- Evidence-Based Prompts
- Posts
- š§ Grow Your Consultancy with AI Workflows
š§ Grow Your Consultancy with AI Workflows
Scaling a consultancy means more clients, more project management, and less time to do the actual work. AI workflows can conduct research, design experiments and generate robust insightsāfreeing you to focus on what matters most.

Someone asked me during my last ChatGPT webinar:
āCan you clone yourself into an AI consultant?ā
Scaling a consultancy means more clients, more project management, and more consultantsāunless you automate key parts of the process.
So Iām learning how to create an AI consultant that asks the right questions, challenges assumptions, and delivers evidence-based ideas for behaviour changeājust like a PhD-Level Behavioural Designer.
The key to building effective AI agents?
Mastering AI workflows.
If you want to scale your consultancy, practicing AI workflow techniques like āprompt chainingā is essential.
Consistently, the most successful implementations use simple, composable patterns rather than complex frameworks.
Now Iām building simple workflows to train ChatGPT to handle core consulting tasksāconducting research, designing low-cost experiments, and generating insightsāfreeing me up to enjoy flat whites and build relationships.
To scale AI in your consultancy, practice this skill.
A Consultantās Guide to Prompt Chaining:

šµļø Call 1: Define the Behavioural Challenge
š Prompt: "You are a senior behavioural science consultant guiding a client to define their behavioural challenge using the 5Ws Framework for Behaviour Change (Who, What, When, Where, Why).
Ask the client to answer the following:
Who is the target audience? (Be specific: demographics, psychographics, behavioural traits)
What behaviour do you want to change? (Specify the exact action you want them to take)
When does this behaviour occur? (Situations, triggers, timing)
Where does this behaviour take place? (Context, physical/digital environment)
Why does this behaviour happen or not happen? (Barriers, motivations, underlying drivers)"
š Output: A structured 5Ws definition of the behavioural challenge.
---
š” Call 2: Capture Client Response
Feasibility
š Prompt: Now, collect the clientās responses to the 5Ws Framework and assess whether this is a true behavioural challenge (rather than just an awareness issue or vague behavioural goal).
Check if:
I. The client is targeting a specific behaviour, not just awareness or attitude change.
II. The client has identified at least one clear reason why this is in the best interest of the target audience.
š¢ Decision Gate (Pass/Fail Routing):
PASS ā Move to Call 3 (Explore Hidden Influences with COM-B).
FAIL ā Route to Call 2A (Refine Challenge Definition).
š“ Call 2A: Refine Challenge Definition (Routing if Failed)
š Prompt:"The clientās challenge lacks specificity or is more about awareness than behaviour change.
Guide them to refine it by:
Identifying a concrete action or decision to influence.
Pinpointing a psychological or environmental barrier to overcome.
Ensuring the goal is measurable (e.g., percentage increase in adoption, frequency of behaviour, etc.).Ask them to reframe their challenge based on this."*
š Output: A clearer behavioural challenge.
ā”ļø Return to Call 2 for another Pass/Fail check.
---
šµļø Call 3: Explore Hidden Influences with the COM-B Model
š Prompt: "Now that we have a well-defined behavioural challenge, use the COM-B Model to explore the hidden influences driving or preventing the behaviour.
Break down the challenge by:
Capability (C): Does the target audience have the psychological or physical ability to perform the behaviour?
Opportunity (O): Are there external factors (social norms, environment, access, time) making this behaviour easier or harder?
Motivation (M): What intrinsic and extrinsic motivations drive or hinder this behaviour?
For each component, generate questions for further research and identify key behavioural barriers or enablers.
š Output: A COM-B model breakdown revealing psychological, environmental, and motivational barriers.
---
š” Call 4: Identify Research Needs
š Prompt:"Based on the COM-B analysis, determine what critical unknowns still exist and recommend research methods to fill these gaps."
š Output: A research roadmap linking COM-B model gaps to recommended research methods for client.
š§Ŗ Footnote
Evidence evolves over time and will NOT work across all contextsāstart with a small experiment to find out what works for your specific learning objective.
(P.S. Dear readers, donāt forget to use the coupon code āEVIDENCEā to claim your $50 discount.)