The Pawanax Engineering Method: Workflow Discovery to Governed AI

The Pawanax engagement method is the four-stage sequence Pawanax uses for AI consulting work: Diagnose, Deploy, Govern and Compound. It starts with AI workflow discovery, moves into a scoped build, and then covers governed AI deployment and a shared review of results. The consultancy page gives each stage a stated outcome and publishes fixed prices for the packages that carry it out. It does not publish client results against those stages.

Four stages in one sequence

The consultancy page describes the approach in one line: "From the first conversation to the next iteration, we build capability alongside the system." Each stage has a short description and a stated outcome.

The order matters because each stage is meant to hand a clear record to the next one. Discovery produces a shared direction. Deployment works against a scope that was agreed in advance. Governance turns a working system into something a team can run. Compounding uses the same baseline to decide what happens next.

The page does not assign each package to a single stage. The packages below are listed as published offerings, and readers should check each package's inclusions against the stages above.

Diagnose: AI workflow discovery before any tooling

Diagnose is the stage the consultancy FAQ names as the usual starting point. The published Discovery Diagnostic runs for two weeks at a fixed fee of $28,000, which the page shows as KES 3,612,000.

The package includes:

• Stakeholder interviews, up to 8
• A data and workflow audit
• Three prioritised use cases with ROI ranges
• An ODPC/CBK risk skim, labelled on the page as non-legal
• A 90-day roadmap
• Three months of Clinical Team seats

Its deliverables are a diagnostic brief in PDF, a use-case scorecard and an executive readout. The ROI ranges are a deliverable of the discovery work. They are estimates prepared during the engagement, not results Pawanax has reported from client deployments.

For a board or leadership team, the written outputs of this stage set out which workflows to pursue, which to leave alone, and what the roadmap assumes. The page presents the Discovery Diagnostic as a way to "make the next decision a better one" before committing to a larger build.

Deploy and Govern: what the published packages deliver

Deploy and Govern are where the published packages do most of their work. The table below lists the five engagement packages on the consultancy page, with duration, price and the deliverables stated in the package data.

Governance appears in the package inclusions as well as the deliverables. The Enterprise Agent OS package lists an architecture and security review, an MLOps and monitoring baseline, a model risk and ODPC-aligned documentation pack, and change management and training. The Clinical AI Pilot includes a Kiswahili and English prompt and governance pack, clinician training and 30 days of hypercare. The WhatsApp Agent Factory includes human-handoff and escalation paths.

For organisations working on AI workflow automation in Kenya, the data protection point matters early. Kenya's Data Protection Act entered into force on 25 November 2019 and establishes the Office of the Data Protection Commissioner, according to Digital Policy Alert. The consultancy FAQ is explicit about the limits of Pawanax's documentation: ODPC-aware documentation "supports the engagement; it is not a substitute for legal advice, regulatory approval, or clinical validation."

Compound: the baseline and the review

Compound is the stage that follows a deployment. The page describes it as a review of measured outcomes, used to decide what to refine, where to expand and what should wait. The outcome it names is "an ongoing partnership grounded in an agreed baseline."

Two published items support that stage. The Clinical AI Pilot delivers a KPI baseline report, and the Enterprise Agent OS delivers a quarterly value review. The ongoing options are three monthly retainers:

An optional outcome fee of 10–20% of verified Year-1 value, capped, can be agreed only against a mutual KPI baseline. Third-party AI usage is billed separately, and ongoing SaaS fees apply after any included product period.

Published offerings and results not yet published

Some parts of the method are published and priced. Others are not yet published as evidence.

Published offerings

• The five engagement packages in the table above, with the prices and durations shown
• The three monthly retainers
• Day rates on the consultancy page: Principal or Engagement Lead at $2,000 (KES 258,000), AI Architect at $1,250 (KES 161,250), and Delivery Engineer at $700 (KES 90,300) per day
• A project minimum of $15,000 (KES 1,935,000)
• The optional capped outcome fee described above

Not yet published

• Client case studies and client delivery results. The consultancy page states that recognitions and product capability "are distinct from client delivery outcomes." The two recognitions it lists are KEMRI KASH, "Best Health System in Kenya" (February 2026), and the Rice University Catalyst Grant (February 2025, through Kenyatta University). Neither is a measured result of the method.
• Outcome figures for any stage of the method, including how much a Compound review has improved a client's results.
• Internal planning figures behind the price list. These are not published.

The phrase "measured outcomes" in the Compound stage describes the review process Pawanax proposes to run with each client. It is not a reported result.

Evidence class for each capability

Pawanax capability claims need an evidence class. The consultancy page labels its capabilities as "Built" or "Designed" and does not name real customer deployments, so none of the capabilities below is classed as Production-validated.

Frequently asked questions

Is the Pawanax engagement method a published offering?

The method is described on the consultancy page as the Pawanax engagement method. The packages that carry it out are published with prices and durations, from the two-week Discovery Diagnostic to the 36–52 week Sovereign AI Program. The method itself is not listed as a separate product with its own price.

What does the Discovery Diagnostic cost and deliver?

It costs $28,000, or KES 3,612,000, as a fixed fee over two weeks. It delivers a diagnostic brief in PDF, a use-case scorecard and an executive readout. It includes up to eight stakeholder interviews, a data and workflow audit and a 90-day roadmap.

Which Pawanax results are not yet published?

Client case studies and client delivery outcomes are not published on the consultancy page. The page states that recognitions and product capability are separate from client delivery outcomes. Outcome figures for the method's stages and internal planning figures are also not published.

Do you only work in healthcare?

Healthcare is the deepest area of focus, according to the consultancy FAQ. Pawanax also says it builds conversational agents and enterprise AI systems for banks, telcos, insurers, NGOs and public institutions across Africa.

Next step

To see which engagement fits your workflow, read the published packages on the Pawanax consultancy page and open the engagement that matches your decision. The people behind the method are listed on the Derrick Macharia profile.

Sources

• Pawanax AI Consultancy: engagements, method and FAQ
• Kenya Data Protection Act 2019 enters into force, establishing the Office of the Data Protection Commissioner (Digital Policy Alert)

Continue exploring

AI Agents in Kenya: A Practical Guide to Enterprise Automation. How Much Does It Cost to Build an AI Agent in Kenya?.

Cover: Photo: Keshie Muchai / Wikimedia Commons, CC BY-SA 4.0 (https://creativecommons.org/licenses/by-sa/4.0/). Illustrative image; not a photograph of a Pawanax system or client. Credit: Keshie Muchai (Wikimedia Commons user; uploaded as own work). CC BY-SA 4.0. Original source.