.svg)
Overview
Discovery calls are the highest-stakes conversations in B2B sales. In 30 minutes, a rep either uncovers the pain, timeline, budget, and decision process that will fuel the entire deal -- or they waste everyone's time with surface-level questions and leave without the information needed to advance the opportunity. The difference between a great discovery call and a mediocre one is rarely the rep's conversational skill. It is the preparation, frameworks, and follow-up infrastructure behind the call.
For GTM Engineers, discovery is not a soft skill to leave to sales training. It is a systems problem. The quality of a discovery call is directly determined by the pre-call research available to the rep, the frameworks embedded in their workflow, the intelligence captured during the conversation, and the speed and quality of follow-up after it ends. Each of these is an infrastructure layer you can build, measure, and optimize.
This guide covers discovery frameworks that work in practice, how to build question libraries that reps actually use, pre-call research automation, call intelligence infrastructure, and the follow-up workflows that convert good conversations into pipeline progression.
Discovery Frameworks That Work
Frameworks are not scripts. They are structures that ensure reps cover the essential ground without sounding robotic. The GTM Engineer's role is to embed these frameworks into tooling so reps follow them naturally, not because they memorized a slide deck during onboarding.
MEDDPICC for Enterprise
MEDDPICC (Metrics, Economic Buyer, Decision Criteria, Decision Process, Paper Process, Identify Pain, Champion, Competition) remains the gold standard for enterprise discovery. The challenge is not teaching the framework -- it is operationalizing it. Build CRM fields for each MEDDPICC element and create stage gates that require reps to fill them before advancing a deal. A deal cannot move to Technical Evaluation until Pain and Champion fields are populated with substantive data, not placeholder text.
SPICED for Mid-Market
SPICED (Situation, Pain, Impact, Critical Event, Decision) is a lighter framework that works well for mid-market deals where the buying process is less formal. It focuses on what matters most: understanding the prospect's current situation, the specific pain driving urgency, the business impact of that pain, any critical event or deadline creating a time constraint, and how the decision will get made.
The Three-Layer Discovery Model
Regardless of which named framework you adopt, effective discovery operates on three layers:
| Layer | Questions to Answer | What It Tells You |
|---|---|---|
| Situational | What do they use today? How big is the team? What does their process look like? | Whether they fit your ICP and which use cases apply |
| Problem | What is not working? What is the cost of the status quo? What have they tried? | Whether there is a real pain worth solving and how urgent it is |
| Decision | Who else is involved? What is the timeline? What is the budget? How have they bought similar solutions? | Whether the deal is winnable and what the path to close looks like |
Your infrastructure should guide reps through these layers. Build call guides that surface the right questions based on where the conversation sits -- not a static list of 30 questions, but a dynamic guide that adapts based on what has already been answered.
The biggest mistake GTM Engineers make with discovery frameworks is over-engineering them. Reps who feel like they are filling out a form during a conversation come across as interrogators, not advisors. Build the framework into the post-call workflow instead: after the call, prompt the rep to tag which elements were covered, and flag gaps that need follow-up. This captures the structure without disrupting the conversation.
Building Question Libraries That Reps Actually Use
Most question libraries are Google Docs that get shared during onboarding and never opened again. To make question libraries useful, they need to be contextual, searchable, and embedded in the workflow.
Organizing by Persona and Pain
Structure your question library around buyer personas and pain categories, not around your product features. An SDR calling a VP of Sales should see different questions than one calling a Director of IT, even if both are prospects for the same product. Map questions to the persona models you have already built.
Layered Question Sets
Build three tiers of questions for each persona-pain combination:
- Opener questions: Broad, non-threatening questions that get the prospect talking. "Walk me through how your team handles [process] today."
- Drill-down questions: Follow-ups that dig into the implications of what the prospect reveals. "When that breaks down, what does that cost you in terms of [metric]?"
- Commitment questions: Questions that advance the deal by establishing timeline, budget, and next steps. "If we could solve [pain], what would need to happen internally to move forward?"
Embedding in Tooling
The question library should surface inside the tools reps already use. This means integrating it with your real-time coaching platform so relevant questions pop up during live calls, or embedding it in the CRM so that when a rep opens a contact record tagged with a specific persona, the corresponding questions are one click away.
Pre-Call Research Automation
The quality of a discovery call is determined before the call starts. A rep who knows the prospect's tech stack, recent company news, competitive landscape, and likely pain points will run a fundamentally different conversation than one who opens with "so, tell me about your company."
The Research Brief
Automate the generation of a one-page research brief for every booked meeting. This brief should include:
Timing and Delivery
The brief should be delivered 30-60 minutes before the scheduled call. Delivering it at booking time means the rep forgets the details by call time. Delivering it 5 minutes before means no time to review. A calendar-triggered workflow that pushes the brief via Slack or email 45 minutes before the meeting hits the sweet spot.
Track conversion rates from discovery to next stage for calls where the rep used the research brief versus calls where they did not. Most teams find a 15-25% lift in stage progression when reps are well-prepared. That data makes the business case for investing in research automation infrastructure.
Call Intelligence Infrastructure
What happens during the call is only valuable if it is captured, structured, and made actionable. Call intelligence platforms have matured significantly, but the GTM Engineer's job is to build the infrastructure that connects call data to downstream workflows.
Recording and Transcription
Every discovery call should be recorded and transcribed. This is table stakes. The more valuable layer is automated analysis: identifying key topics discussed, questions asked, objections raised, competitor mentions, and next steps committed to. Your call intelligence platform should tag these automatically and push them into the CRM deal record.
Framework Compliance Scoring
Use call intelligence to measure discovery quality at scale. Did the rep cover the key MEDDPICC or SPICED elements? Did they ask about pain, budget, timeline, and decision process? Did they spend more time listening or talking? The ideal talk-to-listen ratio for discovery calls is roughly 40/60 -- the rep talks 40% and listens 60%. Tracking this across the team gives managers coaching data they cannot get from ride-alongs alone.
Insight Extraction for Deal Progression
The real value of call intelligence is not coaching -- it is deal context. When a prospect mentions a competitor, that should automatically trigger competitive intelligence workflows. When they mention a specific pain, that should update the CRM pain field and trigger persona-specific follow-up content. When they give a timeline, that should update the expected close date. Build the connective tissue between what is said on calls and what happens in your GTM systems.
Post-Discovery Follow-Up Automation
The follow-up after a discovery call is where most deals either gain momentum or begin to stall. Speed and relevance are everything. A personalized follow-up email sent within 2 hours of the call outperforms a generic one sent the next day by a wide margin.
Automated Follow-Up Drafts
Build a workflow that generates a follow-up email draft based on call notes and CRM data. The draft should:
- Reference specific pain points discussed during the call
- Attach relevant content (case studies, one-pagers, ROI calculators) based on the persona and pain identified
- Propose clear next steps with specific dates, not "let's reconnect soon"
- Include any other stakeholders mentioned during the call, positioning the multi-thread early
Internal Handoff Documentation
If the deal is progressing to a demo or technical evaluation, the discovery output needs to flow to whoever is handling the next stage. Build an automated handoff brief that includes the discovery findings, the prospect's stated requirements, competitive context, and any technical concerns raised. This prevents the next conversation from re-covering ground that was already explored.
Gap Follow-Up Sequences
Not every discovery call covers everything. Build adaptive sequences that follow up on specific gaps. If the rep covered pain but not timeline, trigger a follow-up that references the pain discussed and asks about decision timing. If budget was not addressed, queue a value-focused touchpoint that establishes ROI before the budget conversation. These gap-filling sequences should be automated based on which CRM fields remain empty after the call.
Common Discovery Mistakes
These are the patterns that call intelligence data consistently reveals across underperforming discovery calls.
- Pitching instead of discovering: Reps who spend more than 40% of the call talking are pitching, not discovering. Your call intelligence platform should flag this automatically.
- Asking questions you should already know: If your pre-call research brief includes the company's size, industry, and tech stack, the rep should not be asking about those during the call. It signals lack of preparation and wastes precious time.
- Not multi-threading early: Discovery is the best time to ask "who else would be involved in evaluating a solution like this?" Reps who wait until later in the deal to multi-thread discover stakeholders too late to influence.
- No clear next step: Every discovery call should end with a specific, time-bound commitment. "I will send you a proposal next week" is not a next step. "I will send the technical requirements doc by Thursday, and we have a 30-minute technical deep-dive scheduled for Monday at 2pm" is.
- Not capturing competitive context: If a prospect is evaluating alternatives, that information needs to be captured and trigger competitive workflows immediately, not discovered weeks later when the deal is at risk.
FAQ
Thirty minutes is the standard for initial discovery. This is enough time to cover situational context, primary pain, and decision process basics without exhausting the prospect's patience. For enterprise deals with complex buying processes, a 45-minute discovery may be necessary, but this should be the exception. If your reps consistently need more than 30 minutes, they are either asking questions they should already know the answers to, or they are trying to cover too much ground in a single conversation.
Yes, for mid-market and enterprise deals. Combining discovery and demo in a single call means you are either shortchanging discovery (rushing through questions to get to the demo) or delivering a generic demo (because you have not had time to customize based on what you learned). For SMB and high-velocity motions, a combined call can work if your product is simple enough to demo in 15 minutes and the buying process is straightforward.
Use call intelligence platforms to track framework compliance (were key elements covered?), talk-to-listen ratios, question depth (surface vs. drill-down), and next-step quality. Then correlate these metrics with deal outcomes. You will find that calls scoring high on discovery quality convert to next stage at 2-3x the rate of low-scoring calls. This data makes discovery quality a measurable, coachable metric rather than a subjective judgment.
The SDR's discovery is qualification-focused: Does this prospect match our ICP? Is there a real pain? Is there a timeline? The AE's discovery goes deeper: What is the business impact of the pain? Who are the stakeholders? What is the decision process? What is the competitive landscape? Your infrastructure should ensure that the SDR's findings are passed to the AE automatically so the AE builds on what is known rather than re-covering the same ground. Read more in our guide to personalizing the sales handoff.
What Changes at Scale
Running discovery for a 5-person sales team is manageable. You can listen to every call, review every follow-up email, and personally ensure that research briefs are comprehensive. At 50 reps running 10-15 discovery calls each per week, you are looking at 500-750 calls. Manual quality assurance is physically impossible.
The first thing that breaks is research consistency. Some reps prepare meticulously; others wing it. The pre-call research pipeline that worked when you had time to monitor it starts producing stale or incomplete briefs because nobody is maintaining the enrichment workflows. Follow-up quality degrades because there is no automated enforcement of timing or content standards.
Octave automates the discovery support infrastructure that breaks at scale. The Call Prep Agent generates comprehensive pre-call briefs by pulling enrichment data, CRM history, and account context from the Library, so every rep walks into a discovery call prepared regardless of whether they had time to research manually. The Enrich Company and Enrich Person Agents ensure the underlying data is current, while Playbooks trigger follow-up sequences through the Sequence Agent based on discovery outcomes -- keeping the post-call workflow as consistent as the pre-call preparation.
Conclusion
Discovery calls are won or lost before the phone rings. The GTM Engineer's job is to build the infrastructure that makes every rep well-prepared, every conversation well-structured, and every follow-up well-timed. Pre-call research automation, embedded question frameworks, call intelligence, and follow-up workflows are the four pillars.
Start with research automation -- it has the most immediate impact on call quality and requires the least behavioral change from reps. Then layer in call intelligence to measure quality and identify coaching opportunities. Finally, build the follow-up automation that converts good conversations into deal progression. Discovery is a system, and systems can be engineered.