you get the artifact itself in a code block with a "Copy" button, and where relevant a "How to deploy" note explaining where it goes. Copy it, deploy it, then come back and record that you did β see the next section.
- Open Client β Recommendations.
- Click "View details" on a card to expand it.
- Read the "How to deploy" note if there is one.
- Click "Copy" to put the artifact on your clipboard, then deploy it.
- Come back and either Measure Impact or Mark shipped.
If the tab says recommendations are generated once an audit completes with gaps, that is exactly what it means β run an audit first. The list is newest-first and there are no filters, so scroll rather than search.
9. "Measure Impact" vs "Mark shipped"
Client β Recommendations, on each card
What it is
Two different buttons for two genuinely different situations. Which one you see is decided for you by the type of fix β you do not choose.
How to use it
Some fixes leave a trace we can go and verify: schema markup, rewritten blog titles, a completed G2 profile, a new citation. Those cards show "Measure Impact" β we re-check the site, confirm the fix is actually live, and if it is, re-run the audit and report what moved. Other fixes leave no verifiable trace β a guest-post pitch, a heading rewrite, a Perplexity action plan. Those cards show "Mark shipped" instead, because there is nothing for us to detect; you are simply telling the system it is done.
- "Measure Impact" β deploy the fix first, then click it. The button reads "Measuringβ¦" while it works, which can take a few minutes.
- If it comes back saying the fix is not detected live yet, that is the honest answer: it has not propagated, or it did not go out. Deploy, then measure again.
- If measuring is blocked, the message tells you why β usually a cooldown since the last measurement, or a monthly cap.
- "Mark shipped" β click it when you have done the work. The status badge flips to "Confirmed Shipped" and the card then offers "Measure Impact".
- Once something has been measured successfully, the result appears as a Proof of Impact card on the Insights tab.
Measured results are described as correlational, and deliberately so. AI engines refresh on very different schedules β Perplexity and Google reflect changes within days, ChatGPT and Claude can take weeks. Movement after a fix is evidence, not proof that this one change caused it. Present it that way to clients and you will never be caught out.
10. Turning on continuous monitoring
Client β Monitoring β Setup
What it is
An audit is a snapshot. Monitoring is the ongoing tracker: a set of questions asked roughly three times a week on average (daily for a direct brand) across the engines on your plan, building a trend line for visibility, share of voice, position and sentiment β for your client and for their competitors.
How to use it
When a client has no monitoring data yet, a three-step wizard appears in place of the monitoring dashboard. Step 1 confirms the brand and which competitors to track, and lets you add more. Step 2 offers AI-suggested questions β all pre-ticked β and a box to add your own, one per line. Step 3 shows the daily credit cost and offers to run the first pass immediately.
- Step 1 of 3 β Brands: confirm the brand name and website, tick the competitors to track, add any that are missing, click "Next".
- Step 2 of 3 β Prompts: untick any suggestion you do not want, type your own in the box below, click "Next".
- Step 3 of 3 β Launch: read the credit line, then either "Run first monitoring pass now" or "Skip β daily runs will pick it up".
Two things surprise people. First, there is no schedule picker and no engine picker β the monitoring pass runs daily while your credits last, and the engines come from your plan. Second, there is no "Finish" button; choosing run or skip on Step 3 is the end. On credits, the wizard does the maths for you: one credit is one prompt, on one engine, for one day β so ten prompts across five engines is fifty credits a day. Your remaining balance is shown under Agency β Settings β Monitoring.
11. Managing the questions you track
Client β Monitoring β Prompts
What it is
The list of questions being asked on this client's behalf every day, with how each one is performing. This is the tab you tune over time as you learn what actually matters for the account.
How to use it
The table shows each prompt with its topic, tags, country, status and results β responses, visibility, citations, sentiment and when it last ran. You can sort by Visibility or Citations, and filter by topic, tag, engine or status. Click any row to open a drill-down showing the raw answers each engine gave, day by day, including who got mentioned and which sources were cited. That drill-down is the single most persuasive thing to show a skeptical client β it is the actual AI output, not a summary.
- Add one question at a time with "+ Add prompt" (prompt text, plus optional topic, tags and country).
- Add many at once with the CSV importer at the bottom β download the template first; the header must be exactly prompt_text,country,topic,tags and tags are pipe-separated.
- Use "Suggested prompts" at the bottom and click "Generate" to get AI-proposed questions from the client profile, then Accept or Dismiss each one.
- Pause a prompt to stop spending credits on it without losing its history; Archive to retire it; Unarchive brings it back paused, so it does not silently consume your cap.
- "Run now" queues an extra monitoring pass outside the daily schedule.
Your plan caps how many prompts can be active, and the portal does not show a counter β you find out by hitting it, either as an error when adding or as a "skipped" row in a CSV import. Pause or archive prompts you no longer need before adding a batch of new ones. Note that the default view shows active and paused prompts; archived ones only appear when you filter for them.
12. Sources and competitors
Client β Monitoring β Sources, and Client β Monitoring β Competitors
What it is
Sources answers "where are the AI engines getting their information from?" Competitors answers "how does my client stack up against the other brands in the same answers?"
How to use it
On Sources you get the domains and URLs being cited in answers about this market, how often, a 14-day trend, and whether each one is your client's own site, a competitor's, or a third party. Below the table, "Gap analysis" lists sources that appear in these AI answers but do not cite your client β that list is a ready-made outreach target list. On Competitors you get a benchmark table comparing every tracked brand on visibility, share of voice, average position and sentiment, with your client's row highlighted and badged "you." Below that, "Market brief" holds the research on this client's market: "Research the market" asks Perplexity and Gemini to search the web (once a day per client) and writes who the client is, up to three tiers of competitors with why each competes, where the client wins and where it faces pressure, and the sources, and Track on a competitor proposes it for monitoring. "Manage brands" is where you add or edit the brands being tracked, and "Suggested competitors" proposes brands to track. "Find competitors" lists the brands the AI already mentions in this client's answers, and each suggestion says which list it came from. Accept or Dismiss each one.
Watch this one: there are two separate competitor lists and they do not sync. The competitors in the client's Settings and Intake drive the audit. The tracked brands on the Monitoring β Competitors tab drive monitoring. Adding a rival in one place does not add it in the other β if you want it in both, enter it in both. Also note that on Average position, lower is better, which the screen does not spell out.
13. Alerts β getting told without logging in
Client β Monitoring β Alerts
What it is
Three rules that email you when something meaningful happens to a monitored client, so you are not relying on remembering to check.
How to use it
Each rule is its own card with an on/off switch, a threshold where relevant, and its own list of email recipients. Set them per client. The badge on each card says whether it is running on the platform default or a value you customized.
- "Visibility drop" β fires when the client's own visibility falls by at least the number of points you set.
- "Share-of-voice overtake" β fires the moment a competitor's share of voice passes your client's. No threshold to set; it is on or off.
- "Citation lost" β fires when a source that used to cite them goes quiet for at least the number of days you set.
- Add recipients by typing an email and clicking "Add"; remove one with the Γ on its chip. Click "Save" on that card.
Leave a threshold blank to use the built-in default β the field will read "Default". Alert history below the rules is read-only: there is no acknowledge, snooze or dismiss, so treat it as a log rather than an inbox.
14. Giving your client their own login
Client β Portal (requires Manage Clients)
What it is
A white-labeled portal where your client signs in and sees their own scores and progress under your branding. It is how you stop being the human API between the platform and the customer.
How to use it
Enter the client contact's email address and click "Invite." They are sent a sign-in email immediately β there is no password to create or share, they click the link in the email. You can invite up to ten people per client. The table below shows everyone invited, their status and when they last signed in.
- Client β Portal.
- Type the contact's email and click "Invite". A sign-in email goes out straight away.
- Use "Copy" next to the portal link if you would rather send the link yourself, e.g. in your own onboarding email.
- On an active viewer's row: "Preview" opens their view so you can check what they see before a call; "Resend link" sends a fresh sign-in email; "Revoke" removes access and ends any open session.
Preview starts a short-lived session as that viewer and opens the portal in a new tab β worth doing once before you tell a client their portal is ready, so you see exactly what they will see. Your branding for that portal comes from Agency β Settings.
15. The free audit tool β setting up your lead magnet
Agency β Free-audit setup
What it is
A free, white-labeled audit you put on your own website or share as a link. A prospect enters their website and email, gets an instant AI-visibility score, and lands in your Pipeline as a lead. It is a lead-generation tool that happens to demo your product.
How to use it
The page starts with a master switch β your audit is either live or off. Turn it on, then work down: "Share your audit" gives you a direct link and an embed snippet for your site. "Customize" sets the headline, subheadline, and your call-to-action button and link. "Form fields & copy" controls what the prospect fills in β website and work email are always asked, name and brand name are optional, and you can add up to ten questions of your own. "Report template" lets you restyle the report they receive using tokens for live data. HubSpot connection lives here too, if you want leads pushed straight into your CRM as contacts.
- Switch the audit on at the top of the page.
- Copy the direct link, or the embed code for your website.
- Set your headline, subheadline and CTA button under "Customize", then "Save changes".
- Adjust the form and its wording under "Form fields & copy", then "Save form".
- Optionally connect HubSpot β the on-screen instructions walk through generating an access token.
Setup lives under Agency; the leads it produces live under Pipeline. That split is deliberate β configuring the tool and working the leads are different jobs done at different times β but it does catch people out. If you are looking for prospects, you want Pipeline.
16. Working your free-audit leads
Pipeline β Leads
What it is
The inbox for everyone who has run your free audit. Each lead arrives with a real score already attached, which means you are starting a sales conversation holding evidence.
How to use it
Filter by status along the top β All, New, Contacted, Converted, Dismissed. The table shows brand, contact, score, status and when they submitted, with a link to the report they were sent. Click a row to open the detail panel on the right, which breaks their score down by dimension and shows their answers to any custom questions you added.
- Click a lead to open the detail panel.
- "View report" opens the same report the prospect received β read it before you call.
- "Mark contacted" once you have reached out, so the list reflects reality.
- "Resend email" if they say it never arrived.
- "Convert to client" turns the lead into a real client record β you fill in industry, location and competitors, and it drops you straight into their new workspace.
- "Dismiss" for leads that are not a fit.
Converting creates the client but does not create an intake or run a full audit β the free audit is a lighter snapshot. Treat conversion as the start of the normal flow: fill in Intake, then Trigger Audit.
17. Team and roles
Agency β Team, and Agency β Roles
What it is
Who from your agency can sign in, and what each of them is allowed to do. Team is the people; Roles is the permission sets you assign to them.
How to use it
On Roles you will find built-in system roles, which are read-only, plus any custom roles you create. A role is a name, a choice of whether it covers all clients or only specific ones, and a set of ticked permissions β Manage Users, Manage Roles, View Clients, Manage Clients, Run Audits, View Audits, View Reports, Manage Free Audits, Manage Settings, Manage Billing. On Team you invite people and assign them one of those roles.
- Agency β Roles β "+ New role" to build a permission set, e.g. an Account Manager who can run audits but not change agency settings.
- Tick "All clients", or leave it unticked to make the role client-scoped.
- Agency β Team β "+ Invite teammate": enter their email, set an initial password, choose the role.
- If the role is client-scoped, tick which clients they should see.
- Use "Edit" on a team row to change someone's role or client scope later, or "Deactivate" to block sign-in without deleting them.
You can only grant permissions you hold yourself β anything above your own level appears ticked-out and unselectable, with no explanation on screen. That is the reason, not a bug. Also note the invite sets an initial password that you share with the teammate; they sign in with it directly.
18. Your branding and agency settings
Agency β Settings
What it is
The white-label controls. What you set here is what your clients see in their portal and on the PDF reports you hand over β your name, your logo, your colors.
How to use it
Set your agency name, logo, primary and secondary colors, and the display name used on outbound emails. Your subdomain is shown but locked. If you are on a plan that supports a custom domain, enter it here and the verification is handled for you after you set up the CNAME. Below the branding card, the Monitoring card shows how many monitoring credits you have used against your allowance for the current period.
The logo field takes a URL, not a file β host the image somewhere first (SVG or PNG, under 200KB) and paste the link. Get branding right before you invite your first client to their portal, since that is the first thing they see.
A good default rhythm for a new client: add them, fill in Intake properly, trigger an audit, walk the report with them using the report guide, work the top recommendations, then turn on monitoring so the next conversation is about a trend rather than a snapshot.