The failure mode is predictable and we have watched it happen at enough clients to name it precisely. A department stands up one agent on a no-code tool because it demoed well. Then another department does the same, on a different tool, with different credentials and no shared audit trail. Within a year IT is holding thirty ungoverned agent experiments and a backlog they cannot service, and nobody can answer the one question a regulator will ask first: who approved this agent to take this action.
The fix is not to ban no-code agents — the productivity case for citizen-developer agent building is real. The fix is to build the platform and the governance overlay before the second agent ships, not after the thirtieth. B20 selects a platform on evidence, wires identity, RBAC, and audit logging into its foundation, and only then builds the first 1–3 production agents on top of it. Agent-first gets you a demo. Platform-first gets you a place to put the next fifty.
Six streams,
ending in 1–3 shipped agents.
Platform evaluation and foundation build front-load weeks 1–4. Integration and the governance overlay overlap through weeks 3–9. First agents and compliance handover close weeks 6–12.
Platform evaluation & selection
Vendor-neutral scorecard run against your readiness assessment, existing stack, sovereignty requirements, and integration needs. Selection signed before Stream 02 begins.
Foundation build
Identity, RBAC, secrets management, audit logging, and observability wired in from day one — the scaffolding every agent built afterward inherits.
Enterprise integration
ERP, CRM, ITSM, ticketing, and identity provider connectors within scope, built against the platform's gateway rather than point-to-point.
Governance overlay
Approval gates, evaluation harnesses, drift monitoring, and deployment limits — built into the platform, not offered as a configuration option a citizen developer can skip.
First 1–3 production agents
Built from the A11 readiness backlog, tested through UAT with the named business owner for each agent, and cut over on an agreed date.
Compliance evidence & handover
PDPL/ADHICS/AI Charter evidence pack assembled, operator training delivered, and 30/60/90-day post-handover check-ins scheduled.
Twelve weeks maximum.
Eight minimum. Four phases.
Phase count is fixed. Duration flexes with platform procurement lead-time and integration count. Milestones are signed gates — not aspirations.
Platforms scored,
not on analyst-quadrant position.
Every platform decision runs a six-criteria scorecard in weeks 1–2. Each criterion scored 1–5 against evidence — analyst reports are read but never the final word. Signed by your IT lead before Phase 2 begins.
From agent-by-agent
to governed platform.
A typical pre-engagement state has ad-hoc agent experiments scattered across departments — no shared platform, no governance, an IT backlog growing behind a team too small to clear it. The engagement replaces the scatter with a single governed platform and a first wave of shipped agents.
Reference pattern. Some engagements retain an existing agent trial as one of the first three production agents once it passes the governance overlay. What always gets built new is the platform foundation and governance overlay. Never the underlying business systems the agents call into.
A government first-wave,
shipped.
Representative pattern for a UAE government entity under a federal AI-adoption mandate, with three candidate use cases identified through prior readiness discovery. Ranges reflect target outcomes NexITC underwrites in scope for this class of engagement. N=1 — illustrative composite, not a specific client.
Five artifacts,
each with signed acceptance.
Every deliverable has documented acceptance criteria signed at engagement kickoff. Nothing more, nothing less.
Platform Selection Scorecard
Six-criteria evaluation signed by IT leadership, with documented evidence behind every score.
Platform Foundation Build
Identity, RBAC, secrets management, audit logging, and observability wired and tested.
Enterprise Integration
ERP, CRM, ITSM, ticketing, and identity provider connectors, built through the platform gateway.
Governance Overlay & Compliance Evidence Pack
Approval gates, evaluation harnesses, drift monitoring, deployment limits, and the PDPL/ADHICS/AI Charter evidence pack.
First 1–3 Production Agents
The agents that prove the platform works with your workflows, not vendor demos — built, UAT'd, and handed over with operator training.
Six outcome metrics,
measured pre and post.
Success is not “the platform is licensed.” It is measured against six specific outcomes captured in a baseline report at engagement start and re-measured at steady state.
Honest scoping.
B20 is a fit when specific conditions are met. It is not a fit when other conditions are. We say so before the scope conversation, not after the commercial commitment.
A11 or an equivalent discovery exercise has already surfaced candidate use cases. Without it, Phase 1 has nothing concrete to build the first agents against.
Named and scoped, not a general ambition to “do agents.” The platform selection scorecard is graded against real workload, not a hypothetical one.
A named counterpart who owns identity, RBAC, and integration decisions alongside us — this is not a hands-off procurement.
Someone accountable for UAT and sign-off on each of the 1–3 agents shipped in Stream 05.
Available for bi-weekly reviews and empowered to approve the platform decision once the scorecard is signed.
Start with A11 Agentic AI Readiness & Use-Case Discovery.
For one bespoke agent without onward self-service, look at B14 Agentic Workflow Agent Build™.
Sequence B21 Sovereign AI Platform Build alongside B20 when the platform itself must run inside a sovereign boundary.
Eight weeks is our minimum for a defensible governance overlay. We will not compress safety testing to hit a shorter deadline — the gate stays, or the timeline moves.
Fixed fee.
Milestone-based.
Total engagement fee agreed in the scope statement. Not time-and-materials. Not day rate. Every engagement is preceded by a scope conversation to ensure fit before commitment.
Five, most asked.
Q_01Are you tied to a specific agent platform?
Q_02What if we already have an agent platform in place?
Q_03How is this different from B14 Agentic Workflow Agent Build™?
Q_04How does governance work in production?
Q_05What comes next?
One name
on the engagement letter.
A named Practice Lead is accountable for delivery, commercial outcomes, and the client relationship throughout the engagement. Not a project manager who disappears after kickoff. Not a partner who nods at the SOW and vanishes.
Practice Lead — AI
Present at every phase gate, every scope decision, every difficult conversation. Available for 30/60/90-day post-handover check-ins as part of the engagement.
Including scope amendments.
Signs off all 5 deliverables.
With executive sponsor.
Authorised to negotiate.
CEO within 24 hours.
30/60/90-day check-ins.
Prior. Peer. Next.
Agentic AI Readiness & Use-Case Discovery
3-week discovery of agent-suitable use cases. Produces the 1–3 candidate agents B20 builds first. Ideal precondition.
Agentic Workflow Agent Build™
Peer build for one custom-coded agent when platform-based no-code doesn't fit the workflow complexity or when a specific integration needs custom work.
Managed Agent Operations
Ongoing platform operations, agent lifecycle support, drift monitoring, and safety-test cadence after B20 hands over.
Thirty minutes.
No slide deck.
A structured 30-minute scope conversation with the Practice Lead. You describe the candidate use cases, the systems they touch, and who owns each one. We describe whether B20 is the right engagement — and if not, what is.