AI Pre-Call Briefing Tool: How I Prep Support Agents Before Every Call

AI Pre-Call Briefing Tool: How I Prep Support Agents Before Every Call

A pre-call briefing tool is an internal AI system that reads a customer’s context, order history, prior tickets, call notes and order notes, before a support call and generates a short brief for the agent. It replaces the manual scramble of digging through tabs at the start of every call, so the conversation opens informed instead of fumbling.

Before and after: on the left an agent with four scattered tabs for Shopify orders, helpdesk tickets, the call log and order notes, three to five minutes of searching on every call, opening with a flat how can I help you; on the right the same agent handed a brief card for Maria R., repeat customer, with the birthday gift note in green, the model's guess about a changed delivery address in amber, and an opener written as a question, read in five seconds before the dial tone

I build back-office AI tools for a DTC ecommerce brand on Shopify. This is a concept-level write-up of one of them: a pre-call briefing tool, sometimes called pre-call intelligence or agent assist. It is one of the highest leverage and least glamorous pieces of AI I have shipped, and it is the one I would build first for any store that still phones its customers.

The problem: agents waste minutes scavenging before every call

Before a support agent calls a customer, they spend a few minutes searching across systems to answer basic questions: what did this person order, did something go wrong, have they contacted us before. That prep time, multiplied across every call, is a real cost, and it pushes agents to improvise through the opening of the conversation.

The person making the call is good at talking to people. They should not be spending that time acting as a search engine. Across a day of calls the lost minutes add up, and call quality suffers because the agent is piecing context together live instead of listening.

The part that is harder to see on a timesheet is what happens when the search is cut short. An agent with four tabs open and a customer on the line will stop at the first order they find, which is not always the right one, and open the call with a question the customer already answered in a ticket last month. The customer hears a business that does not remember them.

💡
The expensive failure is not the three minutes. For a brand that sells on being personal, it is the customer hearing a business that does not remember them.

What goes into the brief

The brief pulls from four places: the order system for what was bought and where it went, the helpdesk for prior tickets and how they ended, the call log for past calls and their outcomes, and the order notes field, which is where customers write the human details. Each source is trimmed to the last twelve months and ordered by recency before the model sees it.

Shopify knows the orders. The helpdesk knows the complaints. The phone system knows the calls. None of them agree on who the customer is: the same person is an email address in one, a phone number in another, and a name with a typo in the third. The first version of the tool matched on a single identifier and silently missed a large share of the history.

💡
Match customers on more than one identifier, with a fallback for each. That was the single biggest improvement to the brief's accuracy, before the model was involved at all.

The order notes field earns its place. It is where a customer types “birthday gift for my mom” or “please leave with the neighbour” or “second time ordering, the last one was perfect”, and nobody on the team reads it unless something goes wrong. Feeding it to the brief turned a field that existed for the warehouse into the warmest thing on the agent’s screen.

What stays out: marketing events, email opens, abandoned carts, browsing. The model will happily summarise all of it and the brief becomes a report.

💡
The test for every field: would an agent say it out loud on the call? If not, it is noise.

What the brief looks like

The output is six lines. Four are facts read from a system: who the customer is, their last order, their prior contact, and any note they left. Two are the model’s inferences: the likely reason for the call and a suggested opener. The agent reads the whole thing in about five seconds.

What a pre-call brief looks like: four fact lines in green above a rule, customer, last order, prior contact and order note, then a model inference label and two amber lines below it, likely reason and suggested opener, with a footer explaining that everything above the line is read from a system and everything below it is a guess the agent can check in one question

Six lines was not the first design:

Fixed structure beats fluent prose for anything that has to be read in five seconds, a lesson I have since applied to every internal tool I have built.

The “likely reason” line is the one agents valued most and the one that caused the most trouble, which is covered below.

The model: cheap and fast, on purpose

The tool runs a language model over each customer’s context before the call and produces the brief. I used a fast, inexpensive model tier deliberately, because it runs on every single call, and a model that is good enough and costs almost nothing per run beats a smarter, pricier one you cannot afford to call constantly.

A brief costs well under a cent to generate. At that price nobody has to decide whether a call deserves one, which is the whole point: the tool only works if it runs on every call without anyone thinking about it. The same brief on a top tier model would cost ten to thirty times more, be slower, and read about the same, because summarising a few hundred words of structured data into six lines is not a hard problem for any current model.

The tier is Gemini Flash, the same one that scores the calls, and the two tools have moved together: 2.5 Flash when they were built, 3.5 Flash since Google retired 2.5 in October 2026.

The brief is text in and text out, so unlike call scoring it does not need Gemini’s audio input, and it would run on Claude Haiku 5.5 or GPT-5 nano for a similar or lower cost. I keep it on Gemini because one provider, one key and one billing line is simpler for a two tool system, and the swap test I ran for the call scoring tool covered this one at the same time.

💡
For high frequency internal tools, cost per run decides whether the tool ships, not raw model quality. The expensive models earn their price on the call scoring tool that reads a full transcript and judges tone. For a six line brief, the fast tier is the right tool.

The prompt is short. It does three things:

The second instruction is the one that did the work; see the section on where it broke.

How it is built: the plumbing, not the model

The language model was the easy part. The hard part was data engineering: pulling context from four systems, matching customers across them, deciding what belongs in a brief versus what is noise, and getting the output onto the agent’s screen before the call connects. The model was a small fraction of the work; integrating it into a real workflow was the rest.

How the brief is built: four sources, Shopify, the helpdesk, the call log and order notes, feed an assembler that pulls, dedupes, trims to the last twelve months and orders by recency, then a fast tier model call at under a cent per run, then the brief card in the agent's screen before the dial tone

The brief is generated ahead of the call, not when the agent asks for it, so it is already on screen by the time the call connects.

💡
The agent must never watch a spinner. A brief that is already there gets read; a brief that is loading gets skipped. This mattered more for adoption than anything in the prompt.

Where the build time went: 45 percent pulling and cleaning data from four systems, 20 percent deciding what belongs in a brief, 15 percent getting it into the agent's screen on time, 10 percent the facts versus inference layout, 10 percent the prompt and model

Roughly half the time went into getting the data out of the four systems and agreeing on who the customer was. A fifth went into deciding what belongs in a brief, which sounds like a product question and was mostly a data question: which events are signal, how far back to look, what to do with a customer who has forty orders. The prompt and the model took about a tenth.

💡
If you budget for the prompt and model only, you will ship something that reads beautifully and arrives late with the wrong customer on it. Getting a decent summary out of a model is close to trivial now. Getting the right inputs to it, and its output in front of the right person at the right moment, is where the engineering lives, across every internal tool I have built.

Why it is worth building: time, and then warmth

The obvious benefit is time saved per call. The larger benefit is warmth: order notes contain human details, like an order being a birthday gift, that no agent has time to surface live. When the brief surfaces them, the agent can open with personal context, which turns a transaction into a good customer experience.

A note might say an order is a birthday gift for the customer’s mom. The brief surfaces it, so the agent opens with “and happy birthday to your mom” instead of a flat “how can I help you.” That detail was in the data the whole time. The agent just could not dig it out fast enough on their own.

Here is how small that detail is, and how much it changes. A customer places her third order, a gift box shipped to a different address from the previous two, with “birthday gift for my mom” typed in the notes. The parcel is delivered. Two days later she calls.

Without the brief: the agent finds her by phone number, sees the latest order, and opens with “how can I help you today”. She explains that her mom did not get it. The agent asks for the order number, finds the address change, and spends the next two minutes establishing what the brief would have told them before the ring.

With the brief: the agent sees the gift note above the line and the address change flagged below it, and opens with “Hi Maria, and happy birthday to your mom, did the box reach her alright?” The customer’s first sentence is “no, actually”, the agent already has the delivery scan open, and the call is about fixing it from the first ten seconds.

Same agent, same data, same customer. The difference is whether a line she typed into a box three weeks earlier reached the person talking to her. Multiply that by the share of orders that carry a note, a gift message, a second address or a prior ticket, which for a gifting heavy store is most of them, and the warmth becomes the thing customers mention in reviews.

💡
The tool did not create anything. It moved what the business already knew to the moment it was useful. Agent assist is pitched as efficiency, and prep did drop from minutes to seconds, but the higher value outcome is a better conversation.

Where it broke, and how I fixed it

The tool would sometimes state a guess as confidently as a fact. A wrong assumption about why the customer was calling appeared in the same tidy tone as the verified order details, and an agent who trusted the brief could open the call mis-oriented. The fix was to separate facts from inferences in the brief, in the prompt and on the screen, so the agent knew which parts were solid and which to confirm.

In the first week, the “likely reason” line was right often enough that agents started treating it as fact. Then it was wrong on a call where it mattered, the agent led with the wrong assumption, and the customer had to correct them. One incident like that and the team’s trust in the whole brief dropped, including the four lines that were always right.

The fix was two changes:

Agents now read the top for what is true and the bottom for a head start. The suggested opener is phrased as a question for the same reason: “ask whether the gift arrived in time” cannot embarrass anyone, “the gift arrived late” can.

💡
A confident model is not a careful one. Keeping the human able to tell verified data from model inference is now the first thing I design into any tool where a model's output lands in front of a customer facing person.

Rolling it out and how to judge it

The tool went to two agents first, with a weekly five minute check on whether the briefs were right and useful. The measures that matter, in order: is the first line correct, how often is the inference confirmed on the call, prep time per call, how many calls open with a personal detail, and cost per brief.

How to judge a pre-call briefing tool: first line correct from a weekly spot check, inference confirmed on the call from a tick box after each call, prep time per call from queue timestamp to dial, calls opened with a detail from the call scoring tool, and cost per brief from model usage divided by calls

💡
If cost per brief goes above a cent, the tier is wrong or the context is too long. Trim the input rather than paying for a bigger model.

The recording, and why disputes got easier

Every call is recorded with the customer’s consent, announced at the start, and transcribed. The transcript is attached to the customer and feeds the brief for the next call. It also turned out to be the thing that settles chargebacks, “but you promised” calls and complaints about agents, because the record of what was said is no longer whoever remembers it loudest.

What the recording settles: chargebacks, where the transcript shows the agent confirming items, total and address with a timestamp and goes in as evidence; what was promised, where the next brief quotes the actual line from the last call; and agent fairness, where the call scoring tool has already rated tone and a manager reviews one recording instead of relying on who said what

The recording was built for the call scoring tool, not for disputes. The dispute use came on its own:

None of that works without consent, and the consent has to be real. The notice is read at the top of every call, in the customer’s language, and if the customer says no the recording stops and future briefs run on order data alone. Transcripts sit under the same access rules as the helpdesk, are not sent to any model provider that trains on inputs, and the brief carries only what a customer would reasonably expect the business to remember about their own orders.

💡
A customer who learns the business kept a recording they did not agree to has a bigger complaint than the one they called about. Consent first, every call, in their language.

What it does not do

The brief does not talk to the customer, does not write the ticket, and does not decide anything. It reads, summarises and hands over. That boundary was deliberate: the agent stays responsible for the call, the model’s guesses are labelled as guesses, and when the brief is wrong the cost is one awkward question rather than a wrong refund.

The tools that get into trouble are the ones where the model’s output goes to the customer without a person between them; the reasoning for keeping it internal is in the most useful AI I built was not customer facing.

It also does not replace the helpdesk search. For a complicated case the agent still opens the tabs. The brief handles the ninety percent of calls where the context fits in six lines, so that the agent’s attention goes to the ten percent that do not.

What this cost and what it would take to build yours

This is a concept-level case study of an internal tool I built. It describes the approach and lessons, not proprietary implementation details. The version described here is a small amount of code, a handful of integrations, one prompt and one small screen inside the agent’s existing tools.

💡
Start with a spreadsheet export and a prompt pasted into a chat window. That was the first prototype, and it was enough to show the team what a brief could look like and make them want one.

This tool is one half of the support work. The other half, scoring every call after it happens, feeds the brief with what the last call was actually about. Both sit in the wider set of AI workflows I run on a Shopify store, and the AI customer support post covers where a brief fits against chatbots and ticket triage.

If you want one built for your own support team, the automation page covers scope and cost, or book a call and bring a recent call that went badly.

Frequently asked questions

Can AI prepare an agent for an upcoming customer call with full context?
That is what this tool does. It gathers the customer's orders, order notes, prior tickets and past calls, has a fast model write a six line brief with facts above the line and inferences below, and puts it on the agent's screen before the call connects. Reading it takes about five seconds. It works for support calls, sales follow ups and any call where the person on the other end expects you to already know who they are.
What is a pre-call briefing tool?
An internal AI tool that reads a customer's context, order history, prior support tickets, call notes and order notes, before a support call and generates a short brief for the agent. It replaces manual pre-call research so the agent opens the call already knowing who they are talking to and why.
What is pre-call intelligence?
Another name for the same idea: using automation or a language model to assemble relevant customer context ahead of a call, so the agent does not have to search for it while the customer waits. Sales teams use the term for prospect research; in support it means order and ticket history.
Which AI model is best for a pre-call briefing tool?
For a tool that runs on every call, a fast, low cost model tier is usually the right choice over a slower, more capable one. The task is summarising a few hundred words of structured data into six lines, which is not hard for current models, so cost per run and latency matter more than raw capability. Mine runs for under a cent per brief.
What is the hardest part of building an agent assist AI tool?
The integration, not the model. Pulling clean context from several systems that disagree about who the customer is, filtering out noise, and getting the output onto the agent's screen before the call connects is ordinary data engineering, and it took most of the build time.
How do you keep an AI briefing tool from misleading agents?
Separate verified facts from model inferences in the output, visually and in the prompt. When agents can see which lines came from a system and which are the model's guess, they trust the solid parts and confirm the rest with one question on the call. Without that split, one confident wrong guess costs the tool its credibility.
Why record support calls, and what does it have to do with the brief?
The recording, taken with the customer's consent and announced at the start of the call, is transcribed and attached to the customer. It feeds the next brief with what the last call was actually about, and it settles disputes: a chargeback gets the transcript of the customer confirming the order, a promised refund gets the line where it was promised, and a complaint about an agent gets reviewed from the recording rather than from two accounts of it.
Does a pre-call brief need the customer's permission?
The brief only reads data the business already holds about its own customers and shows it to the staff who would have looked it up anyway. Keep it inside the same access rules as the helpdesk, do not send customer data to a model provider that trains on inputs, and keep a log of what was generated for each call.
Can a Shopify store build this without a developer?
A basic version, yes. Shopify's order data and most helpdesks are reachable from Zapier or Make, and a model call is one step in either. The version described here needed code because it pulls from four systems, dedupes across them and has to finish in a few seconds, but the first prototype was a spreadsheet and a prompt.