The 30/60/90 pilot plan
What happens between "yes, let us pilot" and the renewal decision, with the shared activities the customer owns as well as the ones we do.
A pilot runs three months. Solution Architect is accountable for onboarding enablement: demos, platform setup and getting the customer's team working in the product. Customer Success owns the relationship and every communication. Product is accountable for adoption, meaning whether it sticks after SA has enabled them.
Use this as the shared plan with the customer. CS sends it in week one and runs the review at each checkpoint. For the full ownership matrix and the end-to-end flow, see Process & Timeline.
The shape
Week 0: before the clock starts
Nothing below begins until these are true. Full list in the Onboarding SOPs.
With a written success criterion, not a free trial. This is the handover artefact.
Masters and adaptations for creative, channels for ORM, accounts for paid. Not credits.
CS is our single point of contact. One SPOC customer-side plus an escalation contact, and the weekly call in the diary.
If their IT gates access. Longest pole in enterprise, so start it now.
Sales walks CS, Product and SA through the problem statement and the success criterion. CS owns the relationship from here.
Platform, ad accounts, social handles, CRM. SA says what is needed; CS is the one who asks the customer.
Weeks 1 and 2: Brand Brain, every module
Every agent reads the graph before it acts, so nothing else can start properly until it exists. Two weeks is the target from the day client inputs land. The three lanes run in parallel and the client lane is the critical path, which is why the request list goes out on day one.
- Day 1Client-lane request list sentSolution Architect
Brand guidelines, legal-approved disclaimers, negative keywords, product assets, layered sub-brand and platform rules, escalation matrix.
- Days 1 to 5Crawl lane runs, zero client effortSolution Architect
Products, descriptions, CTAs, published rates, disclaimers as a seed, competitor set and their ads, review and marketplace sources.
- Days 3 to 8Hybrid lane drafted for validationSolution Architect
Brand voice, personas, content angles and claim boundaries. We draft, the client corrects.
- Days 8 to 12Sources connected and freshness verifiedSolution Architect
No stale feeds before go-live. This is visible on the Sources screen.
- Days 10 to 14Calibration outputs reviewed with the clientProduct + SA
Against their own rulebook. Skills created for anything the defaults keep missing.
- End of week 2Client signs off the graphCustomer Success
Recorded. This is the gate into module configuration.
Exit criteria: graph signed off, sources fresh, calibration accepted, refresh cadence agreed.
Built in parallel, so CS can run the account
Every task with owner, status, blocker and target date. CS answers 'where are we?' from this, without asking anyone.
Actives against seats, surfaces opened, work created, goal movement, usage against estimate. CS spots a stalling account in week three.
Weeks 3 to 12: the module tracks
Run only the tracks the customer bought. Each has its own first wow moment, and hitting that moment early is what carries the pilot.
AI Designer
- Week 3ConfigureSolution Architect
Approval flows mirroring their sign-off chain, compliance rule packs, in-scope apps, Library folders per sub-brand.
- Week 3Wow moment: first real campaign setSolution Architect
One master in, the full adaptation set out, inside a day. This is the moment that sells the pilot internally.
- Weeks 4 to 6Volume through the real workflowSolution Architect
Brief, generate, Compliance Checker, Approvals, Library. Restate in writing that iterations do not count against the cap.
- Weeks 6 to 10Localisation and placement breadthProduct + SA
Second language, second platform, marketplace or print if in scope.
- Weeks 10 to 12Throughput and cycle-time evidenceProduct
Activity screen numbers against the week-1 baseline.
Brand Moderator
- Week 3Collect and configureSolution Architect
Tone guide, approved templates, escalation matrix, complaint taxonomy, privacy masking, blocklist. CRM or ticketing connected.
- Week 4Shadow mode, one full weekProduct + SA
Non-negotiable. The agent drafts, humans approve everything, nothing auto-posts. This is what catches appreciation comments creating tickets.
- Week 5Wow moment: a day cleared to zeroSolution Architect
The queue emptied inside the SLA with drafted replies. The community team feels it immediately.
- Weeks 6 to 10Calibrate and expand channelsProduct + SA
Feed rejections back into rules and playbook. Add channels from the agreed list as they come live.
- Weeks 10 to 12Response time and coverage evidenceProduct
Against the week-4 baseline, plus the Action History compliance trail.
Consumer Insights
- Week 3Connect and configureSolution Architect
Channels, open-web sources, sub-brands, competitor set, executive or spokesperson keywords.
- Week 3Wow moment: the category rankSolution Architect
"You are #7 of 11 and here is who took the share." Nobody has shown them that number before.
- Weeks 4 to 8Themes, spikes and sub-brand spreadProduct
Turn a spike into an explained event, and find the sub-brand quietly dragging the average.
- Weeks 8 to 12The board-ready numberProduct + SA
Agree the single figure the CMO will quote, and the movement since kickoff.
AI Paid Media
- Week 3Connect and reconcileSolution Architect
Ad accounts, GA4, MMP, offline conversions where the cycle needs it. Reconcile spend and conversions against their own platform before anyone sees a number.
- Week 3Goals enteredProduct + SA
From the POC KPI sheet, at the agreed scope. This turns the Overview metrics table into performance against intent.
- Week 4Wow moment: the first real SignalSolution Architect
A critical signal on their own account, at ad-group level, with the evidence and a recommended action. Usually something they had not caught.
- Weeks 5 to 8Morning-report habit switchedSolution Architect
Their daily check moves to Overview. This is the adoption test that matters most.
- Weeks 8 to 12Efficiency and time-returned evidenceProduct
Goal achievement rate plus hours per week returned.
Reporting
- Week 3Capture their formatSolution Architect
If they are replacing an incumbent, get a copy of that report and match it before improving on it.
- Weeks 4 to 5Dashboards and saved reports builtSolution Architect
Role-based splits where leadership and operators need different views.
- Week 5Wow moment: the report arrives without anyone building itSolution Architect
Friday morning, in the inbox, matching their format. For several accounts this is the actual purchase.
- Weeks 6 to 12Confirm delivery, widen distributionProduct
Check the run history rather than assuming. Add leadership to the recipient list.
Checkpoints
Day 30
Completed in weeks 1 to 2, not in progress. If it is not done, everything downstream is at risk.
Our numbers match their platform. Do not add scope until this is true.
At least one per module bought, and the customer has seen it.
Written and agreed. See goal setting.
Access is organisation-wide. If only one person uses it, the rollout has not happened.
With owners and dates, in the shared tracker.
Day 60
Not supplemented, stopped. The morning check, the reply queue, the manual report.
Adoption is the phase-two job and the phase-two failure mode.
Even partial. Direction matters more than magnitude at this point.
Someone has actually received the report, not just been sent it.
Rejections and edits are changing rules, skills and playbooks.
Who will be in the room, and what they need to see.
Day 90
Against the original written criterion, in the customer words.
Not a spike around review time.
Someone else will say it works, unprompted.
Not an estimate. Use the pricing model and the scraping calculator.
Volunteering what did not land is what makes the rest believable.
An honest stop beats a drifting pilot.
Shared activities across all 13 weeks
A pilot where only Zocket has tasks is a pilot the customer is watching rather than running. Put this table in the pilot document.
| Activity | Cadence | Owner | Customer |
|---|---|---|---|
| Weekly customer call | Weekly | CS runs it, goals first. Product brings adoption numbers | SPOC attends, decisions made in the call |
| All customer communication | Continuous | CS, always. PM and SA route through CS | Named escalation contact |
| Onboarding tracker | Continuous | SA maintains status, PM owns the format | Prioritises and signs off |
| Product analytics dashboard | Weekly | PM builds and reads it, CS acts on it | n/a |
| Adoption intervention | Weekly | Product spots the dip, CS makes the call | Chases internal usage |
| Demos and training | Weeks 2 to 8 | SA delivers, CS arranges | Sends the right people |
| Feedback on output | Continuous | SA feeds it into rules, skills, playbooks | Says what was wrong and why |
| Requirements | Months 1 and 2 | PM + SA coordinate and deliver, CS communicates | Prioritises |
| 30 / 60 / 90 review | At each checkpoint | CS runs the room, Product brings the evidence | Validates against their own source |
What kills pilots
- No written success criterion. Everything else follows from this.
- Brand Brain not finished in weeks 1 to 2. Module work on a half-built graph produces output the client rejects, and the pilot never recovers the time.
- Numbers that do not reconcile. Fix before adding scope, always.
- Only the SPOC logs in. No advocate at day 90.
- Reporting configured but never delivered. Check the run history.
- Blockers carried silently. A gap disclosed early is a requirement. The same gap discovered by the customer at day 80 is a credibility problem.
- The economic buyer meets the product for the first time at day 90.
- Scope creep instead of depth. A second module while the first is unproven doubles the surface area and halves the evidence.
Adapting the shape
- Enterprise and BFSI: add two to four weeks before week 1 for security review and access. Kotak Neo's own process is a two to three week evaluation across several team leads before pricing is discussed, so align to theirs rather than imposing ours.
- Creative pilots: scope in masters and adaptations, and state in writing that iterations do not count against the cap. That is the live Lenovo structure and it removes the biggest unspoken fear.
- ORM pilots: the shadow-mode week sits inside weeks 3 to 4 and is not negotiable.
- Agency or multi-stakeholder: issue agency roles in week one. A pilot where the agency cannot log in stalls on the agency timeline rather than yours.