Retention
Make high-value functionality easier for customers to discover and use inside existing workflows.
For private equity & software investors
Make AI part of the value-creation plan.
Vertical SaaS assets are valued on retention, pricing power and defensibility. AI moves all three, but only when it reaches production inside the product customers already pay for. You sponsor the agenda. The work happens inside the portfolio company, with its own CTO and team.
- 8
- AI builds inside products that were already live
- 5
- of them vertical B2B SaaS companies
- 2018
- the year Datics started in data science and AI
Value-creation levers
Retention
Expansion
Defensibility
Operational leverage
ROI and payback
Scored on value, feasibility and payback before anything is built.
Investment case
Where AI can move the investment case
We work backwards from the outcomes an owner underwrites - revenue, retention, operating leverage and payback - not from a technology roadmap.
Expansion
Create intelligent capabilities that increase product value and may support premium functionality or new SKUs.
Defensibility
Combine domain workflows, product logic and customer context in ways generic AI products cannot easily reproduce.
Operational leverage
Automate high-friction work inside the product rather than adding service overhead around it.
ROI and payback
Prioritize AI capabilities where the expected commercial or operating benefit can be measured against the cost, risk and time required to ship.
Value x Feasibility x Payback
Not every AI capability deserves investment. We prioritize the ones with a credible path to production and payback.
Value
How much could the capability affect retention, expansion, pricing power, cost-to-serve and customer adoption?
Feasibility
Can the existing architecture, data, workflows, integrations and permissions support it without disproportionate modernization?
Payback
What is the estimated implementation cost, time to production, expected measurable benefit and likely payback period?
The real constraint
Management usually knows what to do. The product is what gets in the way.
Nothing here is a failure of intent. It is the normal operating reality of a mature vertical SaaS business.
- Management understands the AI agenda, but priorities compete for the same engineers.
- The architecture is mature, so meaningful AI needs product work before model work.
- The roadmap is already committed to customers for the next two quarters.
- The internal team has strong domain engineers and limited AI product experience.
- Experiments get built, demo well and never clear security or procurement.
Our job is to translate an ownership-level AI mandate into a product roadmap the CTO agrees with and a capability customers actually use.
Why sponsors use us
We are comfortable with the products that make AI hard
Legacy architecture, accumulated business logic, integration debt and regulated customers are the normal starting conditions in vertical SaaS. They are also our default operating environment, and often the asset's real AI advantage.
Realistic assessment
We separate AI opportunity that can ship this year from the parts that need architecture work first.
Production bias
The measure is a capability customers use, not a proof of concept in a board deck.
We work with the team, not around it
The portfolio company's engineers keep ownership of their product. We bring the AI product experience they have not had to hire for.
What we do
Sprint, unlock, then capability waves
An AI-Native Sprint sets the baseline and ranks the capability portfolio, unlock changes only what those capabilities require, and each wave ships production AI while building foundations the next wave reuses.
- 01
AI-Native Sprint
Establish the product, workflow and architecture baseline, prioritize the AI capability portfolio and define the first wave.
- 02
Unlock
Change only the parts of the product the prioritized AI capabilities actually require. No speculative rewrites.
- 03
AI Capability Waves
Ship capability into production wave by wave. Each wave reuses the foundations of the last, so later capabilities cost less.
AI opportunity assessment
A grounded read on where AI can create measurable value, the likely commercial or operating upside, the cost and complexity to ship, and the expected path to payback.
Targeted product preparation
Focused changes to data, integrations, permissions and architecture where needed to support the prioritized AI capability safely.
Production AI capabilities
Copilots, agent harnesses and intelligent workflows shipped inside the product customers already use, not demonstrated beside it.
To see what this looks like on screen inside a portfolio company product, view the before and after showcase on what we do.
Portfolio multiplier
Different products. A repeatable AI product playbook.
Every portfolio company has a different codebase and architecture. But opportunity assessment, agent governance, security patterns, permission models, evaluation methods and implementation lessons do not need to start from zero each time. We carry those patterns across relevant vertical SaaS portfolio companies.
Assessment framework
The same structured read on AI opportunity and architecture readiness, applied company by company.
Governance and security patterns
Tenancy, permissions, approvals and audit trails designed once and adapted, not reinvented.
Agent and evaluation practices
Harness design, evaluation and observability practices that transfer even when the stacks do not.
Delivery playbooks
A production path already proven through security review elsewhere in the portfolio.
Packaged offers
Two ways sponsors engage us.
Datics runs as a software factory: fixed scopes, repeatable delivery pods and a written definition of done. Portfolio work is packaged the same way.
Portfolio AI Screen
Fixed fee per company2 weeks per asset
We score each asset on value, feasibility and payback, then rank where AI spend goes first across the portfolio.
A ranked spend plan an investment committee can act on.
Screen a portfolioPortfolio Pod
Monthly retainerShared across 2 to 3 companies
One delivery pod working across several portfolio companies, with the agent harness built once and reused at every asset instead of paid for repeatedly.
Harness cost amortised across the portfolio, not duplicated.
Scope a portfolio podExperience with mature software assets
Brownfield is our standing operating environment, not an occasional engagement: software we did not write, customers who cannot be disrupted and products that must improve without betting the asset on a rewrite. Eight AI builds inside live products, five of them vertical B2B SaaS, were all delivered under those conditions.
Reviewing a vertical SaaS asset?
Start with one portfolio company. We identify the highest-value AI opportunities, estimate the cost, complexity and risk to ship them, and show where the likely return justifies investment.