AI SDR for B2B SaaS: Routing Demo Requests, Trials, and Pricing Page Leads

DialSwarm team · October 17, 2026 · 6 min read

The inbound problem in SaaS

In B2B SaaS, the hand-raisers are usually right there in your data. Someone books a demo, starts a trial, or fills out the contact form on the pricing page. The hard part is reaching them while they are still evaluating, and getting them to the right person.

Most teams split this between SDRs, who qualify, and AEs, who demo and close. In practice SDRs spend much of the day in a loop: dial, no answer, voicemail, email, log, repeat. When a launch or a campaign doubles inbound volume, the loop falls behind, and the warmest leads wait the longest.

DialSwarm takes the first pass. The agent calls consented leads, never claims to be human, says it's an AI when asked, asks your qualifying questions, and warm transfers ready buyers to an SDR or AE. The rest of this guide is about doing that well in a SaaS context.

Lead sources and what each one needs

Demo requests. The strongest signal. The person asked to talk. Call within moments, on a Fast schedule, and transfer quickly. Many of these should skip most qualifying and go straight to a human.

Pricing page contact forms. Often a buyer comparing options or with a specific question about seats, security, or contracts. Good fit for a short call that captures the question and routes it.

Free trial signups. Mixed intent. Some are serious evaluators, some are students and competitors. Call only those who opted in to calls, and use product signals to decide who gets called first (more below).

Content and webinar leads. Early stage. Many did not ask for a call. If they did opt in, use a Steady schedule and a soft script. If they did not, leave them to email nurture.

Outbound target accounts. People who never raised a hand. An AI agent should not cold dial these. Leave outbound prospecting to your human SDRs and their own judgment and compliance process.

Routing rules: who gets the transfer

Routing is where SaaS teams most often get it wrong. A warm transfer to the wrong rep is worse than no transfer at all.

Write routing rules before launch:

• By segment. Company size answers usually decide whether a lead goes to an SMB rep, a mid-market AE, or an enterprise team.
• By territory. If AEs own regions, route on location.
• By existing account. If the caller's company is already a customer, route to the account manager, not a new business rep.
• By question type. Security questionnaires, procurement, and technical integration questions go to a person who can answer them, not to the agent.

When nobody in the right group is available, the agent books time on their calendar or creates a follow-up task. It should never transfer an enterprise buyer to whoever happens to be free just to make the transfer happen.

Qualifying questions that respect the buyer's time

B2B buyers hate long discovery calls with someone who cannot help them. Keep the AI's part to what you need for routing.

For demo requests: - What prompted you to look at this now? - Roughly how many people on your team would use it? - Would you like to talk with someone now, or book a time?

For pricing page leads: - What question can we answer about pricing or plans? - How many seats are you thinking about? - Is there a timeline you are working toward?

For trial users: - How is the trial going so far? - Is there something you are trying to set up that we can help with? - Would a quick call with a product specialist be useful?

Avoid BANT-style interrogation on a first call. Budget and authority questions land better coming from an AE who has built some rapport.

Trial signups: product signals first

Not every trial signup deserves a call, and calling every one wastes connected calls and annoys people who just wanted to poke around.

Use what you know. If your product sends events to HubSpot or through webhooks, filter for signals that suggest serious evaluation: invited teammates, connected a data source, hit a usage limit, visited the pricing page during the trial. Route only those leads, who also opted in to calls, to the agent.

A Zapier or Make step can do this filtering. So can your own code, using DialSwarm's API and webhooks. The point is to call the people most likely to want help, at the moment they are likely to want it.

B2B calling is not a free pass

A common belief in SaaS is that calling business contacts is unregulated. That is not safe to assume.

• Many business contacts give a cell number. The FCC's February 2024 ruling confirmed that AI-generated voices are artificial voices under the TCPA, so calls using them need prior express consent, written for telemarketing. Source: FCC Declaratory Ruling FCC-24-17 (accessed Oct 17, 2026). The rules on artificial voice calls to wireless numbers do not turn off just because the person was at work when they filled out your form.
• The TSR's business exemption has limits. The FTC's guide describes how the Telemarketing Sales Rule treats business-to-business calls, including exceptions. See Complying with the Telemarketing Sales Rule (accessed Oct 17, 2026). The FTC has amended the rule over time, so check the current text with counsel rather than relying on an old summary.
• International leads. Calling contacts outside the US brings other countries' privacy and calling laws into play. Get advice before including them.
• Calling hours. Telephone solicitations are limited to 8 a.m. to 9 p.m. in the called person's local time under 47 CFR 64.1200 (accessed Oct 17, 2026). Your leads span time zones. Schedule accordingly.

The simple standard: put a clear consent checkbox on demo, trial, and pricing forms that mentions calls may use AI, store the record, and honor stop requests. This is general information, not legal advice.

Integrating with your stack

• HubSpot. DialSwarm syncs with HubSpot today, so call outcomes, transfers, and notes land on the contact record. Other CRMs are not supported directly yet.
• Zapier and Make. Use them to send new form fills or product events to DialSwarm and to pass outcomes to other tools, like a Slack alert when a transfer happens.
• API and webhooks. For teams with engineering time, these give full control over which leads are called and what happens with results.

Reps can listen in live on any call and coach the agent mid call by typing or talking. Recordings are available if you opt in, which helps with SDR coaching too.

What SDR managers should change

If you add an AI SDR and change nothing else, your SDR team will feel threatened and your metrics will look strange. A few adjustments help.

• Change the scorecard. Stop measuring dials. Measure meetings held, pipeline created, and speed from hand-raise to conversation.
• Move SDRs up the funnel. With the dial loop handled, SDRs can take transfers, run sharper discovery, and spend time on research-heavy outbound to target accounts, where human judgment matters.
• Let SDRs own the agent. They write the questions and tune the routing. They are training an AI version of themselves to do the tedious part, not their replacement.
• Review calls weekly. Pick five transferred calls and five that were not. Adjust.

DialSwarm bills connected calls only, so unanswered dials don't count. You get 10 free calls with no card, and current plans are on the pricing page. Unanswered dials do not count.

A two-week pilot plan

Week one: connect only demo requests. Have one SDR listen in on the first ten calls and note where buyers hesitate. Tighten the questions and routing rules. Week two: add pricing page leads with their own short script. At the end, compare time to first conversation, transfers accepted by AEs, and meetings held against the two weeks before launch. If AEs say transfers are poorly qualified, adjust before adding trial signups.

Questions