The First 90 Days with a GTM Intelligence Layer: A Practical Onboarding Playbook

Three phases. Each has a distinct job. Each builds on the one before.

The phases compound — Phase 2 depends on Phase 1 being properly complete, and Phase 3 depends on Phase 2 running for long enough to produce readable data. Teams who skip Phase 1 and start with Phase 2 produce weak baseline data. Teams who try to evaluate the platform too early are measuring a cold start against a benchmark that reflects a calibrated system. The two are not comparable.

Phase 1 is straightforward: create your Offering, set up your AI SDR, and launch. Wyra generates the Offering from your profile and content — you review and refine it. Once the AI SDR completes warmup (around Day 14–15) and your first Offering is Finalized, your first campaign is ready to go.

Your profile is the foundation’s foundation

Before you upload anything, audit your company profile. The intelligence layer analyzes your company continuously — your website, case studies, public content, and any documents you upload as Artifacts. That analysis is only as good as the source material it has access to.

If your website says “we help companies with cloud solutions” and your actual offering is AWS data platform modernization for mid-size healthcare organizations, the intelligence layer generates outreach angles anchored to the generic version of your positioning. Correct the profile first. Be specific: which ecosystem, which practice areas, which customer types, which outcomes you’ve delivered. Specificity in the profile produces specificity in the intelligence layer.

Upload Artifacts before building Offerings — this order matters

Artifacts are the documents Wyra extracts intelligence from: case studies, capability decks, testimonials, ROI studies, implementation guides, certifications. Uploading them before you build your Offerings is not a preference — it is the correct sequencing. The common mistake: teams build the Offering first, see low AI confidence scores, then upload Artifacts to improve them section by section. The right order is Artifacts first, then Offerings.

Build your first two Offerings — and actually Finalize them

An Offering is a structured pitch playbook. It is the intelligence layer’s brief for every campaign and every list enrichment run. It is not a deck, not a product description, and not a summary of your capabilities. It is a structured argument for why a specific type of buyer should engage with you.

Create your AI SDR on Day 1 — not Day 14

Most teams create their AI SDR in Week 3. They don’t realize they’ve burned two weeks of warmup time they’ll never get back.

The AI SDR requires a warmup period of approximately two weeks after creation. During warmup, the agent operates at reduced capacity to build a sending reputation — this protects deliverability long-term. The warmup runs automatically once the AI SDR is created. But it can only start when you create it. If you want your first campaign live by Day 35, create the AI SDR on Day 1.

Your Phase 1 campaigns are running. The job of Phase 2 is to engage with what they generate — the replies, the auto-replies, the first real conversations — and to extract the patterns that will improve everything that follows. If you’re launching additional campaigns against new Offerings or segments, this phase is where that happens too. The first intelligence cycle is calibration data. Read it as such.

Launch additional campaigns and upload new lists as you go

A list in Wyra is a CSV of LinkedIn profile URLs — a deliberate set of prospects you’ve identified because they match a specific targeting hypothesis: the right industry, the right role, with an observable trigger (a cloud migration announcement, a new partnership, a funding round, a leadership hire, a job posting for cloud architects).

For your first list, start narrower than feels comfortable. 200–500 contacts from a single vertical or a single role-type at companies with a specific trigger tells you something readable. A 2,000-contact broad-sweep list tells you noise. This first list size is distinct from total AI SDR outreach volume — as the campaign sequences run, the AI SDR touches each contact multiple times across channels. Total leads reached accumulates beyond the list size over the campaign’s runtime.

When you upload your list, the platform will prompt you to select a context Offering. This selection determines the quality of what enrichment produces. With a Finalized Offering selected, enrichment generates offering-contextual research for each lead: a relevance score, an outreach strategy, inferred pain points, a tailored CTA, discovery questions, and objections specific to what you sell — grounded in the Offering’s content. Select the Offering you built in Phase 1.

After enrichment completes, review 20–30 profiles in the Interactions view before launching. Check the relevance scores. Read the AI-generated outreach strategies. If the strategies look generic — if they could describe any company in the vertical without modification — your Offering’s Pain Points section may not be specific enough. Go back to the Offering, add specificity or an additional Artifact, and re-enrich a sample before launching.

Launch your first campaign — get these settings right the first time

The Campaign Wizard runs five steps: Offering selection, Persona, Idea, Messages, and AI SDR assignment.

Two campaign paths:if you’re using an uploaded list, the Persona step is skipped — your audience is defined by who’s in the list. If you’re running from Wyra’s Intelligence database, the Persona step appears. For Phase 2, start with your own uploaded list — tighter control, more readable results.

Running multiple campaigns in Phase 2 is fine— in fact expected if you have two Finalized Offerings. Run one campaign per Offering so results stay readable. What to avoid is running two campaigns against the sameOffering with different persona or message settings simultaneously — you won’t be able to read which variable drove the difference.

Run the campaign for 3–4 weeks before drawing conclusions

The most common Phase 2 mistake is adjusting the campaign in Week 2 because early results are slower than expected. Two weeks is not a reliable signal window. LinkedIn connection acceptance rates take time to develop. Email deliverability improves as the sending infrastructure warms. Measuring email performance at Day 10 produces a number that tells you almost nothing about the campaign’s actual trajectory.

Let the campaign run. Check daily for connection acceptance rates, replies (respond promptly — use the Interactions view’s AI reply suggestions as starting points), and patterns in who’s engaging. Don’t adjust after a few days — look for a consistent pattern across at least two weeks before changing anything.

The first 90 days are not about learning a tool. They’re about building the intelligence foundation that the tool runs on.

Phase 3 is where the intelligence layer starts paying compound returns — because you now have real data from a real campaign to improve from. Phase 3 without the prior work is iteration without a real baseline.

By Day 61, you can have multiple campaigns running or completed. Phase 3 is not about launching a second campaign for the first time — it is about reading the patterns across the campaigns you have run, improving your Offerings based on what you learned, and launching new, calibrated campaigns from that foundation.

Read the Interactions layer with intent

Export the full Interactions CSV from your Phase 2 campaigns. Look for three patterns:

• Which vertical or company type had the highest engagement rate? If healthcare buyers replied at twice the rate of manufacturing buyers, your Offering maps more directly to the healthcare buyer’s situation. Build a healthcare-specific Offering (or sharpen the healthcare sections) before your next campaign in that vertical.
• Which role replied most?If VP of IT replied at higher rates than CTO, adjust persona settings on future campaigns — or examine whether the messaging should shift to speak more directly to executive outcomes.
• What did the early replies actually say?Read them. Fifteen replies with a consistent theme is enough to act on. “We already have a partner for this” is a competitive positioning signal — add a competitive differentiation angle to your Offering’s objections section. “Can you give me more detail on how this works in a regulated environment?” is a signal that compliance knowledge isn’t prominent enough in your messaging.

Improve your Offerings based on what you learned

Make targeted improvements before your next set of campaigns:

• If pain points aren’t landing in specific industries, upload an Artifact that speaks directly to that industry’s specific context
• If objections are appearing consistently in replies, enrich the Objections section with the actual language buyers are using
• If social proof is thin or non-specific, add a case study Artifact with verifiable outcomes and re-run AI enrichment on the Social Proof section

Re-Finalize the Offering after improvements. The version connected to active campaigns is frozen — the improved version becomes the basis for new campaigns. Do not modify an Offering that has active campaigns running against it.

Launch calibrated campaigns from what you’ve learned

Each new campaign from this point should be a deliberate improvement, not a restart. Improve one or two variables based on what the data showed: an updated Offering, adjusted persona settings, or a new targeted list built around the trigger types that produced the best engagement in earlier campaigns.

If patterns showed that one specific vertical dramatically outperformed others, build a dedicated Offering for that vertical. A vertical-specific Offering will consistently outperform a general Offering applied to a narrow list.

Establish your weekly intelligence review rhythm

By Day 90, the platform should be a weekly habit. The rhythm is 20–30 minutes:

• Review new Outreach Ideas generated since the prior week — like or dislike each one
• Check Interactions for replies that need responses
• Review AI SDR activity
• Queue any new lists that are ready for upload

The intelligence layer improves continuously with feedback. Liking and disliking Outreach Ideas trains the model toward the angles that match your actual GTM motion. A team that provides this feedback consistently at the weekly review has a materially better-tuned intelligence layer at Day 180 than a team that provides no feedback at all.

Most teams who see disappointing results in the first 60 days have made one of four errors. These are fixable — but finding them early is far cheaper than discovering them after weeks of flat results.

On reply rate and what to expect

Industry averages for B2B cold outreach run 1–3% across most verticals. Research-led outreach — with a well-structured Offering, targeted list, and calibrated persona — consistently outperforms this range. But early campaign results are not the reliable indicator of where your outreach performance will settle.

The reliable indicator is trajectory: are later campaigns outperforming earlier ones, after the Offering improvements from Phase 2 data have been applied? If yes, the intelligence loop is working and the platform is compounding. If campaigns are flat on an unchanged Offering, the diagnosis is usually upstream — an Offering that hasn’t been refined, or a Why Now that’s still generic. Fix upstream before expanding volume.

Your next 7 days