A new pattern showed up in our SME client sales call data through 2025. Booked calls were holding steady. Close rate was dropping. Sales teams were reporting the same complaint across categories. The buyers showing up to calls felt different.
Less informed. Less prepared. Sometimes outright confused about what they had booked. Some of them turned out not to be the actual decision maker. A few admitted they hadn’t booked the call themselves.
The pattern had a name once we tracked it. AI agents were booking sales calls on behalf of buyers. The leads being sent through agent-bookings were converting at roughly half the rate of buyer-booked calls.
The trend is small now. It’s growing fast. By 2027 a meaningful share of every B2B and high-consideration B2C call your sales team takes will be agent-booked. This guide explains what’s happening, how to spot agent-booked calls, and how to qualify them differently so your close rate stays healthy.
What’s Actually Happening
ChatGPT, Claude, Gemini and the new wave of AI shopping agents (Perplexity Comet, OpenAI’s Operator, Anthropic’s Claude with tool use) can now browse the web, fill in forms and make bookings on behalf of users. This capability went from research demo in 2024 to production-ready consumer feature through 2025 and 2026.
The way it shows up in your business is straightforward. A buyer asks ChatGPT “find me a good Google Ads agency that works with B2B SaaS” or “book me a demo with the top 3 marketing automation platforms”. The agent searches, picks candidates, navigates to their booking pages and fills in the call request form on the user’s behalf.
The buyer gets back a list of booked calls. Your sales team receives the lead. Your CRM logs it as a new opportunity. From your side it looks like a normal inbound. From the buyer’s side they may not even remember which calls they booked, who they spoke to, or why they’re on the call.
Why Agent-Booked Calls Convert At Half The Rate
Five reasons explain the conversion gap between buyer-booked and agent-booked calls.
Reason 1: The Buyer Hasn’t Self-Selected
A buyer who booked a call themselves has actively chosen your business over the alternatives they saw. They’ve done some research. They’ve decided your offering is worth their time.
An agent-booked call lacks that self-selection. The agent picked your business from a list. The buyer didn’t necessarily evaluate you against alternatives. The intent signal is weaker.
Reason 2: The Buyer Doesn’t Know What They Booked
Agent-booked calls often arrive with the buyer slightly confused about the context. They forgot which calls were booked. They don’t remember reading your website. They might think you do something different from what you actually do.
Sales reps used to walking into a call where the buyer has visited the website and read the case studies have to start the conversation from scratch.
Reason 3: The Booker Might Not Be The Decision Maker
The user prompting the AI agent might be a researcher, an assistant, an exec’s chief of staff, or the exec themselves. Agent-booked calls increasingly arrive with someone who has authority to research but not authority to buy.
This was always true for some inbound leads. It’s becoming the norm for agent-booked ones.
Reason 4: The Pre-Call Research Is Shallow
Agent-led research is fast but shallow. The agent finds your business, reads the homepage, summarises the offering, books the call. It doesn’t read your blog. It doesn’t watch your founder video. It doesn’t dig into the case studies.
The buyer arrives with the agent’s summary, not with the depth a self-researched buyer brings.
Reason 5: The Buying Process Is Earlier
Agent-booked calls tend to arrive earlier in the buyer’s process. The agent is doing initial scoping. The buyer is exploring options. Both of these are healthy signals, but they don’t convert at the same rate as buyers further down the funnel.
How To Spot An Agent-Booked Call
Five signals together usually indicate an agent-booked call. None are conclusive alone. Together they’re a reliable read.
Signal 1: The Booking Form Was Filled In With Suspiciously Tidy Data
Real buyers fudge their email address, use a casual title in the company field, or skip optional fields. AI agents fill in every field with structured, well-formatted data. Names match LinkedIn. Email domain matches the company exactly. Job title is the formal version.
If every field looks like it was copy-pasted from a structured data source, the lead was probably agent-booked.
Signal 2: The Booking Confirmation Email Was Not Forwarded To Anyone
A real buyer often forwards the calendar invite to a colleague or their EA for visibility. An agent-booked call goes to the user’s inbox and stays there. You can sometimes see this via call invite acceptance patterns or via the buyer’s behaviour in the call confirmation chain.
Signal 3: The Buyer Cannot Recall Specific Pre-Call Research
In the first 60 seconds of the call, ask “what did you read on our site before booking?” or “what made you pick us versus the alternatives you looked at?”. A self-researched buyer answers immediately and specifically. An agent-booked buyer hesitates or gives a vague answer that maps to the homepage summary an AI might have generated.
Signal 4: The Buyer Booked Multiple Similar Calls In The Same Window
Look at the booking time stamp. If the lead came in as part of a cluster of 3 or 4 similar bookings across competitor sites within the same 5-minute window, it’s almost certainly an agent run.
Signal 5: The User Agent Or Referrer Headers Show Agent Traffic
For technical teams, the booking form submission’s HTTP headers often reveal the agent. Look for user agent strings containing “OpenAI”, “Anthropic”, “Perplexity” or “Operator”. Look for referrer URLs from agent platforms. These patterns are still being established but the signal is becoming visible.
How To Qualify Agent-Booked Calls Differently
The fix isn’t to refuse the calls. Agent-booked leads can still close. They just need a different qualification approach.
Approach 1: Add A Mandatory “Pre-Call Brief” Field To Your Booking Form
Require the booker to write a 1-2 sentence description of what they want to discuss. Agents will fill this in, but the answer will be generic and tied to the homepage summary. Real buyers write specific, business-context answers. The difference is visible at a glance.
Approach 2: Send A Pre-Call Confirmation Email That Requires A Real Response
Send the calendar invite plus a short confirmation email asking the booker to reply with 2-3 of their specific challenges they want to discuss. Agents often don’t process the reply. Real buyers do.
If 24 hours before the call you haven’t received a reply, flag the booking as likely agent and adjust the qualification approach.
Approach 3: Open The Call With A Specific Context Question
Don’t open with “tell me about your business”. Open with “what specifically did you want help with that brought you to this call today?”. Agent-booked buyers struggle with this question. Real buyers handle it easily.
Approach 4: Have A Different First-Call Agenda For Agent-Booked Leads
If you’ve flagged the lead as agent-booked, run a shorter, more discovery-led first call. The goal isn’t to close. The goal is to determine whether there’s a real buying process underneath the agent booking.
If yes, schedule a proper sales conversation for the next step. If no, exit gracefully without burning a 30-minute slot.
Approach 5: Track Agent-Booked Versus Buyer-Booked Conversion Rates Separately
Add a field to your CRM that flags agent-booked calls. Track close rate by booking type. Once you have 30+ of each, you’ll have hard data on the conversion gap for your specific business. Some categories will be worse than others.
How To Adapt Your Marketing For The Agent-Booking Era
Adaptation 1: Make Your Homepage Agent-Readable
AI agents scrape the homepage to produce a quick summary of your business for the buyer they’re booking on behalf of. If your homepage is vague, the agent’s summary will be vague. If it’s specific, the summary will be specific.
The same five-lever AEO/GEO work that makes content cite-able for ChatGPT makes it usable for agents booking on behalf of buyers.
Adaptation 2: Add Structured Booking Pages With Clear Service Definitions
Agents prefer booking pages that are explicit about what the call is for, how long it takes and who it’s for. Generic “book a discovery call” pages produce generic agent bookings. Specific “30-min Google Ads audit for £2k-£15k/mo accounts” pages produce more qualified agent bookings.
Adaptation 3: Build A Different Sales Cadence For Agent-Booked Leads
Your existing post-booking nurture probably assumes the buyer self-booked. Update the sequence to ask early-funnel questions and confirm context for agent-booked leads. The lead quality improves measurably when the cadence matches the actual buying process.
5 Mistakes Most SMEs Are Making With Agent-Booked Calls
Mistake 1: Treating Agent-Booked Calls As Equivalent To Buyer-Booked
The two have different conversion economics. Mixing them in the same pipeline reporting masks the underlying performance gap. Track them separately.
Mistake 2: Optimising Marketing Against Mixed Conversion Data
If your ad spend is being optimised against a conversion signal that mixes buyer-booked and agent-booked at different conversion rates, your spend optimisation is being skewed. Update your offline conversion imports to distinguish the two where possible.
Mistake 3: Not Asking The Pre-Call Context Question
The fastest way to identify an agent-booked call is to ask “what brought you here today specifically?” in the first 60 seconds. Most teams skip this and discover the agent booking 20 minutes into a wasted call.
Mistake 4: Trying To Refuse Agent-Booked Calls Outright
Some teams have started blocking agent bookings entirely. This is short-sighted. Agent-booked calls still close at meaningful rates, just lower than buyer-booked. Refusing them blocks a growing share of legitimate pipeline.
Mistake 5: Not Tracking The Trend Over Time
Agent-booked call share is growing roughly 30% to 50% quarter on quarter through 2025 and 2026. If you’re not tracking it as a separate metric, you’ll miss the inflection point where it becomes a majority of your inbound calls.
Your 7-Day Agent-Booking Readiness Plan
- Day 1: Add a “How did you book this call?” field to your booking form with a free-text option. Most agents won’t fill it in or will fill in generic text.
- Day 2: Add a CRM field flagging booking type. Train your sales team on the five signals of agent bookings.
- Day 3: Audit your last 30 booked calls for the five signals. Estimate the current agent-booking share.
- Day 4: Update your booking confirmation email to include a context-confirmation question and require a reply.
- Day 5: Train your sales team on the first 60-second context question and the different first-call agenda for flagged leads.
- Day 6: Add a tracked metric to your monthly sales report. Agent-booked vs buyer-booked close rate.
- Day 7: Update your homepage and core service pages with AEO-grade direct answers, specific service definitions and clear booking page descriptions. Agent-readability becomes a marketing priority.
That’s a complete agent-booking adaptation plan. In one week. Most SME sales teams haven’t even noticed the trend yet.
Frequently asked questions
What does it mean when an AI agent books a sales call?
AI shopping agents like ChatGPT with web browsing, Perplexity Comet, OpenAI’s Operator and Anthropic’s Claude can now navigate websites, fill in forms and book calendar slots on behalf of users. The user prompts the agent (“book me demos with the top 3 marketing automation platforms”) and the agent completes the bookings autonomously.
How can I tell if a sales call was booked by an AI agent?
Five signals together usually indicate an agent booking. Tidy structured form data, no confirmation forwarding to colleagues, vague answers to “what did you read before booking”, clustered booking times across competitor sites, and agent-related user agent or referrer headers on the form submission.
Why do AI agent-booked calls convert at a lower rate?
Five reasons. The buyer hasn’t self-selected your business. The buyer often doesn’t fully remember what they booked. The user prompting the agent may not be the decision maker. The pre-call research is shallow. The buyer is typically earlier in the buying process than self-booked leads.
Should I block AI agent bookings on my calendar?
No. Agent-booked calls still close at meaningful rates, just lower than buyer-booked. Blocking them prevents a growing share of legitimate pipeline. The better approach is to qualify them differently with a pre-call brief field, a context-confirmation email and a different first-call agenda.
How do I qualify an agent-booked call?
Open with a specific context question in the first 60 seconds (“what brought you here today?”). Require a 1-2 sentence pre-call brief on the booking form. Send a confirmation email that requires a real reply. Run a shorter discovery-led first call to determine if there’s a real buying process underneath.
How is this changing my Google Ads conversion data?
If a meaningful share of your inbound leads are now agent-booked, your form-submission conversion data is being skewed by the lower close rate on those leads. Update your offline conversion imports to flag booking type. Smart Bidding then optimises against the cleaner signal of buyer-booked calls or weighted conversion values.
How fast is the agent-booking trend growing?
Across SME B2B and high-consideration B2C accounts SF Digital monitors, agent-booked share of total inbound calls is growing roughly 30% to 50% quarter on quarter through 2025 and 2026. By 2027 it will likely be a majority of inbound calls in many B2B categories.
The Buyer You’re Talking To May Not Have Booked The Call
The agent-booking trend is small now and growing fast. Within 18 to 24 months a meaningful share of every B2B sales call will be booked by an agent acting on behalf of a buyer who may not even remember the booking.
The sales teams that prepare for this win the next two years. The ones that don’t will watch their close rate quietly decline while assuming the leads have just got worse.
The fix is structured. Spot the agent bookings. Qualify them differently. Track them separately. Adapt your marketing to be agent-readable. Update your homepage to be agent-summarisable.
Start the audit this week. By month three you’ll have hard data on what share of your bookings are agent-led and how their conversion rate compares. By month six you’ll have a sales process that handles both equally well and a marketing engine that pulls qualified leads from both buyer-led and agent-led paths.
The buyer you’re talking to may not have booked the call. The agent that booked it is sitting in their inbox. Get ready to sell to both.
Written by Uzair Kharawala
Founder of SF Digital Studios, a Google Premier Partner agency. Uzair has been helping businesses grow with Google Ads since 2002 and teaches it every week on YouTube.
Get our best Google Ads tips by email
New articles and videos, straight to your inbox. Unsubscribe any time.