Essentials
01One agent in production, steady volume.
Coverage, 24 h
Business hours, U.S. Central
Human review, one week
Weekly health review
Forward Deployed Pods
Production is the only demo that counts.
A pod that owns the outcome, measured against a KPI we sign first.
- weeks to production
- 8 to 12
- KPI signed first
- 1
- days to exit, on request
- 5
DaCodes delivery standard, Sep 2026
The model
Not consulting. Not staffing.
Three ways to buy AI delivery. Only one ends in production.
DaCodes
Forward Deployed Pod
- Deliverable
- A system in production
- Consulting: DecksStaffing: Hours
- Ownership
- Of the business outcome
- Consulting: Of the reportStaffing: Of the task
- Where they work
- Inside your operation: your repo, your data
- Consulting: At their officeStaffing: In your backlog
- How it is measured
- The business KPI we signed
- Consulting: Project deliveryStaffing: Utilization
- When it ends
- The system stays, your team is trained
- Consulting: You keep a PDFStaffing: The knowledge leaves
| Dimension | Traditional consulting | Staff augmentation | DaCodesForward Deployed Pod |
|---|---|---|---|
| Deliverable | Decks | Hours | A system in production |
| Ownership | Of the report | Of the task | Of the business outcome |
| Where they work | At their office | In your backlog | Inside your operation: your repo, your data |
| How it is measured | Project delivery | Utilization | The business KPI we signed |
| When it ends | You keep a PDF | The knowledge leaves | The system stays, your team is trained |
Pod anatomy
Three people own the result.
One pod per account, inside your operation.
The pod
- 01
Forward Deployed Engineer
1 FTEBuilds, integrates, owns the outcome.
- 02
Deployment Strategist
0.5 FTEDefines the use case and the metric.
- 03
AI Engineer
1 FTETools, MCPs, prompts and evals.
Reviewed daily by the shared bench
AI Solutions Architect
0.5 FTE
QA and security
0.25 FTE
Customer Success
0.25 FTE
Typical dedication per account. Source: DaCodes pod composition, Sep 2026. Dedication varies with scope.
The four phases
Four phases, one signed number.
Nothing is built before the KPI is signed.
01
Discovery Express
2 sessions · No cost
Two sessions with the people who run the process.
02
Blueprint
1 to 2 weeks · Fixed fee
Workflow, data, architecture, ROI, and the KPI Contract.
03
Pilot
8 to 12 weeks · Fixed by scope
Built in your systems, a demo every week.
04
Scale & Managed
Continuous · Monthly retainer
Handover, then a retainer on evals, cost and drift.
DaCodes delivery standard, Sep 2026. Library agents need 2 to 4 weeks of tuning in the pilot.
Commitment gate · the KPI decidesHatched: the upper end of a rangeKPI Contract
We sign the number first.
- 01
One KPI, one baseline, one target, signed before kickoff.
- 02
Measured on your data, A/B against the baseline.
- 03
At the commitment gate the number decides: no movement, no scale-up, no retainer.
KPI ContractDraft
Example clause
4.2 Committed KPI
Baseline today
35%
Target at the gate
60%
tier 1 tickets resolved without a human
- 4.2.1Baseline and target. 35% of tier 1 tickets resolved without a human today. 60% at the end of the pilot, CSAT at or above the baseline.
- 4.2.2Commitment gate. If the target is missed, the client ends the engagement with no scale-up fee and no retainer, and keeps everything built.
Illustrative. Your numbers come from the Blueprint.
Client
DaCodes
Signed before kickoff
Handover
Your team keeps everything.
Your code, your cloud, your IP, from the first sprint.
- 01
Repo and deployment in your accounts
- 02
Prompts and evals versioned
- 03
Runbooks for rollback and incidents
- 04
Two-day workshop
- 05
Exit plan executed in five days
We'd rather you stay because it works.
Managed services
Someone has to watch the agent.
Models change, data drifts, costs creep. The retainer is mandatory after the pilot.
Agent health · Tier 1 support
Demo dataResolved without a human
12 wk63.6%
Cost per run
12 wk$0.041
This week
- Eval regressions214 of 214 passing
- Model updateRe-evaluated before rollout
- Input driftNew intent: invoice copies
Drift alert sent to #agents-ops · 09:14
Owner: AI Engineer · review in the daily
Professional
02Several agents or a customer-facing channel.
Coverage, 24 h
Extended hours, engineer on call
Human review, one week
Daily review, cost and drift alerts
Enterprise
03Regulated operations or agents that act on money.
Coverage, 24 h
24x7 by agreement
Human review, one week
Continuous, alerts in your channels
Every tier runs eval regressions and a quarterly business review. No prices here by design: the tier is sized in the Blueprint.
Deployment capacity
A full pod for one U.S. salary.
- what a Forward Deployed Engineer opening pays in the U.S.
- $300K to $550K
Source: Alvarez & Marsal and GetPerspective, 2026
- engineers behind every pod
- 250+
Source: DaCodes, Sep 2026
Offices
U.S. Central- Mérida
- Mexico City
- Houston
- Scottsdale
Overlapping hours
Mérida, Mexico City and Houston on U.S. Central. Scottsdale on Mountain time.
English B2+ or C1
Verified in hiring. Demos and runbooks in English.
Entities in Texas and México
You contract in the U.S. We are REPSE registered in México.
Playbook
The Forward Deployed Playbook
PDF12 pagesv1 ships 21 Oct 2026
How we take one process from discovery to a signed KPI in production.
Fill in the form and a person from our team sends it to you once it ships.
Loading the form
FAQ
The questions every operator asks.
Which process do we automate first?
Two discovery sessions, no cost, and you leave with two or three quick wins whether or not you build with us.
We reply in under one business day.
Loading the form