Before the agency brief
Fewer concepts to develop, and better ones. Test the candidates first, then brief the agency on the messages that held up with early adopters, evidence-led physicians and cautious followers alike.
Veltis Foresight: test the decision before the market does.
Foresight runs structured research interviews between an AI interviewer and AI physician personas. Explore how physicians might respond to your message, a competitor launch or new evidence before you commit budget or brief primary research. Twenty candidates in, three or four worth the money out.
- Agents in every interview
- 2
- Persona model layers
- 5
- Evidence tiers
- 4
- Turns before brands are named
- 5
- Excel export sheets
- 7
Twenty ideas in the room. Three worth taking to research.
For brand strategy and insights teams. Foresight augments your market research and your agency rather than replacing either. Candidates from the brand plan meet physician personas and evidence in simulated interviews, and the short list goes to the agency and to primary research.
Before primary research
Arrive with sharper questions. Foresight shows which segments are fragile and which objections come first, so your research budget confirms the decisions that matter instead of exploring everything.
You plan today and launch into tomorrow.
The market you plan in is not the one you launch into. Today's research toolkit is invaluable for understanding the present, and limited when you need to explore what might happen next.
Primary research is slow
It is expensive, takes weeks and measures attitudes that already exist.
Advisory boards are small
Deep and qualitative, but too few voices to test several scenarios.
Sales data looks back
It tells you what happened, not what happens when a guideline changes or a competitor launches.
Will your positioning survive the next guideline update? How will your segments react to a competitor launch? Which segments are fragile, and how would you defend them? Foresight lets you explore these questions before they arrive.
Two agents, one structured interview.
An AI interviewer questions an AI physician whose profile is a random draw within the ranges you set. Both keep their own memory, so the interview flows like a real one, with follow-ups and pushback.
- Choose brands and personasThe brands in your market, and a persona set built for it.
- Set ranges and evidenceTune persona ranges and channel exposure; upload evidence by tier.
- Draw physiciansEach simulated physician is drawn at random within the ranges, and every draw is logged.
- InterviewThe AI interviewer runs a full interview with every physician.
- AnalyzeA checker agent labels each transcript and extracts the KPIs.
- ReportWatch the run live, read the aggregate report, export to Excel.
Simulated interview · rheumatologist personaSample
AI interviewer
Tell me about the last patient you considered moving to a new advanced therapy. What held you back?
AI physician · community practice, cautious adopter
Access, mostly. If prior authorization takes three weeks, I stay with what I know. The efficacy story has to be clearly better, not slightly.
AI interviewer
What would "clearly better" look like to you?
AI physician
Head-to-head data, or at least real-world numbers in patients who failed one biologic. And someone who can tell me what the payer mix looks like in my area.
- Adoption intent
- Low until access improves
- Barrier
- Prior authorization time
- Evidence need
- Head-to-head or real-world data
Physician personas you can tune and audit.
Every simulated physician is built from five layers. You set the ranges; the engine draws within them and logs every draw, so a study can be reviewed and re-run.
Identity
Specialty, demographics and practice context: academic or community, patient volume and payer mix.
Beliefs
Therapeutic preferences, brand attitudes and the evidence each physician trusts most.
Decision logic
How the physician weighs efficacy, safety, access and cost when choosing a treatment.
Memory
What the physician has seen and heard, carried across the interview and across time steps.
Context and openness
How receptive the physician is to new evidence once exposed, from early adopter to cautious follower.
Randomized, logged draws
Each physician is a random draw within your ranges. The draw log travels with the results, so no physician is hand-picked.
Each physician knows only what they would have seen.
Not every physician has seen your latest data. Foresight models that gap, so simulated answers reflect realistic awareness rather than perfect knowledge.
Evidence tiers
Upload trials, papers, guidelines and brand materials, tagged by tier. Each simulated physician retrieves only what their tier permits.
Exposure modeling
Rep visit frequency, congress attendance and channel touchpoints decide how likely each physician is to have seen new evidence.
Knowledge Builder
Trials, guidelines and competitive evidence are structured and indexed for retrieval, so answers are grounded in your material.
Change one assumption. Watch what moves.
Run the same sample of physicians through a sequence of time steps as the market changes. Compare scenarios side by side and see which KPIs shift, in which segments.
Guideline change
Load the expected update and see which segments move first, and which hold.
Competitor launch
Introduce the competitor's evidence and positioning, and measure unaided preference before and after.
New data readout
Test how a new trial result lands with early adopters compared with cautious followers.
Results you can inspect, not a black box.
Every number traces back to a transcript, and every transcript traces back to a logged persona draw.
KPIs by segment
Adoption rate, eligibility threshold, unaided first-mention preference and a 0 to 10 rating for each brand.
Findings and unmet needs
Per-brand findings, recurring objections, unmet needs and quality flags extracted from every interview.
Live run monitor
Watch interviews complete in real time, browse individual simulations and read the aggregate summary.
Full transcripts
Every interview is kept with its prompts, persona draw and extracted findings.
Seven-sheet Excel export
Overview, per simulation, conversations, brand findings, report detail, exposure detail and results by physician archetype.
Editable prompts
Analysis and KPI prompts can be tailored for each client and therapy area.
Studies teams run with Foresight.
We size each study to the decision: from a quick message test to a study across hundreds of simulated physicians.
Message pre-testing
Test several versions of a core message and its supporting claims before they go to agency or MLR.
Competitive response
Explore how each segment responds to a competitor launch, and which of your messages hold up.
Launch readiness
Map the objections physicians raise first, so Coach scenarios and field materials prepare reps for them.
120-question strategy review
A structured study across audience understanding, evidence, value, access, competitive response, objections and channel.
Start with 90 days. We build the physician personas and evidence for your market, then run a series of studies on the decisions in your brand plan.
What Foresight is, and what it is not.
Foresight is simulation. Its personas are AI physicians built from realistic ranges, not real people, and its results are not real physician data.
Use it to explore hypotheses, sharpen questions and pre-test messages quickly. Confirm the decisions that matter with human research, and use Foresight to decide which questions that research should ask.
How we work with you.
Three decisions brand teams test in Foresight, and what a study gives them back.
Launch positioning with twenty candidates
We run every candidate message past hundreds of simulated physicians drawn from your market. You take the three or four that held up to the agency for development and to primary research for confirmation.
A competitor launches next quarter
We load the competitor's evidence and positioning and run the same simulated physicians through the launch. You see which segments move first, which of your messages hold, and what the field needs to say.
A guideline update is expected
We introduce the expected change as a time step and measure adoption intent and unaided preference before and after, segment by segment, so the brand plan is ready before the guideline is.
Bring us the decision you need to test.
We design the study with you: personas, evidence, scenarios and the KPIs that matter to your brand plan.