Skip to content

Inspect and curate the AI context

The AI generates its email drafts and recommendations from your context, meaning the data you maintain in Flino. To make sure the result fits you and your deal, it pays to look at what the AI currently “sees”. That is what the AI context badge on the offer detail page is for.

On the detail page of an offer, below the title, there is a badge labelled “AI context: Well prepared” (or “Please improve” or “Missing fundamentals”, depending on the state). Clicking it opens the Review AI context sheet with the subtitle:

“Flino currently uses these details for follow-ups and AI suggestions.”

The badge tells you at a glance whether the AI has enough material or whether you should add more.

BadgeMeaning
Well preparedThe AI has enough context for good suggestions. Tooltip: “The AI has enough context for good suggestions.” with a list of the sources available
Please improveEnough for usable suggestions, but gaps hold back the quality
Missing fundamentalsImportant basics are missing, suggestions become generic

For “Please improve” and “Missing fundamentals”, the tooltip names what is missing: “For better AI suggestions, the following is missing: …”, for example scope of work, call notes, a call or reaction, a customer reply or contact notes.

At the very top, the Add AI briefing area invites you to note briefly what Flino should know about this opportunity. Quick note (marked Recommended) opens the note dialog, Log a call the call dialog. You can type or dictate in both.

Below that, How the AI uses your context shows how Flino reads the state of the conversation: Goal, Interest, Tone, Last reply (Positive, Question, Objection, Neutral or Rejection), CTA and Objection.

Then come the sources the AI reads from, each with the most recent entries:

  • Scope of work: the description of the offer. You edit it on the offer page (“Edit via offer details”). If it is missing, the sheet says “No scope of work set. This info clearly improves AI quality.”
  • Important notes: notes you created with Use for follow-ups.
  • Call analyses: logged calls and analysed conversations. For analysed conversations you see the detected Objections, the question “Did I get this right?” and the Edit result button. Source: Log reactions and Analyze a conversation.
  • Reactions: what the customer said over other channels (WhatsApp, phone, …). This section appears as soon as there are reactions.
  • Contact notes: notes stored on the contact itself. You edit them in the contact details.
  • Last customer reply: the most recent reply from the email thread.

For each section you see whether the content is currently used (badge “Used by AI”) or empty (e.g. “No important notes available.”). Individual entries show what they influence, for example “influences: Timing · CTA” for a call.

Prioritisation: what the AI should weight or ignore

Section titled “Prioritisation: what the AI should weight or ignore”

For notes, calls, call analyses and reactions you set the priority per entry:

  • Important: the AI weights this entry more heavily. Toast: “Marked as important”
  • Standard: neutral weighting. Toast: “Set to standard”
  • Ignore: the AI no longer uses this entry. Toast: “Excluded from AI context”

This helps when, for example, an older call is outdated and only gets in the way: set it to Ignore. The last customer reply cannot be re-weighted.

The AI context is fed from several sources. The cleaner you maintain them, the better the suggestions:

  • Incoming emails are picked up through your Gmail or Outlook connection and flow into the context automatically.
  • Manually assigned emails from the Inbox are added as soon as you assign them to an offer.
  • Reactions over other channels (phone, WhatsApp, SMS, LinkedIn, in person) are recorded via the reaction dialogs. Call notes are taken over with their summary.
  • Notes on the offer or contact are a direct field for free text.
  • Scope of work: in the Scope of work card on the offer page you describe the offer or have it read from the offer PDF with Extract scope of work from PDF.
  • Capture all customer communication in Flino. Including the call in which your customer half-confirmed the deal. Log it via Log call, otherwise the AI cannot see it.
  • Set important notes to “Important”. Not all notes are equally relevant; flag explicitly what the AI should keep in mind.
  • Set outdated information to “Ignore”. When the situation has changed, an old hint will otherwise skew the recommendation.
  • Fill in the scope of work. It is the basis for every personalised email. Without it, the AI writes generic text.
  • Quickly review the context before important emails. A click on the badge takes seconds and often makes the difference between “fits” and “I have to rewrite this”.

For the AI draft in the Send offer dialog and in the AI suggestion tab of the reply composer, Flino shows what the draft is based on (“Based on: …”). If something is missing, a short hint appears, e.g. “Tip: A described scope of work clearly improves AI quality.” or “Tip: A call note improves follow-up quality.” With “Missing fundamentals”, a clear warning lists the missing details.